Complete Handouts Eco607 (1-44)

Mathematical Economics II (ECO607) VU
Lesson 01
INTRODUCTION TO DYNAMIC ANALYSIS
TOPIC 001: NEED AND IMPORTANCE OF DYNAMIC ANALYSIS IN ECONOMICS
Pre-requisites: Basic Mathematical Economics

Types of Economic Analyses: Static, Comparative-Static & Dynamic
Static: ‫ساکن‬, Not moving (over time)
Comparative-Static: ‫مقابلتا ً ساکن‬, Moving (due to exogenous variables)
Dynamic: ‫متحرک‬, Moving (over time)
COMPARISON
# Static/Comparative-Static
Grounds Dynamic Analysis
Analysis
Time dimension is absent as all
Time element has pivotal role.
Time economic variables refer to the same
1 Variables can occur in different
dimension point in time – Static
points of time.
economy/Timeless economy.
Studies the process by which
Studies only a particular point of equilibrium is achieved. Therefore,
2 Equilibrium
equilibrium. it deals with both, equilibrium and
disequilibrium.
Doesn’t show the path of change –
Shows the path of changes – like a
3 Transition only the condition(s) of equilibrium,
‘movie’ of the market.
like a ‘still picture’ of the market.
‘Static model’ evaluates the ‘Dynamic model’ delineates the
endogenous variables that satisfy time path of some variable, on the
Nature of
4 some specified equilibrium basis of the rate of change.
Model
condition(s) e.g.𝑸𝒅 = 𝑸𝒔 , 𝑨𝑫 = 𝑆𝑎𝑣𝑖𝑛𝑔(𝑡𝑖𝑚𝑒) and
𝑨𝑰, 𝑺 = 𝑰 𝐼𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡(𝑡𝑖𝑚𝑒) etc.
Doesn’t fully represent real-life
situations.
Based on the unrealistic assumptions
of perfect competition, perfect Closer to reality.
Real-life
5 knowledge, etc. Considers that all economic
Situation
Important economic variables like variables as changeable.
fashions, population, models of
production, etc. are assumed to be
constant.
6 Complexity Easy but oversimplified Complex but realistic
Rationale for Dynamic Analysis: Time is ‘unstoppable’ and affect variables of all disciplines
including those of economics.
Better to conduct analysis over time rather than one point in time only.
Examples of Dynamic Analysis in Economics:
1. Decline in global GDP due to Covid-19 during first half of 2020.
2. Rise in unemployment in Pakistan due to Covid-19 during first half of 2020.
3. Decline of fuel (Petrol) prices due to Covid-19 during first half of 2020.
4. Change of share value over time.
5. Daily activity in Pakistan Stock Exchange (PSX).
______________________________________________________________________________
©Copyright Virtual University of Pakistan 1
TOPIC 002: DYNAMIC ANALYSIS: EXAMPLE FROM POPULATION

𝟏
𝒅𝑯
Say, population size 𝐻 changes over time at the rate: 𝒅𝒕
= 𝒕 −𝟐
𝟏
𝒅𝑯
What time path of population 𝑯 = 𝑯(𝒕) can yield the rate of change 𝒅𝒕 = 𝒕−𝟐 ?
Here 𝑯(𝒕) can be called time path as ‘population’ depends on ‘time’.
𝟏 𝟏 𝟏
𝐝𝑯(𝒕) 𝟏 𝒅𝑯
If 𝑯(𝒕) = 𝟐𝒕𝟐, then its derivative is 𝒅𝒕
= 𝟐 (𝟐) 𝒕−𝟐 ⇒ 𝒅𝒕
= 𝒕 −𝟐 .
𝟏
𝒅𝑯
Interpretation: If rate of change of population over time is 𝒅𝒕
= 𝒕−𝟐, then the relationship (time
𝟏
path) between population and time is 𝑯(𝒕) = 𝟐𝒕𝟐 .
𝒅𝑯
However, same 𝒅𝒕 can be had with a large number of similar 𝑯(𝒕):
𝟏 𝟏
𝒅𝑯 −
If 𝑯(𝒕) = 𝟐𝒕𝟐 ± 𝟏𝟓 then 𝒅𝒕
= 𝒕 𝟐
𝟏 𝒅𝑯 𝟏
𝑯(𝒕) = 𝟐𝒕𝟐 ± 𝟗𝟗 again 𝒅𝒕 = 𝒕−𝟐
… …
… …
… …
𝟏 𝟏 𝟏
𝒅𝑯
𝑯(𝒕) = 𝟐𝒕𝟐 ± 𝒄 or 𝑯(𝒕) = 𝟐𝒕𝟐 + 𝒄∗ again 𝒅𝒕
= 𝒕 −𝟐
*
instead of mentioning −𝒄, −𝒗𝒆 sign is merged into 𝒄.
𝒄 is used as general notation for all possible constant values.
It’s better to call 𝒄 as the ‘arbitrary constant’.
As arbitrary is something which is based on ‘random choice’, rather than any reason.
𝒄 is arbitrary as it can assume any value at random.
𝟏
𝑯(𝒕) = 𝟐𝒕𝟐 ± 𝒄 is time path of population, however, without determination of 𝑐 the time path will
remain ‘general’ in nature and not ‘definite’.
Therefore, we try to determine value of 𝒄, which is possible.
We need the starting point of the function for it – Initial condition.
Initial condition: Value of the dependent variable when independent variable is zero. Here
dependent variable is 𝑯 and independent variable 𝒕.
Therefore, initial condition for function 𝑯(𝒕) is 𝑯(𝟎) or 𝑯𝒐.
Now for the initial condition, general time path 𝑯(𝒕) becomes:
𝟏 𝟏
𝑯(𝒕) = 𝟐𝒕𝟐 + 𝒄 ⇒ 𝑯(𝟎) = 𝟐(𝟎)𝟐 + 𝒄 ⇒ 𝑯(𝟎) = 𝒄 ⇒ 𝒄 = 𝑯(𝟎)
Value of arbitrary constant 𝒄 is 𝑯(𝟎).
𝟏
Plugging it back in ‘general’ time path: 𝑯(𝒕) = 𝟐𝒕𝟐 + 𝒄
𝟏
We get ‘definite’ time path: 𝑯(𝒕) = 𝟐𝒕𝟐 + 𝑯(𝟎)
𝟏
𝑯(𝒕)
⏟ = 𝟐𝒕
⏟𝟐 + 𝑯(𝟎)
⏟
𝑫𝒆𝒇𝒊𝒏𝒊𝒕𝒆 𝑻𝒆𝒓𝒎 𝒊𝒏𝒗𝒐𝒍𝒗𝒊𝒏𝒈 𝑰𝒏𝒊𝒕𝒊𝒂𝒍
𝑻𝒊𝒎𝒆 𝑷𝒂𝒕𝒉 𝒕𝒊𝒎𝒆 𝑷𝒐𝒑𝒖𝒍𝒂𝒕𝒊𝒐𝒏
Arbitrary constant determined is not arbitrary any more 𝒄 = 𝑯(𝟎).
______________________________________________________________________________
Multiple arbitrary constants can exist for which initial conditions should be available.
Complete Map of the function – Solution of the dynamic model.
We can experiment some value of 𝑯(𝟎). e.g. 𝑯(𝟎) = 𝟏𝟎𝟎.
𝟏
𝑯(𝒕) = 𝟐𝒕𝟐 + 𝟏𝟎𝟎
______________________________________________________________________________
Lesson 02
INTEGRATION AND RULES
TOPIC 003: INTRODUCTION TO INTEGRAL CALCULUS: INDEFINITE INTEGRALS
RATIONALE FOR INTEGRATION:

History Perspective of Integration
Calculus is divided into differential and integral calculus. In this discussion, we are concerned with
integral calculus. Economists often face the mathematical problem of finding a function from its
derivative.
This process of reconstructing a function from its derivative can be regarded as the ‘inverse’ of
differentiation. Mathematicians call this process integration.
There are simple formulas that have been known since ancient times for calculating the area of
any triangle and polygon. Over 4000 years ago, however, the Babylonians were concerned with
accurately measuring the area of plane surfaces, like circles, that are not bounded by straight
lines. Solving this kind of area problem is related to integration. This topic will also include some
important applications of integrals that economists are expected to know.
Mathematical Formulation
Let 𝑭(𝒙) be the ‘primitive/original function’ while 𝒇(𝒙) be the ‘derivative’, rather than 𝒇′ (𝒙).
Under differentiation, process is from primitive function to derived function: 𝑭(𝒙) ⇒ 𝒇(𝒙).
However, in Integration the shoe is on the other foot – From derived function to back to primitive
function. 𝒇(𝒙) ⇒ 𝑭(𝒙). Therefore, Integration is also called anti-derivative.
Elongated ‘S’ ∫ 𝒇(𝒙) 𝒅𝒙 represents the integral sign.
𝒅𝒙 is written after the function here 𝒇(𝒙) showing that function is integrated w.r.t variable 𝒙.
Function to be integrated is known as ‘integrand’.
Differentiation and Integration are opposite of each other:
𝒅
𝑭(𝒙) = 𝒇(𝒙) ⇒ ∫ 𝒇(𝒙) 𝒅𝒙 = 𝑭(𝒙) + 𝒄
𝒅𝒙
𝒄 is arbitrary constant of integration.
Multiple parentage of the integrand – multiple possibilities of 𝒄.
TOPIC 004: BASIC RULES OF INTEGRATION: CONSTANT FUNCTION RULE & CONSTANT
MULTIPLE RULE
1. Constant Function Rule: It is known that a constant function can be written as:
𝒚 = 𝒇(𝒙)
𝒚 = 𝒂𝒙𝟎
Where 𝒂 is a constant.
𝒚=𝒂
Integrating w.r.t. 𝒚 and 𝒙 on L.H.S and R.H.S, respectively:
∫ 𝒚 𝒅𝒚 = ∫ 𝒂 𝒅𝒙
∫ 𝒚 𝒅𝒚 = 𝒂𝒙 + 𝒄
𝒄 is the constant of integration.
Numerical example: Let 𝒚 = 𝒇(𝒙)
𝒇(𝒙) = 𝒂
Let 𝒂 = 𝟏, then 𝒚 = 𝟏.
______________________________________________________________________________
𝒚 = ∫ 𝒇(𝒙) 𝒅𝒙 = ∫(𝟏) 𝒅𝒙 = (𝟏)𝒙 + 𝒄
𝑭(𝒙) = ∫(𝟏) 𝒅𝒙 = 𝒙 + 𝒄
2. Constant Multiple Rule: If a constant (𝒂) is being multiplied with a function, its integration
shall appear like this:
𝒚 = 𝒂 ∙ 𝒇(𝒙)
∫ 𝒚 𝒅𝒚 = ∫{𝒂 ∙ 𝒇(𝒙)} 𝒅𝒙
∫ 𝒚 𝒅𝒚 = 𝒂 ∙ ∫ 𝒇(𝒙) 𝒅𝒙
∫ 𝒚 𝒅𝒚 = 𝒂 ∙ {𝑭(𝒙)} + 𝒄 [∵ ∫ 𝒇(𝒙) 𝒅𝒙 = 𝑭(𝒙)]

And 𝑭(𝒙) is the primitive.
Therefore, similar to differentiation, a constant that appears as a multiple of a function in
an integral, comes out of the integral expression.
Numerical example: Let 𝒂 = 𝟓, then 𝒚 = 𝟓.
𝒚 = 𝟓 ∙ 𝒇(𝒙)
∫ 𝒚 𝒅𝒚 = ∫{𝟓 ∙ 𝒇(𝒙)} 𝒅𝒙
∫ 𝒚 𝒅𝒚 = 𝟓 ∙ ∫ 𝒇(𝒙) 𝒅𝒙
∫ 𝒚 𝒅𝒚 = 𝟓 ∙ {𝑭(𝒙)} + 𝒄 [∵ ∫ 𝒇(𝒙) 𝒅𝒙 = 𝑭(𝒙)]

Visualizing the integrand before after integration i.e. 𝒇(𝒙) and 𝑭(𝒙).
______________________________________________________________________________
Before integration the function was a constant function with degree zero (𝟎) and after it has
become a linear function with degree one (𝟏).
TOPIC 005: POWER RULE OF INTEGRATION
Power Rule: When one is to integrate a function that has power other than −𝟏, power rule is
applicable:
Assuming that the power is symbolized by 𝒏:
𝒙𝒏+𝟏
∫ 𝒙𝒏 𝒅𝒙 = 𝒏+𝟏
+ 𝒄 , where 𝒏 ≠ −𝟏
Parametric Stipulation: Now it is clear why 𝒏 ≠ −𝟏. Putting 𝒏 = −𝟏 in the formula:
𝒙𝒏+𝟏
∫ 𝒙𝒏 𝒅𝒙 = +𝒄
𝒏+𝟏
𝒙−𝟏+𝟏
∫ 𝒙(−𝟏) 𝒅𝒙 = +𝒄
−𝟏 + 𝟏
𝒙𝟎
∫ 𝒙(−𝟏) 𝒅𝒙 = +𝒄
𝟎
𝟏
∫ 𝒙(−𝟏) 𝒅𝒙 = + 𝒄
𝟎
∫ 𝒙(−𝟏) 𝒅𝒙 = ∞ + 𝒄
∫ 𝒙(−𝟏) 𝒅𝒙 = ∞ (undefined)
Numerical Example: 𝒚 = 𝒇(𝒙) = 𝒙𝟑
Integrating on both sides
∫ 𝒚 𝒅𝒚 = ∫ 𝒙𝟑 𝒅𝒙
______________________________________________________________________________
𝒙𝟑+𝟏
∫ 𝒚 𝒅𝒚 = +𝒄
𝟑+𝟏
𝒙𝟒
∫ 𝒚 𝒅𝒚 = +𝒄
𝟒
𝒙𝟒
Visualizing the pre and post graphs of the integrand; i.e. 𝒇(𝒙) = 𝒙𝟑 and 𝑭(𝒙) = 𝟒
+ 𝒄 with
(𝒄 = 𝟏).
Power Rule and Constant Function Rule

One can use power rule to derive constant function rule:
Assuming the standard form of constant function:
𝒚=𝒂
Where 𝒂 = 𝟏 is a constant.
Rewriting the function: 𝒚 = (𝟏)𝒙𝟎
Integrating w.r.t x on R.H.S. ∫ 𝒚 𝒅𝒚 = ∫(𝒙𝟎 ) 𝒅𝒙
𝒙𝟎+𝟏
Applying power rule: = 𝟎+𝟏
+𝒄
𝒙𝟏
= +𝒄
𝟏
∫ 𝒚 𝒅𝒚 = 𝒙 + 𝒄
Which is the constant function rule with constant (𝒂) = 𝟏.
______________________________________________________________________________
TOPIC 006: DERIVATIVE OVER FUNCTION RULE OF INTEGRATION
When the integrand happens to be in a ratio between derivative and its primitive, then the
integration is done using Derivative over Function Rule.
𝒇′ (𝒙)
∫ 𝒅𝒙 = 𝐥𝐧|𝒇(𝒙)| + 𝐜
𝒇(𝒙)
The rule states that if integrand is a derivative-primitive ratio, then the integration is the natural
logarithm of the primitive.
In other words, the primitive lies in denominator while its derivative is placed in the numerator.
𝟒𝒙
Numerical Example(s): Assuming a function: 𝒚 = 𝒙𝟐 +𝟏
Integrating w.r.t. 𝒙 on R.H.S.
𝟒𝒙
∫ 𝒚 𝒅𝒚 = ∫ 𝟐 𝒅𝒙
𝒙 +𝟏
As per the rule the primitive lies in denominator, therefore, 𝒇(𝒙) = 𝒙𝟐 + 𝟏 The derivative of 𝒇(𝒙)
is 𝒇′ (𝒙) = 𝟐𝒙.
The numerator does not have exact term of 𝒇′ (𝒙) i.e. 𝟐𝒙. Therefore, we create it by re-writing the
numerator.
𝟐(𝟐𝒙)
∫ 𝒚 𝒅𝒚 = ∫ 𝟐 𝒅𝒙
𝒙 +𝟏
𝟐𝒙
∫ 𝒚 𝒅𝒚 = 𝟐 ∫ 𝟐 𝒅𝒙
𝒙 +𝟏
𝒅 𝟐
𝒅𝒙 (𝒙 + 𝟏)
∫ 𝒚 𝒅𝒚 = 𝟐 ∫ 𝒅𝒙
(𝒙𝟐 + 𝟏)
Recalling and applying the derivative over function rule:

𝒇′ (𝒙)
∫ 𝒅𝒙 = 𝐥𝐧|𝒇(𝒙)| + 𝐜
𝒇(𝒙)
𝟐𝒙
∫ 𝒚 𝒅𝒚 = 𝟐 ∫ 𝟐 𝒅𝒙
𝒙 +𝟏
∫ 𝒚 𝒅𝒚 = 𝟐(𝒍𝒏|𝒙𝟐 + 𝟏|) + 𝒄
𝟒𝒙
Visualizing the pre and post graphs of the integrand; 𝒇(𝒙) = 𝒙𝟐+𝟏 and 𝑭(𝒙) = 𝟐(𝒍𝒏|𝒙𝟐 + 𝟏|) + 𝒄
with (𝒄 = 𝟏).
______________________________________________________________________________
______________________________________________________________________________
Lesson 03
INTEGRATION AND RULES (CONTINUED)
TOPIC 007: EXPONENTIAL FUNCTION RULE OF INTEGRATION
Exponential Function Rule states that if integrand is an exponential function, then its integration
is equal to:
𝒇(𝒙)
𝒂𝒇(𝒙)
∫𝒂 𝒅𝒙 = ′ +𝒄
𝒇 (𝒙) 𝐥𝐧|𝒂|
Numerical Example:
𝟐𝟑𝒙
∫ 𝟐𝟑𝒙 𝒅𝒙 =
𝒅
(𝟑𝒙) ∙ 𝐥𝐧|𝟐|
𝒅𝒙
𝟑𝒙
𝟐𝟑𝒙
∫ 𝟐 𝒅𝒙 =
𝟑 ∙ (𝟎. 𝟔𝟗𝟑)
𝟐𝟑𝒙 𝟐𝟑𝒙
∫ 𝟐𝟑𝒙 𝒅𝒙 = 𝟏.𝟑𝟖𝟔
+ 𝒄 or ∫ 𝟐𝟑𝒙 𝒅𝒙 = 𝟑 ∙ 𝐥𝐧|𝟐| + 𝒄
𝟐𝟑𝒙
Visualizing the pre and post graphs of the integrand; i.e. 𝒇(𝒙) = 𝟐𝟑𝒙 and 𝑭(𝒙) = with
𝟑 ∙ 𝐥𝐧|𝟐|
(𝒄 = 𝟎).
TOPIC 008: NATURAL EXPONENTIAL FUNCTION RULE OF INTEGRATION
𝒆𝒇(𝒙) 𝒆𝒇(𝒙)
Natural Exponential Function Rule: ∫ 𝒆𝒇(𝒙) 𝒅𝒙 = 𝒇′ (𝒙) 𝐥𝐧|𝒆| + 𝒄 = 𝒇′ (𝒙) + 𝒄
Natural Exponential Function Rule states that if integrand is a natural exponential function, then
its integration is calculated similar to Exponential Function Rule:
𝒆𝒇(𝒙)
∫ 𝒆𝒇(𝒙) 𝒅𝒙 = +𝒄
𝒇′ (𝒙) 𝐥𝐧|𝒆|
______________________________________________________________________________
𝒆𝒇(𝒙)
∫ 𝒆𝒇(𝒙) 𝒅𝒙 = +𝒄
𝒇′ (𝒙) × 𝟏
𝒆𝒇(𝒙)
∫ 𝒆𝒇(𝒙) 𝒅𝒙 =+𝒄
𝒇′ (𝒙)
Numerical Example: Assuming a numerical example: 𝒚 = 𝟏𝟑𝒆𝟒𝒙
Integrating w.r.t. 𝒙 on R.H.S.
= ∫ 𝟏𝟑𝒆𝟒𝒙 𝒅𝒙
𝒆𝟒𝒙
= 𝟏𝟑 { }
𝒅
(𝟒𝒙)
𝒅𝒙
𝒆𝟒𝒙
= 𝟏𝟑 ( )
𝟒
𝟏𝟑 𝟒𝒙
∫ 𝟏𝟑𝒆𝟒𝒙 𝒅𝒙 = 𝒆 +𝒄
𝟒
𝟏𝟑 𝟒𝒙
Visualizing the pre and post graphs of the integrand, i.e. 𝒇(𝒙) = 𝟏𝟑𝒆𝟒𝒙 and 𝑭(𝒙) = 𝒆 +𝒄
𝟒
with (𝒄 = 𝟏).
TOPIC 009: SUM-DIFFERENCE RULE OF INTEGRATION
Sum-Difference rule states that the integration of an integral which is either a sum/difference or
two or more functions, is equal to the sum/difference of the integrals of those functions.
For convenience we assume only two functions here; i.e. 𝒇(𝒙) = 𝒈(𝒙) ± 𝒉(𝒙)
∫ 𝒇(𝒙) 𝒅𝒙 = ∫{𝒈(𝒙) ± 𝒉(𝒙)} 𝒅𝒙 = ∫ 𝒈(𝒙) 𝒅𝒙 ± ∫ 𝒉(𝒙) 𝒅𝒙 + 𝒄
Numerical Example(s): Assuming a function:
𝒇(𝒙) = (𝟑𝒙𝟑 − 𝒙 + 𝟏)
Integrating w.r.t 𝒙.
∫ 𝒇(𝒙) 𝒅𝒙 = ∫(𝟑𝒙𝟑 − 𝒙 + 𝟏) 𝒅𝒙
______________________________________________________________________________
Here, instead of two functions, there are three functions; i.e. ∫ 𝒈(𝒙), 𝒉(𝒙) and 𝒊(𝒙). Moreover,
both of the arithmetic operations i.e. sum and difference, are used. Therefore, the formula shall
be:
∫ 𝒇(𝒙) 𝒅𝒙 = ∫{𝒈(𝒙) − 𝒉(𝒙) + 𝒊(𝒙)} 𝒅𝒙
∫{𝒈(𝒙) − 𝒉(𝒙) + 𝒊(𝒙)} 𝒅𝒙 = ∫ 𝒈(𝒙) 𝒅𝒙 − ∫ 𝒉(𝒙) 𝒅𝒙 + ∫ 𝒊(𝒙) 𝒅𝒙 + 𝒄
= ∫ 𝟑𝒙𝟑 𝒅𝒙 − ∫ 𝒙 𝒅𝒙 + ∫ 𝟏 𝒅𝒙
= 𝟑 ∫ 𝒙𝟑 𝒅𝒙 − ∫ 𝒙 𝒅𝒙 + ∫ 𝟏 𝒅𝒙
𝒙𝟑+𝟏 𝒙𝟏+𝟏
= 𝟑( )− +𝒙+𝒄
𝟑+𝟏 𝟏+𝟏
𝟑𝒙𝟒 𝒙𝟐
𝑭(𝒙) = − +𝒙+𝒄
𝟒 𝟐
𝟑𝒙𝟒
Visualizing the pre and post graphs of the integrand, i.e. 𝒇(𝒙) = (𝟑𝒙𝟑 − 𝒙 + 𝟏) and 𝑭(𝒙) = 𝟒
−
𝒙𝟐
+ 𝒙 + 𝒄 with (𝒄 = 𝟏).
𝟐
______________________________________________________________________________
TOPIC 010: MORE COMPLEX RULES: INTEGRATION BY SUBSTITUTION
Integration of a product or quotient of two differentiable functions can be done via ‘Integration by
Substitution’.
If integrand can be expressed as a constant-multiple 𝒏 of another function 𝒖 and its derivative
𝒅𝒖
𝒅𝒙
.
Integration of a product or quotient of two differentiable functions of 𝒙, such as:
∫ 𝒇(𝒙) 𝒅𝒙 ⇒ ∫ 𝒏 ∙ {𝒈(𝒙) ∙ 𝒉(𝒙)} 𝒅𝒙
Elaborating the expression:
𝑝𝑟𝑜𝑑𝑢𝑐𝑡 𝑜𝑓 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
( )
𝑎𝑛𝑑 𝑖𝑡𝑠 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒
⏞
∫ 𝒇(𝒙) 𝒅𝒙 ⇒ ∫ 𝒏
⏟ ∙ 𝒈(𝒙) ∙ 𝒉(𝒙)
⏟ 𝒅𝒙
(≠0)
ℎ ′ (𝑥)
{ }
𝑝𝑟𝑜𝑑𝑢𝑐𝑡 𝑜𝑓 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
( )
𝑎𝑛𝑑 𝑖𝑡𝑠 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒
⏞′ (𝒙)
∫ 𝒇(𝒙) 𝒅𝒙 ⇒ ∫ 𝒏
⏟ ∙ 𝒉 ∙ 𝒉(𝒙) 𝒅𝒙
(≠0)
{ }
Method of substitution: Breakup 𝒇(𝒙) into a product of two function – 𝒖 = 𝒉(𝒙) and its derivative
𝒅𝒖
𝒅𝒙
= 𝒉′ (𝒙).
= ∫ 𝒇(𝒙) 𝒅𝒙
𝒅𝒖
= ∫ 𝒖 ( ) 𝒅𝒙
𝒅𝒙
= ∫ 𝒖 𝒅𝒖
= 𝑭(𝒖) + 𝒄
Hack: Look for the function and its derivative in the function.
Example 5: ∫ 𝟏𝟐𝒙𝟐 (𝒙𝟑 + 𝟐) 𝒅𝒙

𝑖𝑡𝑠 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒
(𝑤𝑖𝑙𝑙 𝑏𝑒 𝑞𝑢𝑎𝑑𝑟𝑎𝑡𝑖𝑐 )
Intuition: ∫ 𝟏𝟐𝒙𝟐
⏟ ⏟ ⏞
(𝒙𝟑 + 𝟐) 𝒅𝒙
𝑞𝑢𝑎𝑑𝑟𝑎𝑡𝑖𝑐 𝑐𝑢𝑏𝑖𝑐
( ) (𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛)
𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
𝟑
Derivative of (𝒙 + 𝟐):
𝒅 𝟑
(𝒙 + 𝟐) = 𝟑𝒙𝟐
𝒅𝒙
However, in the given example, we have 𝟏𝟐𝒙𝟐 instead of 𝟑𝒙𝟐 .
So, we breakup 𝟏𝟐𝒙𝟐 into 𝟒 × 𝟑𝒙𝟐 .
= ∫ 𝟏𝟐𝒙𝟐 (𝒙𝟑 + 𝟐) 𝒅𝒙
= ∫⏟ ⏟𝟐 (𝒙
𝟒 × 𝟑𝒙 ⏟ 𝟑 + 𝟐) 𝒅𝒙
𝑛 𝑑𝑢 𝑢
( )
𝑑𝑥
𝒅𝒖
= ∫ 𝟒 × ( ) 𝒖 𝒅𝒙
𝒅𝒙
______________________________________________________________________________
= 𝟒 ∫(𝒖) 𝒅𝒖
𝒖𝟐
= 𝟒(
)+𝒄
𝟐
= 𝟐𝒖𝟐 + 𝒄
=𝟐 𝒖⏟ 𝟐+𝒄
(𝑥 3 +2)2
Restoring the expression in terms of 𝒙:
𝟐
∫ 𝟏𝟐𝒙𝟐 (𝒙𝟑 + 𝟐) 𝒅𝒙 = 𝟐(𝒙𝟑 + 𝟐) + 𝒄
Verification (via differentiation):
𝒅 𝟐
{𝟐(𝒙𝟑 + 𝟐) + 𝒄}
𝒅𝒙
𝒅 𝟐 𝒅
{𝟐(𝒙𝟑 + 𝟐) + 𝒄} = 𝟐 ∙ {𝟐(𝒙𝟑 + 𝟐)} { (𝒙𝟑 + 𝟐)} + 𝟎
𝒅𝒙 𝒅𝒙
𝒅
{∫ 𝟏𝟐𝒙𝟐 (𝒙𝟑 + 𝟐) 𝒅𝒙} = 𝟒(𝒙𝟑 + 𝟐)(𝟑𝒙𝟐 )
𝒅𝒙
𝟐 𝟑 𝟐 𝟑
𝟏𝟐𝒙
⏟ (𝒙 + 𝟐) = 𝟒(𝟑𝒙 )(𝒙 + 𝟐)
𝑜𝑟𝑖𝑔𝑖𝑛𝑎𝑙 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
𝟏𝟐𝒙𝟐 (𝒙𝟑 + 𝟐) = ⏟
𝟏𝟐𝒙𝟐 (𝒙𝟑 + 𝟐)
𝑜𝑟𝑖𝑔𝑖𝑛𝑎𝑙 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
𝑖𝑠 𝑟𝑒𝑠𝑡𝑜𝑟𝑒𝑑
TOPIC 011: MORE COMPLEX RULES: INTEGRATION BY PARTS
Rationale for integration by parts:

Example 5: ∫ 𝟒𝒙(𝒙 + 𝟏)𝟑 𝒅𝒙
(𝑖𝑡𝑠 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑤𝑖𝑙𝑙)
𝑛𝑜𝑡 𝑏𝑒 𝑙𝑖𝑛𝑒𝑎𝑟
Intuition: ∫ 𝟒𝒙
⏟ ⏟ ⏞
(𝒙 + 𝟏)𝟑 𝒅𝒙
𝑙𝑖𝑛𝑒𝑎𝑟 𝑐𝑢𝑏𝑖𝑐
(𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛) (𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛)
(𝒙 𝟑
Derivative of + 𝟏) :
𝒅
(𝒙 + 𝟏)𝟑 = 𝟑(𝒙 + 𝟏)𝟐
𝒅𝒙
However, in the given example, we have 𝟒𝒙
⏟ instead of ⏟𝟑(𝒙 + 𝟏)𝟐 .
𝑙𝑖𝑛𝑒𝑎𝑟 𝑞𝑢𝑎𝑑𝑟𝑎𝑡𝑖𝑐
Coefficient 𝒏 can be developed however, degree of function cannot be changed.
As Integration by Substitution fails here, we resort to a superior alternative ‘Integration by Parts’.
Formula for Integration by Parts:

Used when integrand is a product or quotient of differentiable functions of 𝑥 and Integration by
substitution fails.
______________________________________________________________________________
Derivation of the Rule: Reversing Product Theorem

𝒅
{𝒈(𝒙)𝒉(𝒙)} = 𝒈(𝒙)𝒉′ (𝒙) + 𝒉(𝒙)𝒈′ (𝒙)
𝒅𝒙
Taking Integral:
𝒅
∫ {𝒈(𝒙)𝒉(𝒙)} 𝒅𝒙 = ∫{𝒈(𝒙)𝒉′ (𝒙)} 𝒅𝒙 + ∫{𝒉(𝒙)𝒈′ (𝒙)} 𝒅𝒙
𝒅𝒙
{𝒈(𝒙)𝒉(𝒙)} = ∫{𝒈(𝒙)𝒉′ (𝒙)} 𝒅𝒙 + ∫{𝒉(𝒙)𝒈′ (𝒙)} 𝒅𝒙
𝒈(𝒙)𝒉(𝒙) − ∫{𝒈(𝒙)𝒉′ (𝒙)} 𝒅𝒙 = ∫{𝒉(𝒙)𝒈′ (𝒙)} 𝒅𝒙
∫{𝒉(𝒙)𝒈′ (𝒙)} 𝒅𝒙 = 𝒈(𝒙)𝒉(𝒙) − ∫{𝒈(𝒙)𝒉′ (𝒙)} 𝒅𝒙

⏟
𝐹𝑜𝑟𝑚𝑢𝑙𝑎 𝑓𝑜𝑟 ‘𝐼𝑛𝑡𝑒𝑔𝑟𝑎𝑡𝑖𝑜𝑛 𝑏𝑦 𝑃𝑎𝑟𝑡𝑠’
Example: ∫ 𝟒𝒙(𝒙 + 𝟏)𝟑 𝒅𝒙

Hack: Consider smaller function as 𝒉(𝒙), while more complicated one as 𝒈′ (𝒙).
∫ 𝟒𝒙
⏟ ⏟ + 𝟏)𝟑 𝒅𝒙
(𝒙
(𝑠𝑖𝑚𝑝𝑙𝑒𝑟) 𝑚𝑜𝑟𝑒
(𝑐𝑜𝑚𝑝𝑙𝑖𝑐𝑎𝑡𝑒𝑑 )
∫ 𝟒𝒙
⏟ ⏟(𝒙 + 𝟏)𝟑 𝒅𝒙
ℎ(𝑥) 𝑔′ (𝑥)
Formula for Integration by Substitution:
𝑔𝑖𝑣𝑒𝑛 𝑔𝑖𝑣𝑒𝑛 𝑔𝑖𝑣𝑒𝑛
⏞⏞
∫ {𝒉(𝒙) 𝒈′ (𝒙)} 𝒅𝒙 = 𝒈(𝒙) ⏞ − ∫{ 𝒈(𝒙)
⏟ 𝒉(𝒙) ⏟ 𝒉 ⏟′ (𝒙) } 𝒅𝒙
𝑚𝑖𝑠𝑠𝑖𝑛𝑔 𝑚𝑖𝑠𝑠𝑖𝑛𝑔 𝑚𝑖𝑠𝑠𝑖𝑛𝑔
Given extracted terms; 𝒉(𝒙) = 𝟒𝒙 , 𝒈′ (𝒙) = (𝒙 + 𝟏)𝟑 .

Missing Terms
𝒈(𝒙) 𝒉′ (𝒙)
𝒅
𝒈(𝒙) = ∫{𝒈′ (𝒙)} 𝒅𝒙 𝒉′ (𝒙) = {𝒉(𝒙)}
𝒅𝒙
𝒅
𝒈(𝒙) = ∫(𝒙 + 𝟏)𝟑 𝒅𝒙 = {𝟒𝒙}
𝒅𝒙
(𝒙 + 𝟏)𝟒 𝒉′ (𝒙) = 𝟒
𝒈(𝒙) = + 𝒄𝟏
𝟒
Plugging values in the formula:

⏞ 𝒈′ (𝒙)} 𝒅𝒙 = 𝒈(𝒙)
∫ {𝒉(𝒙) ⏞ − ∫{𝒈(𝒙)
⏟ 𝒉(𝒙) ⏟ 𝒉′ (𝒙)} 𝒅𝒙
(𝒙+𝟏)𝟒 𝟒
⏞ − ∫ [{(𝒙+𝟏) + 𝒄𝟏 } 𝟒] 𝒅𝒙
⏞ ∙ (𝒙 + 𝟏)𝟑 𝒅𝒙 = {
∫ 𝟒𝒙 + 𝒄𝟏 } (𝟒𝒙)
⏟ 𝟒 ⏟ 𝟒
______________________________________________________________________________
(𝒙+𝟏)𝟒 (𝒙+𝟏)𝟒
∫ 𝟒𝒙 ∙ (𝒙 + 𝟏)𝟑 𝒅𝒙 = {𝟒𝒙 𝟒
+ 𝟒𝒄𝟏 𝒙} − ∫ {𝟒 𝟒
+ 𝟒𝒄𝟏 } 𝒅𝒙
= {𝒙(𝒙 + 𝟏)𝟒 + 𝟒𝒄𝟏 𝒙} − ∫{(𝒙 + 𝟏)𝟒 }𝒅𝒙 − ∫ 𝟒𝒄𝟏 𝒅𝒙

(𝒙+𝟏)𝟓
= 𝒙(𝒙 + 𝟏)𝟒 + 𝟒𝒄𝟏 𝒙 − { 𝟓
+ 𝒄𝟐 } − (𝟒𝒄𝟏 𝒙 + 𝒄𝟑 )
(𝒙+𝟏)𝟓
= 𝒙(𝒙 + 𝟏)𝟒 + 𝟒𝒄𝟏 𝒙 − − 𝒄𝟐 − 𝟒𝒄𝟏 𝒙 − 𝒄𝟑
𝟓
𝑦𝑒𝑡 𝑎𝑛𝑜𝑡ℎ𝑒𝑟
𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 (𝒄)
(𝒙 + 𝟏)𝟓 ⏞
= 𝒙(𝒙 + 𝟏)𝟒 −
− (𝒄𝟐 + 𝒄𝟑 )
𝟓
𝟑 𝟒
(𝒙 + 𝟏)𝟓
(𝒙
∫ 𝟒𝒙 ∙ + 𝟏) 𝒅𝒙 = 𝒙(𝒙 + 𝟏) − −𝒄
𝟓
Verification (via differentiation):
𝒅 (𝒙 + 𝟏)𝟓
{𝒙(𝒙 + 𝟏)𝟒 − − 𝒄}
𝒅𝒙 𝟓
𝑝𝑟𝑜𝑑𝑢𝑐𝑡 𝑟𝑢𝑙𝑒 𝑝𝑜𝑤𝑒𝑟 𝑟𝑢𝑙𝑒 𝑝𝑟𝑜𝑑𝑢𝑐𝑡 𝑟𝑢𝑙𝑒 𝑝𝑜𝑤𝑒𝑟 𝑟𝑢𝑙𝑒
𝒅 ⏞
(𝒙+𝟏)𝟓 ⏞
𝟓(𝒙+𝟏)𝟒
𝒅𝒙
{⏞
𝒙(𝒙 + 𝟏)𝟒 − 𝟓
− 𝒄} = ⏞
𝒙 ∙ 𝟒(𝒙 + 𝟏)𝟑 + 𝟏 ∙ (𝒙 + 𝟏)𝟒 − 𝟓
−𝟎
𝒅
{∫ 𝟒𝒙(𝒙 + 𝟏)𝟑 𝒅𝒙} = 𝟒𝒙 ∙ (𝒙 + 𝟏)𝟑 + (𝒙 + 𝟏)𝟒 − (𝒙 + 𝟏)𝟒
𝒅𝒙
𝟒𝒙(𝒙 + 𝟏)𝟑 = ⏟
⏟ 𝟒𝒙 ∙ (𝒙 + 𝟏)𝟑
𝑜𝑟𝑖𝑔𝑖𝑛𝑎𝑙 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝑜𝑟𝑖𝑔𝑖𝑛𝑎𝑙 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
𝑖𝑠 𝑟𝑒𝑠𝑡𝑜𝑟𝑒𝑑
______________________________________________________________________________
Lesson 04
DEMAND AND SUPPLY ANALYSIS USING INTEGRATION
TOPIC 012: DEMAND ANALYSIS USING INTEGRATION BY SUBSTITUTION
The price 𝒑 (dollars) of each unit of a particular commodity is estimated to be changing at the
𝒅𝒑 −𝟏𝟑𝟓𝒙
rate: 𝒅𝒙 = 𝟐
. Here 𝒙 is the Consumer demand in (hundred) units. If demand is (𝒙 =) 𝟒 then
√𝟗+𝒙
price would be (𝒑 =) 𝑷𝑲𝑹 𝟑𝟎. Find demand function 𝒑(𝒙).
a. Find the demand function 𝒑(𝒙).
b. At what price will 300 units be demanded? At what price will no units be demanded?
c. How many units are demanded at a price of PKR 20 per unit?
𝒅𝒑 −𝟏𝟑𝟓𝒙
Given the rate of change of inverse demand function; 𝒅𝒙 = 𝟐
, integration can give inverse
√𝟗+𝒙
demand function:
𝒅𝒑 −𝟏𝟑𝟓𝒙 −𝟏𝟑𝟓𝒙
∫ 𝒅𝒙 = ∫ 𝒅𝒙 ⇒ 𝒑 = ∫ 𝒅𝒙
𝒅𝒙 √𝟗 + 𝒙𝟐 √𝟗 + 𝒙𝟐
−𝟏𝟑𝟓𝒙
𝒑(𝒙) = ∫ 𝒅𝒙 using integration by substitution as basic rules of integration don’t work here.
√𝟗+𝒙𝟐
Let, the relatively higher degree polynomial be substituted:
𝒖 = 𝟗 + 𝒙𝟐
𝒅𝒖 𝒅𝒖
Which implies that: = 𝟐𝒙 ⇒ = 𝒙 𝒅𝒙
𝒅𝒙 𝟐
−𝟏𝟑𝟓𝒙 −𝟏𝟑𝟓
Substituting in the integrand, we get: 𝒑(𝒙) = ∫ 𝒅𝒙 = ∫ (𝒙 𝒅𝒙)
√𝟗+𝒙𝟐 √𝟗+𝒙𝟐
𝟏
−𝟏𝟑𝟓 𝒅𝒖 𝟏𝟑𝟓 −
𝟏 𝟏𝟑𝟓 𝒖−𝟐+𝟏
𝒑(𝒙) = ∫ ( )=− ∫ 𝒖 𝟐 𝒅𝒖 = − ( )+𝒄
√𝒖 𝟐 𝟐 𝟐 −𝟏 + 𝟏
𝟐
𝟏
𝟏𝟑𝟓 𝒖𝟐 𝟏𝟑𝟓 𝟏
𝒑(𝒙) = − ( )+𝒄 = − × 𝟐 𝒖𝟐 + 𝒄
𝟐 𝟏 𝟐
𝟐
𝟏
𝒑(𝒙) = −𝟏𝟑𝟓 𝒖𝟐 + 𝒄
Restoring the solution in terms of 𝒙 instead of 𝒖.
𝟏
𝒑(𝒙) = −𝟏𝟑𝟓 (𝟗 + 𝒙𝟐 )𝟐 + 𝒄
𝒑(𝒙) = −𝟏𝟑𝟓 √𝟗 + 𝒙𝟐 + 𝒄 is a general solution as 𝒄 is still arbitrary in nature.

𝟏
As given, that (𝒑 = 𝟑𝟎)𝒙=𝟒 , therefore, plugging values in 𝒑(𝒙) = −𝟏𝟑𝟓 (𝟗 + 𝒙𝟐 )𝟐 + 𝒄, we get:
𝟏
𝒑(𝟒) = 𝟑𝟎 = −𝟏𝟑𝟓 {𝟗 + (𝟒)𝟐 }𝟐 + 𝒄
𝟏 𝟏
𝟑𝟎 = −𝟏𝟑𝟓 {𝟗 + (𝟒)𝟐 }𝟐 + 𝒄 ⇒ 𝟑𝟎 = −𝟏𝟑𝟓 {𝟐𝟓}𝟐 + 𝒄 = −𝟏𝟑𝟓 × 𝟓 + 𝒄
𝟑𝟎 = −𝟏𝟑𝟓 × 𝟓 + 𝒄 ⇒ 𝒄 = 𝟕𝟎𝟓
Substituting the value of 𝒄 in general solution we get the definite solution:
𝒑(𝒙) = −𝟏𝟑𝟓 √𝟗 + 𝒙𝟐 + 𝟕𝟎𝟓
Graph of 𝒑(𝒙) = −𝟏𝟑𝟓 √𝟗 + 𝒙𝟐 + 𝟕𝟎𝟓
______________________________________________________________________________
c. How many units are demanded at a price of PKR 20 per unit?

Given that 𝒑 = 𝟐𝟎, the definite solution of Inverse demand function becomes:
𝟐
𝟐𝟎 = −𝟏𝟑𝟓 √𝟗 + (𝒙) + 𝟕𝟎𝟓 ⇒ 𝒙 = 𝟒. 𝟎𝟗 demand in (hundred) units
______________________________________________________________________________
TOPIC 013: SUPPLY ANALYSIS USING INTEGRATION BY SUBSTITUTION
The owner of a fast-food chain determines that if 𝒙 thousand units of a new meal item are supplied,
𝒙
then the marginal price at that level of supply is given by 𝒑′ (𝒙) = (𝒙+𝟑)𝟐. Where 𝒑(𝒙) is the price
(in dollars) per unit at which all 𝒙 meal units will be sold. Currently, 𝟓, 𝟎𝟎𝟎 units are being supplied
at a price of 𝑷𝑲𝑹 𝟐. 𝟐𝟎 per unit.
Find the supply (price) function 𝒑(𝒙).
𝒙
𝒑′ (𝒙) =
(𝒙 + 𝟑)𝟐
Find the supply function 𝒑(𝒙).
Solution: Given the rate of change of (inverse) supply function;
′ (𝒙) 𝒅𝒑 𝒙
𝒑 = 𝒅𝒙 = (𝒙+𝟑)𝟐, integration can give (inverse) supply function:
𝒅𝒑 𝒙 𝒙
∫ 𝒅𝒙 = ∫ 𝒅𝒙 ⇒ 𝒑 = ∫ 𝒅𝒙
𝒅𝒙 (𝒙 + 𝟑)𝟐 (𝒙 + 𝟑)𝟐
𝒙
𝒑(𝒙) = ∫ (𝒙+𝟑)𝟐 𝒅𝒙 using integration by substitution as basic rules of integration don’t work here.
Let, the denominator be substituted:
𝒖=𝒙+𝟑
𝒅𝒖
Which implies that: = 𝟏 ⇒ 𝒅𝒖 = 𝒅𝒙
𝒅𝒙
Substituting in the integrand, we get:
𝒙 𝒖−𝟑
𝒑(𝒙) = ∫ 𝒅𝒙 = ∫ (𝒅𝒖) ∵ [𝒖 = 𝒙 + 𝟑 ⇒ 𝒙 = 𝒖 − 𝟑]
(𝒙 + 𝟑)𝟐 (𝒖)𝟐
𝒖−𝟑 𝟏 𝟑 𝟏 𝟏
𝒑(𝒙) = ∫ ( 𝟐 ) 𝒅𝒖 = ∫ ( − 𝟐 ) 𝒅𝒖 = ∫ 𝒅𝒖 − 𝟑 ∫ 𝟐 𝒅𝒖
𝒖 𝒖 𝒖 𝒖 𝒖
(−𝟏) 𝟑
𝒑(𝒙) = 𝐥𝐧|𝒖| + 𝒄𝟏 − {𝟑 × + 𝒄𝟐 } ⇒ 𝐥𝐧|𝒖| + 𝒄𝟏 + − 𝒄𝟐
𝒖 𝒖
𝟑 𝟑
𝒑(𝒙) = 𝐥𝐧|𝒖| + + 𝒄𝟏 − 𝒄𝟐 = 𝐥𝐧|𝒖| + + 𝒄 ∵ [𝒄𝟏 − 𝒄𝟐 = 𝒄]
𝒖 𝒖
Restoring the solution in terms of 𝒙 in place of 𝒖: 𝒖 = 𝒙 + 𝟑
𝟑
𝒑(𝒙) = 𝐥𝐧|𝒖| + + 𝒄
𝒖
𝟑
𝒑(𝒙) = 𝐥𝐧|𝒙 + 𝟑| + +𝒄
𝒙+𝟑
Due to the arbitrary constant ′𝒄′, supply function 𝒑(𝒙) remains general solution.
Assuming value of 𝒄(= 𝟏) in general solution, we get definite solution:
𝟑
𝒑(𝒙) = 𝐥𝐧|𝒙 + 𝟑| + +𝟏
𝒙+𝟑
𝟑
𝒑(𝒙) = 𝐥𝐧|𝒙 + 𝟑| + +𝟏
𝒙+𝟑
______________________________________________________________________________
𝟑
Graph of 𝒑(𝒙) = 𝐥𝐧|𝒙 + 𝟑| + 𝒙+𝟑 + 𝟏
b. If 10,000 meal units are supplied to restaurants in the chain, what unit price should be charged
so that all the units will be sold?
𝟑
𝒑(𝟏𝟎𝟎𝟎𝟎) = 𝐥𝐧|𝟏𝟎𝟎𝟎𝟎 + 𝟑| + +𝟏
𝟏𝟎𝟎𝟎𝟎 + 𝟑
𝟑
𝒑(𝟏𝟎𝟎𝟎𝟎) = 𝐥𝐧|𝟏𝟎𝟎𝟎𝟑| + +𝟏
𝟏𝟎𝟎𝟎𝟑
𝒑(𝟏𝟎𝟎𝟎𝟎) = 𝟏𝟎. 𝟐𝟏
Interpretation: If 𝟏𝟎𝟎𝟎𝟎 meal units are supplied to restaurants in the chain, then unit price will
be PKR 10.21.
TOPIC 014: REVENUE ANALYSIS USING INTEGRATION BY SUBSTITUTION
The marginal revenue from the sale of x units of a particular commodity is estimated to be:
𝟐
𝑹′ (𝒙) = 𝟓𝟎 + 𝟑. 𝟓𝒙𝒆−𝟎.𝟎𝟏𝒙
Dollars per unit where 𝑹(𝒙) is revenue in dollars.
a. Find 𝑹(𝒙), assuming that 𝑹(𝟎) = 𝟎.
b. What revenue should be expected from the sale of 1,000 units?
Given the rate of change of Revenue function;

′ (𝒙) 𝒅𝑹 −𝟎.𝟎𝟏𝒙𝟐
𝑹 = 𝒅𝒙 = 𝟓𝟎 + 𝟑. 𝟓𝒙𝒆 , integration can give Revenue function:
𝒅𝑹 𝟐 𝟐
∫ 𝒅𝒙 = ∫ (𝟓𝟎 + 𝟑. 𝟓𝒙𝒆−𝟎.𝟎𝟏𝒙 ) 𝒅𝒙 ⇒ 𝑹 = ∫ 𝟓𝟎 𝒅𝒙 + ∫ 𝟑. 𝟓𝒙𝒆−𝟎.𝟎𝟏𝒙 𝒅𝒙
𝒅𝒙
𝟐
𝑹 = 𝑹(𝒙) = 𝟓𝟎 ∫ 𝟏 𝒅𝒙 + 𝟑. 𝟓 ∫ 𝒙𝒆−𝟎.𝟎𝟏𝒙 𝒅𝒙
𝟐
𝑹 = 𝑹(𝒙) = 𝟓𝟎𝒙 + 𝒄𝟏 + 𝟑. 𝟓 ∫ 𝒙𝒆−𝟎.𝟎𝟏𝒙 𝒅𝒙
______________________________________________________________________________
Applying integration by parts to solve underlined integrand:

𝒅𝒖 𝒅𝒖 𝒅𝒖
𝒖 = −𝟎. 𝟎𝟏𝒙𝟐 ⇒ = −𝟎. 𝟎𝟏(𝟐𝒙) ⇒ = −𝟎. 𝟎𝟐𝒙 ⇒ = 𝒙 𝒅𝒙
𝒅𝒙 𝒅𝒙 −𝟎. 𝟎𝟐
Rearranging the integrand to get 𝒙 𝒅𝒙 term in:
𝟐
𝑹(𝒙) = 𝟓𝟎𝒙 + 𝒄𝟏 + 𝟑. 𝟓 ∫ 𝒙𝒆−𝟎.𝟎𝟏𝒙 𝒅𝒙
𝟐
𝑹(𝒙) = 𝟓𝟎𝒙 + 𝒄𝟏 + 𝟑. 𝟓 ∫ 𝒆−𝟎.𝟎𝟏𝒙 𝒙𝒅𝒙
Introducing 𝒖 instead of 𝒙.
𝒅𝒖 𝟑. 𝟓
= 𝟓𝟎𝒙 + 𝒄𝟏 + 𝟑. 𝟓 ∫ 𝒆𝒖 ⇒ 𝟓𝟎𝒙 + 𝒄𝟏 + ∫ 𝒆𝒖 𝒅𝒖
−𝟎. 𝟎𝟐 −𝟎. 𝟎𝟐
= 𝟓𝟎𝒙 + 𝒄𝟏 + (−𝟏𝟕𝟓) ∫ 𝒆𝒖 𝒅𝒖
𝒆𝒇(𝒙)
Using Natural Exponential Function Rule: ∫ 𝒆𝒇(𝒙) 𝒅𝒙 = +𝒄
𝒇′ (𝒙)
= 𝟓𝟎𝒙 + 𝒄𝟏 + (−𝟏𝟕𝟓)𝒆𝒖 + 𝒄𝟐 ⇒ 𝟓𝟎𝒙 − 𝟏𝟕𝟓𝒆𝒖 + 𝒄𝟏 + 𝒄𝟐
= 𝟓𝟎𝒙 − 𝟏𝟕𝟓𝒆𝒖 + 𝒄
Restoring the solution in terms of 𝒙 in place of 𝒖: 𝒖 = −𝟎. 𝟎𝟏𝒙𝟐
𝟐
𝑹(𝒙) = 𝟓𝟎𝒙 − 𝟏𝟕𝟓𝒆−𝟎.𝟎𝟏𝒙 + 𝒄
Due to the arbitrary constant ′𝒄′, revenue function 𝑹(𝒙) remains general solution.
Given that 𝒙 = 𝟎, 𝑹(𝟎) = 𝟎, the general solution becomes:
𝟐
𝑹(𝟎) = 𝟓𝟎(𝟎) − 𝟏𝟕𝟓𝒆−𝟎.𝟎𝟏(𝟎) + 𝒄 ⇒ 𝑹(𝟎) = 𝟎 − 𝟏𝟕𝟓𝒆𝟎 + 𝒄
𝑹(𝟎) = 𝟎 − 𝟏𝟕𝟓 + 𝒄 ⇒ 𝟎 = −𝟏𝟕𝟓 + 𝒄 ⇒ 𝒄 = 𝟏𝟕𝟓
Substituting value of 𝒄 in general solution, we get definite solution:
𝟐
𝑹(𝒙) = 𝟓𝟎𝒙 − 𝟏𝟕𝟓𝒆−𝟎.𝟎𝟏𝒙 + 𝟏𝟕𝟓
𝟐
Graph of 𝑹(𝒙) = 𝟓𝟎𝒙 − 𝟏𝟕𝟓𝒆−𝟎.𝟎𝟏𝒙 + 𝟏𝟕𝟓
______________________________________________________________________________
If 𝟏, 𝟎𝟎𝟎 units are being produced then revenue shall be:

𝟐
𝑹(𝟏𝟎𝟎𝟎) = 𝟓𝟎(𝟏𝟎𝟎𝟎) − 𝟏𝟕𝟓𝒆−𝟎.𝟎𝟏(𝟏𝟎𝟎𝟎) + 𝟏𝟕𝟓 = 𝑷𝑲𝑹 𝟓𝟎𝟏𝟕𝟓
Interpretation: If 𝟏, 𝟎𝟎𝟎 units are sold, then the revenue will be PKR 50175.
TOPIC 015: INVESTMENT PORTFOLIO ANALYSIS USING INTEGRATION BY

SUBSTITUTION
An investment portfolio changes value at the rate:

𝑽′ (𝒕) = 𝟏𝟐𝒆−𝟎.𝟎𝟓𝒕 (𝒆𝟎.𝟑𝒕 − 𝟑)
dollars per unit where 𝑽(𝒕) is investment in dollars. Find 𝑽(𝒕).
Given the rate of change of investment portfolio function;
𝒅𝑽
𝑽′ (𝒕) = = 𝟏𝟐𝒆−𝟎.𝟎𝟓𝒕 (𝒆𝟎.𝟑𝒕 − 𝟑), integration can give investment portfolio function:
𝒅𝒕
𝒅𝑽
∫ 𝒅𝒕 = ∫{𝟏𝟐𝒆−𝟎.𝟎𝟓𝒕 (𝒆𝟎.𝟑𝒕 − 𝟑)} 𝒅𝒕 ⇒ 𝑽(𝒕) = 𝟏𝟐 ∫{𝒆−𝟎.𝟎𝟓𝒕 (𝒆𝟎.𝟑𝒕 − 𝟑)} 𝒅𝒕
𝒅𝒕
Converting decimals into ratios for ease.
𝟏 𝟑
− 𝒕 𝒕
𝑽(𝒕) = 𝟏𝟐 ∫ {𝒆 𝟐𝟎 (𝒆𝟏𝟎 − 𝟑)} 𝒅𝒕
𝟏 𝒅𝒖 𝟏
Applying substitution method: 𝒖 = − 𝟐𝟎 𝒕 ⇒ 𝒅𝒕
= − 𝟐𝟎 ⇒ −𝟐𝟎𝒅𝒖 = 𝒅𝒕
𝟏 𝟑
− 𝒕
𝑽(𝒕) = 𝟏𝟐 ∫ {𝒆 𝟐𝟎 (𝒆 𝟏𝟎𝒕 − 𝟑)} 𝒅𝒕
𝟏 𝟑 𝟑 𝟑
− 𝒕 = 𝒖 ⇒ 𝒕 = −𝟐𝟎𝒖 ⇒ 𝒕= × (−𝟐𝟎𝒖) ⇒ 𝒕 = −𝟔𝒖
𝟐𝟎 𝟏𝟎 𝟏𝟎 𝟏𝟎
𝑽(𝒖) = 𝟏𝟐 ∫ {𝒆𝒖 (𝒆 −𝟔𝒖 − 𝟑)} (−𝟐𝟎𝒅𝒖)
𝑽(𝒖) = 𝟏𝟐 × (−𝟐𝟎) ∫{𝒆𝒖 (𝒆−𝟔𝒖 − 𝟑)} 𝒅𝒖
𝑽(𝒖) = −𝟐𝟒𝟎 ∫{𝒆𝒖 (𝒆−𝟔𝒖 − 𝟑)} 𝒅𝒖
𝑽(𝒖) = −𝟐𝟒𝟎 ∫(𝒆−𝟓𝒖 − 𝟑𝒆𝒖 ) 𝒅𝒖 ⇒ 𝟐𝟒𝟎 ∫(𝟑𝒆𝒖 − 𝒆−𝟓𝒖 ) 𝒅𝒖
𝑽(𝒖) = 𝟐𝟒𝟎 {∫(𝟑𝒆𝒖 ) 𝒅𝒖 − ∫(𝒆−𝟓𝒖) 𝒅𝒖}
𝑽(𝒖) = 𝟐𝟒𝟎 {𝟑 ∫(𝒆𝒖 ) 𝒅𝒖 − ∫(𝒆−𝟓𝒖 ) 𝒅𝒖}

𝒆𝒇(𝒙)
Using Natural Exponential Function Rule ∫ 𝒆𝒇(𝒙) 𝒅𝒙 = 𝒇′ (𝒙) + 𝒄
𝒆−𝟓𝒖
𝑽(𝒖) = 𝟐𝟒𝟎 {𝟑 𝒆𝒖 + 𝒄𝟏 − ( + 𝒄𝟐 )}
(−𝟓)
𝒆−𝟓𝒖
𝑽(𝒖) = 𝟐𝟒𝟎 {𝟑 𝒆𝒖 + + 𝒄𝟏 − 𝒄𝟐 }
𝟓
𝒆−𝟓𝒖
𝑽(𝒖) = 𝟐𝟒𝟎 {𝟑 𝒆𝒖 + + 𝒄𝟑 }
𝟓
𝑽(𝒖) = 𝟕𝟐𝟎 𝒆𝒖 + 𝟒𝟖 𝒆−𝟓𝒖 + 𝟐𝟒𝟎 𝒄𝟑
𝑽(𝒖) = 𝟕𝟐𝟎 𝒆𝒖 + 𝟒𝟖 𝒆−𝟓𝒖 + 𝒄
______________________________________________________________________________
𝟏
Restoring the function in terms of 𝒕, instead of 𝒖, as 𝒖 = − 𝟐𝟎 𝒕
𝟏 𝟏
(− 𝒕) −𝟓(− 𝒕)
𝑽(𝒕) = 𝟕𝟐𝟎 𝒆 𝟐𝟎 + 𝟒𝟖 𝒆 𝟐𝟎 +𝒄
𝒕 𝒕
𝑽(𝒕) = 𝟕𝟐𝟎 𝒆−𝟐𝟎+ 𝟒𝟖 𝒆𝟒
+𝒄
Due to the arbitrary constant ′𝒄′, investment portfolio function 𝑽(𝒕) remains general solution.
𝒕 𝒕
Graph of 𝑽(𝒕) = 𝟕𝟐𝟎 𝒆−𝟐𝟎 + 𝟒𝟖 𝒆𝟒 + ⏟
𝒄
( 𝑙𝑒𝑡 )
𝒄=𝟎
______________________________________________________________________________
Lesson 05
DEFINITE INTEGRALS
TOPIC 016: DEFINITE INTEGRALS
By the virtue of indefinite integral, we know:

∫ 𝒇(𝒙) 𝒅𝒙 = 𝑭(𝒙) + 𝒄
However, if starting & ending values from domain (here, values of 𝑥) are chosen then function
has starting & ending points.
Diagrammatically
Here the between the curve and 𝒙 −axis has opposite signs (+𝒗𝒆 & − 𝒗𝒆), therefore, it’s a signed
area. Properly defining, definite integral is the signed area between curve, 𝒙 −axis and lines (𝒙 =
𝒂 and 𝒙 = 𝒃).
Using calculus, the expression and diagram looks like this:
𝒃
𝑨𝒓𝒆𝒂 = ∫ 𝒇(𝒙) 𝒅𝒙
𝒂
Where ′𝒂′ is the lower limit and ′𝒃′ is the upper limit, i.e. (𝒂 < 𝒃).
Solution will be
𝒃
∫ 𝒇(𝒙) 𝒅𝒙 = {𝑭(𝒙) + 𝒄}]𝒃𝒂 = [𝑭(𝒃) + 𝒄] − [𝑭(𝒂) + 𝒄]
𝒂
= 𝑭(𝒃) + 𝒄 − 𝑭(𝒂) − 𝒄 ⇒ 𝑭(𝒃) − 𝑭(𝒂)
______________________________________________________________________________
𝒄 – the arbitrary constant of integration, is cancelled out in definite integral, therefore, usually it is
omitted.
Interpretation: Result of a definite integral will be numerical in nature.
Since, diagrammatically it is ‘Area’, therefore, it has a ‘definite’ numerical value.
Numerical Example # 1:
𝟓
∫ 𝟑𝒙𝟐 𝒅𝒙 = (𝒙𝟑 + 𝒄)]𝟓𝟏 = [(𝟓)𝟑 + 𝒄] − [(𝟏)𝟑 + 𝒄]
𝟏
= 𝟏𝟐𝟓 + 𝒄 − 𝟏 − 𝒄 ⇒ 𝟏𝟐𝟒
𝒃 𝒙 𝒙 𝒃
∫𝒂 𝒌𝒆 𝒅𝒙 = 𝒌(𝒆 )]𝒂 [constant of definite integrals gets omitted]
= 𝒌(𝒆𝒃 − 𝒆𝒂 )
𝟒
𝟒 𝟏
∫𝟎 (𝟏+𝒙 + 𝟐𝒙) 𝒅𝒙 = ( ⏟
𝐥𝐧|𝟏 + 𝒙| ⏟𝟐 )]
+ 𝒙
𝑫𝒆𝒓𝒊𝒗𝒂𝒕𝒊𝒗𝒆 𝒐𝒗𝒆𝒓 𝑷𝒐𝒘𝒆𝒓
𝑭𝒖𝒏𝒄𝒕𝒊𝒐𝒏 𝑹𝒖𝒍𝒆 𝑹𝒖𝒍𝒆 𝟎
𝟐 𝟐
= (𝐥𝐧|𝟏 + 𝟒| − 𝟒 ) + (𝐥𝐧|𝟏 + 𝟎| − 𝟎 )
= (𝐥𝐧|𝟓| − 𝟏𝟔) + (𝐥𝐧|𝟏| − 𝟎) ⇒ (𝐥𝐧|𝟓| − 𝟏𝟔) + (𝟎 −𝟎)
= 𝐥𝐧|𝟓| − 𝟏𝟔
TOPIC 017: DEFINITE INTEGRAL: AREA UNDER CURVE
In addition to calculus-based (precise) method, Area under the curve can be ‘approximated’
graphically by dividing into vertical rectangles.
Diagram shows that the area under the curve between [𝒂, 𝒃], is divided into seven (𝟕) vertical
rectangles. The width of each of these rectangles is equal, however, height is not. Lengths are
𝒇(𝒙𝒐 ), 𝒇(𝒙𝟏 ), 𝒇(𝒙𝟐 ), 𝒇(𝒙𝟑 ), 𝒇(𝒙𝟒 ), 𝒇(𝒙𝟓 ) and 𝒇(𝒙𝟔 ), respectively.
Therefore, the respective areas of each rectangle will be 𝒇(𝒙𝒐 ) × (𝒙𝟏 − 𝒙𝒐 ), 𝒇(𝒙𝟏 ) × (𝒙𝟐 − 𝒙𝟏 ),
𝒇(𝒙𝟐 ) × (𝒙𝟑 − 𝒙𝟐 ), 𝒇(𝒙𝟑 ) × (𝒙𝟒 − 𝒙𝟑 ), 𝒇(𝒙𝟒 ) × (𝒙𝟓 − 𝒙𝟒 ), 𝒇(𝒙𝟓 ) × (𝒙𝟔 − 𝒙𝟓 ), and 𝒇(𝒙𝟔 ) × (𝒙𝟕 −
𝒙𝟔 ).
Adding the area (𝑨) of the seven rectangles, we get overall ‘shaded’ area:
______________________________________________________________________________
𝑨 = 𝒇(𝒙𝒐 ) × ( 𝒙𝟏 − 𝒙𝒐 ) + 𝒇(𝒙𝟏 ) × (𝒙𝟐 − 𝒙𝟏 ) + 𝒇(𝒙𝟐 ) × (𝒙𝟑 − 𝒙𝟐 ) + 𝒇(𝒙𝟑 ) × (𝒙𝟒 − 𝒙𝟑 ) + 𝒇(𝒙𝟒 )

× (𝒙𝟓 − 𝒙𝟒 ) + 𝒇(𝒙𝟓 ) × (𝒙𝟔 − 𝒙𝟓 ) + 𝒇(𝒙𝟔 ) × (𝒙𝟕 − 𝒙𝟔 )
𝑨 = 𝒇(𝒙𝒐 ) × ( ∆𝒙𝒐 ) + 𝒇(𝒙𝟏 ) × (∆𝒙𝟏 ) + 𝒇(𝒙𝟐 ) × (∆𝒙𝟐 ) + 𝒇(𝒙𝟑 ) × (∆𝒙𝟒 ) + 𝒇(𝒙𝟒 ) × (∆𝒙𝟓 ) + 𝒇(𝒙𝟓 )
× (∆𝒙𝟔 ) + 𝒇(𝒙𝟔 ) × (∆𝒙𝟕 )
In summation form:
𝒏
𝑨 = ∑{𝒇(𝒙𝒏 ) × (∆𝒙𝒏 )}
𝒊=𝒐
The greater the value of 𝒏, the smaller will be the rectangles and more precise will be calculation.
This method is usually known as the ‘Riemann Sum’.
Numerical depiction:
𝑨 = 𝒇(𝒙𝟏 ) × (𝒙𝟐 − 𝒙𝟏 ) + 𝒇(𝒙𝟐 ) × (𝒙𝟑 − 𝒙𝟐 ) + 𝒇(𝒙𝟑 ) × (𝒙𝟒 − 𝒙𝟑 ) + 𝒇(𝒙𝟒 ) × (𝒙𝟓 − 𝒙𝟒 ) + 𝒇(𝒙𝟓 )

× (𝒙𝟔 − 𝒙𝟓 )
𝟏 𝟏 𝟏 𝟏
𝑨 = 𝟏 × (𝟐 − 𝟏) + 𝟐 × (𝟑 − 𝟐) + 𝟑 × (𝟒 − 𝟑) + 𝟒 × (𝟓 − 𝟒) + 𝟓 × (𝟔 − 𝟓)
𝟏 𝟏 𝟏 𝟏
𝑨 =𝟏+ + + +
𝟐 𝟑 𝟒 𝟓
𝑨 = 𝟐. 𝟐𝟖𝟑
Interpretation: Approximation of the area under the curve between 𝟏 and 𝟔.
𝟔 𝟏
However, if calculus-based method is adopted: ∫𝟏 ( ) 𝒅𝒙 = 𝟏. 𝟕𝟗𝟐
𝒙
It is observable that Area using Riemann Sum (𝟐. 𝟐𝟖𝟑) is over-approximated as compared to that
using calculus-based method (𝟏. 𝟕𝟗𝟐).
TOPIC 018: PROPERTIES OF DEFINITE INTEGRALS
Property – 1: Interchanging limits of integration changes the sign of the definite integral:
𝒃 𝒂
∫ 𝒇(𝒙) 𝒅𝒙 = − ∫ 𝒇(𝒙) 𝒅𝒙
𝒂 𝒃
𝒃
Proof: Using L.H.S: ∫𝒂 𝒇(𝒙) 𝒅𝒙 = 𝑭(𝒃) − 𝑭(𝒂)
Taking −𝟏 common, we get: = −{−𝑭(𝒃) + 𝑭(𝒂)}
Interchanging places of the two terms: = −{𝑭(𝒂) − 𝑭(𝒃)}
𝒖𝒑𝒑𝒆𝒓 𝒍𝒊𝒎𝒊𝒕
In general form: ∫𝒍𝒐𝒘𝒆𝒓 𝒍𝒊𝒎𝒊𝒕 𝒇(𝒙) 𝒅𝒙 = {𝑭(𝒖𝒑𝒑𝒆𝒓 𝒍𝒊𝒎𝒊𝒕) − 𝑭(𝒍𝒐𝒘𝒆𝒓 𝒍𝒊𝒎𝒊𝒕)}
𝒂
Using the general form write the integral form: − ∫𝒃 𝒇(𝒙) 𝒅𝒙 = R.H.S
Property – 2: Definite integral has a value of zero when the two limits of integration are
identical:
______________________________________________________________________________
𝒂
∫ 𝒇(𝒙) 𝒅𝒙 = 𝟎
𝒂
𝒂
Proof: Using L.H.S: ∫𝒂 𝒇(𝒙) 𝒅𝒙 = 𝑭(𝒂) − 𝑭(𝒂)
𝟎 = R.H.S
Property – 3: Definite integral can be expressed as a sum of a finite number of definite

sub-integrals:
𝒅 𝒃 𝒄 𝒅
∫ 𝒇(𝒙) 𝒅𝒙 = ∫ 𝒇(𝒙) 𝒅𝒙 + ∫ 𝒇(𝒙) 𝒅𝒙 + ∫ 𝒇(𝒙) 𝒅𝒙
𝒂 𝒂 𝒃 𝒄
Where (𝒂 < 𝒃 < 𝒄 < 𝒅)
It can be termed as Additive property of definite integrals.
Graphically speaking
Property – 4: A coefficient in definite integral appears a coefficient outside the integral:

𝒃 𝒃
∫ 𝒌 𝒇(𝒙) 𝒅𝒙 = 𝒌 ∫ 𝒇(𝒙) 𝒅𝒙
𝒂 𝒂
For example:
𝟓
∫ 𝟑𝒄𝒙𝟐 𝒅𝒙 = 𝒄 {(𝒙𝟑 + 𝒄)]𝟓𝟏 } = 𝒄 [{(𝟓)𝟑 + 𝒄} − {(𝟏)𝟑 + 𝒄}]
𝟏
= 𝒄 (𝟏𝟐𝟓 + 𝒄 − 𝟏 − 𝒄) ⇒ 𝒄 (𝟏𝟐𝟒)
Property – 5: Definite integrals follow sum & difference rules of integration as followed by
indefinite integrals:
𝒃 𝒃 𝒃
∫{𝒈(𝒙) ± 𝒉(𝒙)} 𝒅𝒙 = ∫{𝒈(𝒙) 𝒅𝒙} ± ∫{𝒉(𝒙)} 𝒅𝒙 + 𝒄

𝒂 𝒂 𝒂
Numerical Example(s):
𝟐 𝟐
𝟐
𝒙𝟑 (𝟐)𝟑 (−𝟏)𝟑 𝟒𝟔 𝟐𝟔 𝟕𝟐
∫(𝟗 − 𝒙 ) 𝒅𝒙 = 𝟗𝒙 − ] ⇒ {𝟗(𝟐) − } − {𝟗(−𝟏) − } ⇒ ( ) − (− ) ⇒ ⇒ 𝟐𝟒
𝟑 −𝟏 𝟑 𝟑 𝟑 𝟑 𝟑
−𝟏
______________________________________________________________________________
TOPIC 019: IMPROPER INTEGRALS
Improper implies ‘unfitting’ – something not in accordance with accepted standards

Here, mathematically speaking, the element of ‘undefined’ (∞) will be present.
Can have two reasons for undefined answer:
1. Infinite limits of integration
2. Infinite integrand
1. INFINITE LIMITS OF INTEGRATION:

Consider the following two definite integrals:
∞
∫ 𝒇(𝒙) 𝒅𝒙
𝒂
Or
𝒃
∫ 𝒇(𝒙) 𝒅𝒙
−∞
Either of upper or lower limits are ∞ in these cases. Therefore, an improper integral is there.
∞
𝒅𝒙 𝟏∞ 𝟏 𝟏
∫ 𝟐
= − ] ⇒ − {( ) − ( )} ⇒ −{(𝟎) − (𝟏)} ⇒ 𝟏
𝒙 𝒙𝟏 ∞ 𝟏
𝟏
Or
∞
𝒅𝒙
∫ = 𝐥𝐧|𝒙|]∞
𝟏 ⇒ 𝐥𝐧|∞| − 𝐥𝐧|1| ⇒ ∞ − 𝟎 ⇒ ∞
𝒙
𝟏
______________________________________________________________________________
2. INFINITE INTEGRAND
Even if, none of limits [𝒂, 𝒃] are infinite, integrand can become undefined – improper integral.
1
𝒅𝒙
∫ = 𝐥𝐧|𝒙|]1𝟎 ⇒ {𝐥𝐧|𝟏| − 𝐥𝐧|𝟎|} ⇒ {(𝟎) − (∞)} ⇒ −∞
𝒙
𝟎
9
𝒅𝒙 𝟏 9 𝟏 𝟏
∫ = 𝟐 𝒙𝟐 ] ⇒ 𝟐 {(𝟗)𝟐 − (𝟎)𝟐 } ⇒ 𝟔
√𝒙 𝟎
𝟎
Economic applications of improper integrals include a situation with interest from a perpetual bond
or the revenue from an indestructible capital asset such as land.
______________________________________________________________________________
Lesson 06
MARGINAL TO TOTAL ANALYSIS USING INTEGRALS
TOPIC 020: MARGINAL TO TOTAL COST USING INTEGRALS
Given that, 𝑪′ (𝑸) = 𝟐 𝒆𝟎.𝟐𝑸 , 𝑪𝑭 = 𝟗𝟎, 𝑪(𝑸) =?

∫ 𝑪′ (𝑸) 𝒅𝑸 = ∫ 𝟐 𝒆𝟎.𝟐𝑸 𝒅𝑸 = 𝟐 ∫ 𝒆𝟎.𝟐𝑸 𝒅𝑸
𝒆𝒇(𝒙)
∫ 𝒆𝒇(𝒙) 𝒅𝒙 = 𝒇′ (𝒙)
𝒆𝟎.𝟐𝑸
∫ 𝑪′ (𝑸) 𝒅𝑸 = 𝟐 ∫ 𝒆𝟎.𝟐𝑸 𝒅𝑸 ⇒ 𝟐 ( 𝟎.𝟐
)+ 𝒄 ⇒ 𝟏𝟎(𝒆𝟎.𝟐𝑸 ) + 𝒄
Since derivative & integral are inverse operators: 𝑪(𝑸) = 𝟏𝟎 𝒆𝟎.𝟐𝑸 + 𝒄
Indeterminate answer as (𝒄) is still arbitrary.
Use 𝑪𝑭 = 𝟗𝟎 ⇒ 𝑪(𝑸) = 𝟏𝟎 𝒆𝟎.𝟐𝑸 + 𝒄
Initial Condition requires: 𝑸 = 𝟎
𝑪(𝟎) = 𝟏𝟎 𝒆𝟎.𝟐(𝟎) + 𝒄 ⇒ 𝑪(𝟎) = 𝟏𝟎 𝒆𝟎 + 𝒄 ⇒ 𝑪(𝟎) = 𝟏𝟎 + 𝒄
𝑪(𝟎) = 𝟏𝟎 + 𝒄 as 𝑪(𝑸)𝑸=𝟎 = 𝑪(𝟎) = 𝑪𝑭
⇒ 𝑪𝑭 = 𝟏𝟎 + 𝒄 ⇒ 𝒄 = 𝑪𝑭 − 𝟏𝟎 ⇒ 𝒄 = 𝟗𝟎 − 𝟏𝟎 ⇒ 𝒄 = 𝟖𝟎
Substituting value of 𝒄: 𝑪(𝑸) = 𝟏𝟎 𝒆𝟎.𝟐𝑸 + 𝟖𝟎
Graph of 𝑪(𝑸) = 𝟏𝟎 𝒆𝟎.𝟐𝑸 + 𝟖𝟎
TOPIC 021: MARGINAL TO TOTAL REVENUE USING INTEGRALS
Given that, 𝑹′ (𝑸) = 𝟐𝟖 𝑸 − 𝒆𝟎.𝟑𝑸 , 𝑹(𝑸) =?

∫ 𝑹′ (𝑸) 𝒅𝑸 = ∫(𝟐𝟖 𝑸 − 𝒆𝟎.𝟑𝑸 ) 𝒅𝑸 = 𝟐𝟖 ∫ 𝑸 𝒅𝑸 − ∫ 𝒆𝟎.𝟑𝑸 𝒅𝑸
𝟏 𝟏
= 𝟏𝟒𝑸𝟐 + 𝒄𝟏 − (𝟎.𝟑 × 𝟎. 𝟑) ∫ 𝒆𝟎.𝟑𝑸 𝒅𝑸 = 𝟏𝟒𝑸𝟐 − 𝟎.𝟑 ∫ 𝟎. 𝟑 𝒆𝟎.𝟑𝑸 𝒅𝑸
______________________________________________________________________________
Since: ∫ 𝒇′ (𝒙) 𝒆𝒇(𝒙) 𝒅𝒙 = 𝒆𝒇(𝒙) + 𝒄

𝟏 𝟎.𝟑𝑸 𝟏 𝟎.𝟑𝑸
= 𝟏𝟒𝑸𝟐 + 𝒄𝟏 − ( 𝒆 + 𝒄𝟐 ) = 𝟏𝟒𝑸𝟐 − 𝒆 + 𝒄 [∵ 𝒄𝟏 − 𝒄𝟐 = 𝒄]
𝟎.𝟑 𝟎.𝟑
𝟏𝟎
= 𝟏𝟒𝑸𝟐 − 𝟑 𝒆𝟎.𝟑𝑸 + 𝒄
Since derivative & integral are inverse operators: ∫ 𝑹′ (𝑸) 𝒅𝑸 = 𝑹(𝑸)
𝟏𝟎 𝟎.𝟑𝑸
Finally, we have 𝑹(𝑸) = 𝟏𝟒𝑸𝟐 − 𝟑
𝒆 + 𝒄 General Solution
Indeterminate answer as (𝒄) is still arbitrary.
𝟏𝟎 𝟏𝟎
Use initial condition: 𝑸 = 𝟎 ⇒ 𝑹(𝟎) = 𝟏𝟒(𝟎)𝟐 − 𝟑 𝒆𝟎.𝟑(𝟎) + 𝒄 ⇒ 𝑹(𝟎) = − 𝟑 + 𝒄 ⇒ 𝒄 = 𝑹(𝟎) +
𝟏𝟎
[∵ 𝑹(𝟎) = 𝟎]: no revenue at no output.
𝟑
𝟏𝟎
⇒𝒄=
𝟑
𝟏𝟎 𝟏𝟎
Substituting value of 𝒄: 𝑹(𝑸) = 𝟏𝟒𝑸 − 𝟑 𝒆𝟎.𝟑𝑸 + 𝟑 Definite Solution
𝟐
𝟏𝟎 𝟎.𝟑𝑸 𝟏𝟎
Graph of 𝑹(𝑸) = 𝟏𝟒𝑸𝟐 − 𝒆 +
𝟑 𝟑
TOPIC 022: MARGINAL TO TOTAL SAVINGS USING INTEGRALS

𝟏
Given that, 𝑺′ (𝒀) = 𝟎. 𝟑 − 𝟎. 𝟏 𝒀−𝟐 , (𝑺)𝒀=𝟖𝟏 = 𝟎, 𝑺(𝒀) =?
Caveat: (𝑺)𝒀=𝟖𝟏 = 𝟎 not (𝒀)𝑺=𝟖𝟏 = 𝟎 as income must not be zero for a positive level of
savings {(𝒀 ≠ 𝟎)𝑺>𝟎 }.
Given condition is an initial condition of dependent variable in terms of independent variable.
𝟏 𝟏
− −
∫ 𝑺′ (𝒀) 𝒅𝒀 = ∫ (𝟎. 𝟑 − 𝟎. 𝟏𝒀 𝟐 ) 𝒅𝒀 ⇒ ∫ 𝟎. 𝟑 𝒅𝒀 − ∫ 𝟎. 𝟏𝒀 𝟐 𝒅𝒀
𝟏
𝟏
𝒀𝟐
𝑺(𝒀) = 𝟎. 𝟑𝒀 − 𝟎. 𝟏 ( 𝟏 ) + 𝒄 ⇒ 𝟎. 𝟑𝒀 − 𝟎. 𝟐 (𝒀𝟐 ) + 𝒄
𝟐
𝟏
𝑺(𝒀) = 𝟎. 𝟑𝒀 − 𝟎. 𝟐 (𝒀𝟐 ) + 𝒄 Indeterminate savings function
Given (𝑺)𝒀=𝟖𝟏 = 𝟎, therefore substituting 𝒀 = 𝟖𝟏:
𝟏
𝑺(𝟖𝟏) = 𝟎. 𝟑(𝟖𝟏) − 𝟎. 𝟐 (𝟖𝟏𝟐 ) + 𝒄 ⇒ 𝟎 = 𝟐𝟐. 𝟓 + 𝒄 ⇒ 𝒄 = −𝟐𝟐. 𝟓
Substituting value of 𝑐, we get determinate savings function:
______________________________________________________________________________
𝟏
𝑺(𝒀) = 𝟎. 𝟑𝒀 − 𝟎. 𝟐𝒀𝟐 − 𝟐𝟐. 𝟓
𝟏
Graph of 𝑺(𝒀) = 𝟎. 𝟑𝒀 − 𝟎. 𝟐𝒀𝟐 − 𝟐𝟐. 𝟓
TOPIC 023: SHORT RUN PRODUCTION FUNCTION FROM MARGINAL PRODUCTS USING
INTEGRALS
Short-run production functions can be found by using marginal product function.

𝒅
Let 𝑴𝑷𝑳 = 𝒅𝑳 {𝑸(𝑲, 𝑳)} = 𝟏𝟎𝟎𝟎 − 𝟑𝑳𝟐 .
Here 𝑲 shows the constancy of capital in the short run.
𝒅
Integrating w.r.t. 𝑳 on both sides: ∫ {𝑸(𝑲, 𝑳)} 𝒅𝑳 = ∫(𝟏𝟎𝟎𝟎 − 𝟑𝑳𝟐 ) 𝒅𝑳
𝒅𝑳
𝑸(𝑲, 𝑳) = ∫(𝟏𝟎𝟎𝟎) 𝒅𝑳 − ∫(𝟑𝑳𝟐 ) 𝒅𝑳 ⇒ 𝟏𝟎𝟎𝟎 ∫(𝑳𝟎 ) 𝒅𝑳 − 𝟑 ∫(𝑳𝟐 ) 𝒅𝑳

𝑳𝟑
= {𝟏𝟎𝟎𝟎 (𝑳) + 𝒄𝟏 } − {𝟑 ( ) + 𝒄𝟐 } ⇒ 𝟏𝟎𝟎𝟎𝑳 − 𝑳𝟑 + (𝒄𝟏 − 𝒄𝟐 )
𝟑
𝑸(𝑲, 𝑳) = 𝟏𝟎𝟎𝟎𝑳 − 𝑳𝟑 + 𝒄 ∵ 𝒄𝟏 − 𝒄𝟐
______________________________________________________________________________
Graph of 𝑸(𝑲, 𝑳) = 𝟏𝟎𝟎𝟎𝑳 − 𝑳𝟑 + ⏟

𝒄
( 𝒍𝒆𝒕 )
𝒄=𝟏
______________________________________________________________________________
Lesson 07
MARGINAL TO TOTAL ANALYSIS USING INTEGRALS (CONTINUED)
TOPIC 024: CONSUMPTION FUNCTION FROM MPC USING INTEGRALS

𝟏𝟎
Given that, 𝑴𝑷𝑪 = 𝟐𝟎 + 𝟑 (𝑪)𝒀=𝟏𝟔 = 𝟒𝟐𝟎, 𝑪(𝒀) =?
𝒀𝟒
Caveat: (𝑺)𝒀=𝟖𝟏 = 𝟎
⏟ NOT (𝒀)𝑺=𝟖𝟏 = 𝟎 .
⏟
𝑠𝑎𝑣𝑖𝑛𝑔 𝑐𝑎𝑛 ℎ𝑎𝑝𝑝𝑒𝑛 𝑠𝑎𝑣𝑖𝑛𝑔 𝑐𝑎𝑛′ 𝑡 ℎ𝑎𝑝𝑝𝑒𝑛
𝑤𝑖𝑡ℎ 𝑠𝑜𝑚𝑒 𝑖𝑛𝑐𝑜𝑚𝑒 𝑤𝑖𝑡ℎ 𝑛𝑜 𝑖𝑛𝑐𝑜𝑚𝑒
(𝒀≠𝟎⇒𝑺>𝟎) 𝒀=𝟎⇒𝑺≤𝟎
𝟑
𝟏𝟎
∫(𝑴𝑷𝑪) 𝒅𝒀 = ∫ 𝑪′ (𝒀) 𝒅𝒀 = ∫ (𝟐𝟎 + 𝟑 ) 𝒅𝒀 ⇒ ∫(𝟐𝟎)𝒅𝒀 + ∫ 𝟏𝟎𝒀−𝟒 𝒅𝒀
𝒀𝟒
𝟏
𝟏
𝒀𝟒
𝑪(𝒀) = 𝟐𝟎𝒀 + 𝟏𝟎 ( 𝟏 ) + 𝒄 ⇒ 𝟐𝟎𝒀 + 𝟒𝟎 (𝒀𝟒 ) + 𝒄
𝟒
𝟏
𝑪(𝒀) = 𝟐𝟎𝒀 + 𝟒𝟎 (𝒀𝟒 ) + 𝒄 Indeterminate consumption function
Given (𝑪)𝒀=𝟏𝟔 = 𝟒𝟐𝟎, therefore substituting:
𝟏
(𝑪)𝒀=𝟏𝟔 = 𝑪(𝟏𝟔) = 𝟐𝟎(𝟏𝟔) + 𝟒𝟎 (𝟏𝟎𝟒 ) + 𝒄
𝟒𝟐𝟎 = 𝟑𝟐𝟎 + 𝟖𝟎 + 𝒄 ⇒ 𝟒𝟐𝟎 = 𝟒𝟎𝟎 + 𝒄
𝒄 = 𝟒𝟐𝟎 − 𝟒𝟎𝟎 ⇒ 𝒄 = 𝟐𝟎
Substituting value of 𝑐, we get determinate consumption function
𝟏
𝑪(𝒀) = 𝟐𝟎𝒀 + 𝟒𝟎𝒀𝟒 + 𝟐𝟎
𝟏
𝑪(𝒀) = 𝟐𝟎 (𝒀 + 𝟐𝒀𝟒 + 𝟏)
𝟏
Graph of 𝑪(𝒀) = 𝟐𝟎𝒀 + 𝟒𝟎𝒀𝟒 + 𝟐𝟎
______________________________________________________________________________
TOPIC 025: MARGINAL TO TOTAL IMPORTS USING INTEGRALS
Given that, 𝑴′ (𝒀) = 𝟎. 𝟏, (𝑴)𝒀=𝟎 = 𝟐𝟎, 𝑴(𝒀) =?

∫ 𝑴′ (𝒀) 𝒅𝒀 = ∫ 𝟎. 𝟏 𝒅𝒀 ⇒ 𝟎. 𝟏 ∫(𝟏) 𝒅𝒀 ⇒ 𝟎. 𝟏 𝒀
∫ 𝑴′ (𝒀) 𝒅𝒀 = 𝟎. 𝟏 𝒀 + 𝒄
Since derivative & integral are inverse operators: 𝑴(𝒀) = 𝟎. 𝟏 𝒀 + 𝒄
Indeterminate or General Solution as (𝑐) is still arbitrary.
Initial Condition implies: 𝒀 = 𝟎
Use (𝑴)𝒀=𝟎 = 𝑴(𝟎) = 𝟐𝟎 ⇒ 𝑴(𝟎) = 𝟎. 𝟏 (𝟎) + 𝒄
𝟐𝟎 = 𝟎 + 𝒄 ⇒ 𝒄 = 𝟐𝟎
Substituting value of 𝒄 in:
𝑴(𝒀) = 𝟎. 𝟏 𝒀 + 𝒄 becomes
𝑴(𝒀) = 𝟎. 𝟏 𝒀 + 𝟐𝟎 Definite Solution for Imports Function.
______________________________________________________________________________
Graph of 𝑴(𝒀) = 𝟎. 𝟏 𝒀 + 𝟐𝟎
TOPIC 026: MARGINAL TO TOTAL PROFIT USING INTEGRALS
The marginal profit of a certain commodity is 𝝅(𝒒) = 𝟏𝟎𝟎 − 𝟐𝒒 when 𝒒 units are produced. When
10 units are produced, the profit is $700.
𝝅′ (𝒒) = 𝟏𝟎𝟎 − 𝟐𝒒
Moreover, {𝝅(𝒒)}𝒒=𝟏𝟎 = $𝟕𝟎𝟎
𝒅
𝝅′ (𝒒) = 𝒅𝑳 {𝝅(𝒒)} = 𝟏𝟎𝟎 − 𝟐𝒒.
𝒅
Integrating w.r.t. 𝒒 on both sides: ∫ 𝒅𝒒 {𝝅(𝒒)} 𝒅𝒒 = ∫(𝟏𝟎𝟎 − 𝟐𝒒) 𝒅𝒒
𝝅(𝒒) = ∫(𝟏𝟎𝟎) 𝒅𝒒 − ∫(𝟐𝒒) 𝒅𝒒 ⇒ 𝟏𝟎𝟎 ∫(𝟏) 𝒅𝒒 − 𝟐 ∫(𝒒) 𝒅𝒒

𝒒𝟐
𝝅(𝒒) = 𝟏𝟎𝟎𝒒 − 𝟐 ( ) + 𝒄 ⇒ 𝟏𝟎𝟎𝒒 − 𝒒𝟐 + 𝒄
𝟐
𝝅(𝒒) = 𝟏𝟎𝟎𝒒 − 𝒒𝟐 + 𝒄 Indeterminate Profit function.
As given, {𝝅(𝒒)}𝒒=𝟏𝟎 = $𝟕𝟎𝟎, it can be substituted as 𝝅(𝟏𝟎) = $𝟕𝟎𝟎
𝝅(𝟏𝟎) = 𝟕𝟎𝟎 = 𝟏𝟎𝟎(𝟏𝟎) − (𝟏𝟎)𝟐 + 𝒄 ⇒ 𝟕𝟎𝟎 = 𝟏𝟎𝟎𝟎 − 𝟏𝟎𝟎 + 𝒄
𝟕𝟎𝟎 = 𝟗𝟎𝟎 + 𝒄
𝒄 = −𝟐𝟎𝟎
𝝅(𝒒) = 𝟏𝟎𝟎𝒒 − 𝒒𝟐 − 𝟐𝟎𝟎
______________________________________________________________________________
Graph of 𝝅(𝒒) = 𝟏𝟎𝟎𝒒 − 𝒒𝟐 − 𝟐𝟎𝟎
Determinate Profit function.

If 𝒒 = 𝟓𝟎, then
𝝅(𝟓𝟎) = 𝟏𝟎𝟎(𝟓𝟎) − (𝟓𝟎)𝟐 − 𝟐𝟎𝟎 = $𝟐𝟑𝟎𝟎
𝝅(𝟓𝟎) = $𝟐𝟑𝟎𝟎
______________________________________________________________________________
Lesson 08
VARIOUS ECONOMIC APPLICATIONS OF INTEGRALS
TOPIC 027: RATE OF GROWTH OF SHOPS USING INTEGRALS
The rate of growth in the number of shops, 𝒏, in a chain of local grocery stores after t
𝒅𝒏 𝟑
months satisfies 𝒅𝒕 = 𝒕.
√
𝟑 𝟏
𝒏(𝒕) = ∫ 𝒏(𝒕) 𝒅𝒕 ⇒ ∫ 𝒅𝒕 ⇒ ∫ 𝟑𝒕−𝟐 𝒅𝒕
√𝒕
𝟏
𝟏 𝒕−𝟐+𝟏
𝒏(𝒕) = ∫ 𝟑𝒕−𝟐 𝒅𝒕 ⇒ 𝟑 {( ) + 𝒄𝟏 }
𝟏
− +𝟏
𝟐
𝟏
𝒕𝟐 𝟏
𝒏(𝒕) = 𝟑 ( ) + 𝟑𝒄𝟏 ⇒ 𝟔𝒕𝟐 + 𝒄
𝟏
𝟐
𝟏
𝒏(𝒕) = 𝟔𝒕𝟐 + 𝒄 Indeterminate solution for shops’ growth function due to 𝒄.
Introducing initial condition: 𝒕 = 𝟎, in the preceding solution.
𝟏
𝒏(𝟎) = 𝟔(𝟎)𝟐 + 𝒄 ⇒ 𝒏(𝟎) = 𝒄 ⇒ 𝒄 = 𝒏(𝟎)
𝟏
𝒏(𝒕) = 𝟔𝒕𝟐 + 𝒏(𝟎)
Scenario: If number of shops (𝒏) are two in the very beginning (𝒏 = 𝟐)𝑡=0 = 𝒏(𝟎) = 𝟐, then the
definite solution will become:
𝟏
𝒏(𝒕) = 𝟔𝒕𝟐 + 𝟐
Furthermore, if 9 months pass (𝒕 = 𝟗) after ‘now’ (𝒕 = 𝟎) we use the value to 𝒕 = 𝟗 in the shops’
growth function:
𝟏
𝒏(𝟗) = 𝟔(𝟗)𝟐 + 𝟐
______________________________________________________________________________
𝟏
Graph of 𝒏(𝟗) = 𝟔(𝟗)𝟐 + 𝟐
𝒏(𝟗) = 𝟏𝟖
⏟ + ⏟
𝟐
𝑻𝒐𝒕𝒂𝒍 𝑺𝒉𝒐𝒑𝒔 𝑫𝒖𝒓𝒊𝒏𝒈 𝑷𝒓𝒆−𝒆𝒙𝒊𝒔𝒕𝒊𝒏𝒈
𝒂𝒇𝒕𝒆𝒓 𝟗 𝒎𝒐𝒏𝒕𝒉𝒔 𝟗 𝒎𝒐𝒏𝒕𝒉𝒔 𝒔𝒉𝒐𝒑𝒔
𝒏(𝟗) = 𝟐𝟎
Interpretation: The number of shops after lapse of 9 months will be 20, where 2 shops were
already there when the time of analysis started and 18 were established during the span of 9
months.
TOPIC 028: OIL EXTRACTION ANALYSIS USING INTEGRALS
A company begins its extraction of oil from a newly discovered oil field at 𝒕 = 𝟎. The
𝒅𝑵
rate of extraction, measured in thousands of barrels per year, is given by 𝒅𝒕 = 𝟑𝟎𝒕𝟐 − 𝟒𝒕𝟑 .
𝒅𝑵
Given that, 𝒅𝒕 = 𝟑𝟎𝒕𝟐 − 𝟒𝒕𝟑 , {𝑵(𝒕)}𝒕=𝟎 = initial condition, 𝑵(𝒕) =?
𝒅𝑵
∫ 𝒅𝒕 = ∫(𝟑𝟎𝒕𝟐 − 𝟒𝒕𝟑 ) 𝒅𝒕 ⇒ ∫ 𝟑𝟎𝒕𝟐 𝒅𝒕 − ∫ 𝟒𝒕𝟑 𝒅𝒕
𝒅𝒕
𝑵(𝒕) = 𝟑𝟎 ∫ 𝒕𝟐 𝒅𝒕 − 𝟒 ∫ 𝒕𝟑 𝒅𝒕 ⇒
𝒕𝟑 𝒕𝟒
𝑵(𝒕) = 𝟑𝟎 ( + 𝒄𝟏 ) − 𝟒 ( + 𝒄𝟐 ) ⇒ (𝟏𝟎𝒕𝟑 + 𝟑𝟎𝒄𝟏 ) − (𝒕𝟒 + 𝟒𝒄𝟐 )
𝟑 𝟒
𝑵(𝒕) = 𝟏𝟎𝒕𝟑 − 𝒕𝟒 + (𝟑𝟎𝒄𝟏 − 𝟒𝒄𝟐 ) ⇒ 𝟏𝟎𝒕𝟑 − 𝒕𝟒 + (𝒄)
𝑵(𝒕) = 𝟏𝟎𝒕𝟑 − 𝒕𝟒 + 𝒄 ⇒ 𝟏𝟎𝒕𝟑 − 𝒕𝟒 + 𝒄
𝑵(𝒕) = 𝟏𝟎𝒕𝟑 − 𝒕𝟒 + 𝒄
Indeterminate Oil Extraction function due to 𝒄.
Introducing Initial Condition: 𝒕 = 𝟎 on both sides:
𝑵(𝟎) = 𝟏𝟎(𝟎)𝟑 − (𝟎)𝟒 + 𝒄
𝑵(𝟎) = 𝟏𝟎(𝟎) − 𝟎 + 𝒄
______________________________________________________________________________
𝒄 = 𝑵(𝟎) arbitrary constant definitized in a symbolic manner rather than a numeric one.
Substituting value of 𝒄, we get the definite solution for Oil Extraction function:
𝑵(𝒕) = 𝟏𝟎𝒕𝟑 − 𝒕𝟒 + 𝑵(𝟎)
Scenario: If company starts extracting from very beginning (𝒕 = 𝟎) , no oil would be extracted till
at that point of time.
𝒕 = 𝟎 ⇒ 𝑵(𝟎) = 𝟎. Under this scenario, oil extraction function is:
𝑵(𝒕) = 𝟏𝟎𝒕𝟑 − 𝒕𝟒
Graph of 𝑵(𝒕) = 𝟏𝟎𝒕𝟑 − 𝒕𝟒
If company extract oil for four years (𝒕 = 𝟒) , then the number of barrels (in thousands) would
be:
𝟑 𝟒
𝑵(𝟒) = 𝟏𝟎(𝟒) − (𝟒) = 𝟑𝟖𝟒 thousand barrels or 𝟑𝟖𝟒, 𝟎𝟎𝟎 barrels of oil would be extracted
after 4 years to oil extraction process.
TOPIC 029: CAPITAL FORMATION USING INTEGRALS

𝟏
Given that, 𝑰(𝒕) = 𝟑𝒕𝟐, 𝑲(𝟎) = 𝟎, 𝑲(𝒕) =?
𝟏
𝑲(𝒕) = ∫ 𝑰(𝒕) 𝒅𝒕 = ∫ 𝟑𝒕𝟐 𝒅𝒕
𝟏 𝟑
𝑲(𝒕) = ∫ 𝟑𝒕𝟐 𝒅𝒕 ⇒ 𝟐𝒕𝟐 + 𝒄
𝟑
𝑲(𝒕) = 𝟐𝒕𝟐 + 𝒄 Indeterminate Capital accumulation function due to 𝒄.
Introducing Initial Condition: 𝒕 = 𝟎 on both sides:
𝟑
𝑲(𝟎) = 𝟐(𝟎)𝟐 + 𝒄
𝑲(𝟎) = 𝒄 Substituting in Indeterminate Capital accumulation function
𝟑
𝑲(𝒕) = 𝟐𝒕𝟐 + 𝑲(𝟎) Determinate Capital accumulation function
______________________________________________________________________________
𝟑
Graph of 𝑲(𝒕) = 𝟐𝒕𝟐 + 𝑲(𝟎)
⏟
𝑙𝑒𝑡 𝐾(0)=1
TOPIC 030: RETAILER'S STORAGE COSTS USING INTEGRALS
A retailer receives a shipment of 10,000 kilograms of rice that will be used up over a 5-month
period at the constant rate of 2,000 kilograms per month. Storage costs are 1 cent per kilogram
per month. The storage costs function is 𝑺(𝒕) = 𝟎. 𝟎𝟏(𝟏𝟎𝟎𝟎𝟎 − 𝟐𝟎𝟎𝟎𝒕).
How much will the retailer pay in storage costs over the next 5 months?
𝒅𝑺
= 𝟎. 𝟎𝟏(𝟏𝟎𝟎𝟎𝟎 − 𝟐𝟎𝟎𝟎𝒕)
𝒅𝒕
𝑺(𝒕) = ∫ 𝑺(𝒕) 𝒅𝒕 ⇒ ∫{𝟎. 𝟎𝟏(𝟏𝟎𝟎𝟎𝟎 − 𝟐𝟎𝟎𝟎𝒕)} 𝒅𝒕 ⇒ 𝟎. 𝟎𝟏 ∫(𝟏𝟎𝟎𝟎𝟎 − 𝟐𝟎𝟎𝟎𝒕) 𝒅𝒕
= 𝟎. 𝟎𝟏 ∫(𝟏𝟎𝟎𝟎𝟎) 𝒅𝒕 − 𝟐𝟎𝟎𝟎 ∫(𝒕) 𝒅𝒕

𝒕𝟐
= 𝟎. 𝟎𝟏 {(𝟏𝟎𝟎𝟎𝟎𝒕) − 𝟐𝟎𝟎𝟎 ( )} + 𝒄 ⇒ 𝟎. 𝟎𝟏(𝟏𝟎𝟎𝟎𝟎𝒕 − 𝟏𝟎𝟎𝟎𝒕𝟐 ) + 𝒄
𝟐
𝑺(𝒕) = 𝟏𝟎𝟎𝒕 − 𝟏𝟎𝒕𝟐 + 𝒄 Indeterminate solution for Storage cost function due to 𝒄.
Introducing initial condition: 𝒕 = 𝟎, in the preceding solution.
𝑺(𝟎) = 𝟏𝟎𝟎(𝟎) − 𝟏𝟎(𝟎)𝟐 + 𝒄 ⇒ 𝑺(𝟎) = 𝒄 ⇒ 𝒄 = 𝑺(𝟎)
𝑺(𝟎) is the initial level of storage cost, when the ship arrives.
During this time of unloading, the storage is not amassed. Therefore, no storage is incurred.
𝑺(𝟎) = 𝟎
𝟐
𝑺(𝒕) = 𝟏𝟎𝟎𝒕 − 𝟏𝟎𝒕 Definite solution for Storage cost function.
______________________________________________________________________________
Graph of 𝑺(𝒕) = 𝟏𝟎𝟎𝒕 − 𝟏𝟎𝒕𝟐
Scenario: Storage cost during five time-intervals (𝒕 = 𝟓):

The definite solution will become:
𝑺(𝒕) = 𝟏𝟎𝟎(𝟓) − 𝟏𝟎(𝟓)𝟐
𝑺(𝒕) = $𝟐𝟓𝟎
Interpretation: The storage costs during the five months will be equal to $ 250.
______________________________________________________________________________
Lesson 09
CONSUMER AND PRODUCER SURPLUS ANALYSIS USING INTEGRALS
TOPIC 031: CONSUMER WILLINGNESS TO PAY USING INTEGRALS
Suppose that the consumers’ demand function for a certain commodity is 𝒑(𝒒) = 𝟒(𝟐𝟓 − 𝒒𝟐 )
dollars per unit. Find the consumer surplus, if consumer is willing to spend 3 units of the
commodity.
Graphically putting:
Given 𝒒 = 𝟑, therefore 𝒑(𝟑) = 𝟒{𝟐𝟓 − (𝟑)𝟐 } = 𝟔𝟒

Visualizing the graph:
Yellow line demarcates an approximate triangle and a rectangle.

Allotting proper names to triangle ⊿𝑷∗ 𝑬𝑷 and rectangle ▯𝑶𝑸∗ 𝑬𝑷∗ :
______________________________________________________________________________
Extraction of consumer surplus can be understood from the version:
Mathematically speaking:
𝟑 𝟑
Area under the curve – Definite Integral: ∫𝟎 𝒑(𝒒) 𝒅𝒒 = ∫𝟎 {𝟒(𝟐𝟓 − 𝒒𝟐 )} 𝒅𝒒
The shape of the whole shaded area is a Trapezoid 𝑶𝑸∗ 𝑬𝑷. (One pair of sides parallel). Economic
sense of it is the ‘consumers’ willingness to pay’ for 3 units.
Expenditure on 3 Units (Price × Quantity): ▯𝑶𝑸∗ 𝑬𝑷∗ = |𝑶𝑸∗ | × |𝑶𝑷∗ | = |𝑷∗ 𝑬| × |𝑬𝑸∗ |
This is a rectangular shape. Economic sense of it is ‘Actual Expenditure’.
Triangular Area: ⊿𝑷∗ 𝑬𝑷
This is a triangular shape. Economic sense of it is ‘Consumer Surplus’.
______________________________________________________________________________
It is evident from the diagram that the area under the curve is the sum of triangular area and the
rectangular area:
𝟑
∫ 𝒑(𝒒) = ▯𝑶𝑸∗ 𝑬𝑷∗ + ⊿𝑷∗ 𝑬𝑷
𝟎
Therefore, triangular area is the difference of area under the curve and expenditure on the 3
units:
𝟑
⊿𝑷∗ 𝑬𝑷 = ∫ 𝒑(𝒒) 𝒅𝒒 − ▯𝑶𝑸∗ 𝑬𝑷∗
𝟎
𝟑
⊿𝑷∗ 𝑬𝑷 = ∫ {𝟒(𝟐𝟓 − 𝒒𝟐 )} 𝒅𝒒 − {|𝑶𝑸∗ | × |𝑶𝑷∗ |}
𝟎
Putting the formula numerically:
𝟑 𝟑 𝟑
𝒒𝟑
∗ 𝟐
⊿𝑷 𝑬𝑷 = 𝟒 (∫ 𝟐𝟓 𝒅𝒒 − ∫ 𝒒 𝒅𝒒) − (𝟑 × 𝟔𝟒) ⇒ 𝟒 {𝟐𝟓(𝒒]𝟑𝟎 ) − ] } − (𝟏𝟗𝟐)
𝟎 𝟎 𝟑 𝟎
(𝟑)𝟑 − (𝟎)𝟑
⊿𝑷∗ 𝑬𝑷 = 𝟒 [𝟐𝟓(𝟑 − 𝟎) − { }] − (𝟏𝟗𝟐) ⇒ 𝟒[𝟐𝟓(𝟑) − (𝟗)] − 𝟏𝟗𝟐
𝟑
⊿𝑷∗ 𝑬𝑷 = 𝟒{𝟐𝟓(𝟑) − (𝟗)} − 𝟏𝟗𝟐 = 𝟕𝟐
Consumer Surplus: ⊿𝑷∗ 𝑬𝑷 = 𝟕𝟐
______________________________________________________________________________
TOPIC 032: CONSUMER SURPLUS USING INTEGRALS
𝟖𝟎
The demand function of a good is given by 𝑷 = 𝟑 dollars per unit. Shade the area which gives
√𝑸
the consumer’s surplus at 𝑸 = 𝟔𝟒 and use integration to calculate its value.
𝟖𝟎 𝟖𝟎
Given 𝑸 = 𝟔𝟒, therefore 𝑷(𝟔𝟒) = 𝟑 = = 𝟐𝟎
√𝟔𝟒 𝟒
Visualizing the graph for Consumer Surplus:
We are looking for consumer surplus.
Allotting proper names to triangle ⊿𝑷∗ 𝑬𝑷 and rectangle ▯𝑶𝑸∗ 𝑬𝑷∗ :
𝟔𝟒 𝟔𝟒 𝟖𝟎
Consumer’s willingness to pay (Area under the curve): ∫𝟎 𝑷(𝑸) 𝒅𝑸 = ∫𝟎 ( 𝟑 ) 𝒅𝑸
√𝑸
Actual Expenditure (Equilibrium Price × Equilibrium Quantity): ▯𝑶𝑸∗ 𝑬𝑷∗ : = |𝑶𝑸∗ | × |𝑶𝑷∗ |
Triangular Area: ⊿𝑷∗ 𝑬𝑷 is the ‘Consumer Surplus’.
______________________________________________________________________________
It is evident from the diagram that the area under the curve is the sum of triangular area and the
rectangular area:
𝟔𝟒
𝟖𝟎
∫ ( 𝟑 ) 𝒅𝑸 = (|𝑶𝑸∗ | × |𝑶𝑷∗ |) + ⊿𝑷∗ 𝑬𝑷
𝟎 √𝑸
Triangular area is the difference of area under the curve and actual expenditure on the 64 units:
𝟔𝟒
𝟖𝟎
⊿𝑷∗ 𝑬𝑷 = ∫ ( 𝟑 ) 𝒅𝑸 − (|𝑶𝑸∗ | × |𝑶𝑷∗ |)
𝟎 √𝑸
Putting the formula numerically:
𝟏 𝟔𝟒
𝟔𝟒 𝟏 𝑸−𝟑+𝟏
⊿𝑷∗ 𝑬𝑷 = 𝟖𝟎 {∫ 𝑸−𝟑 𝒅𝑸} − (𝟔𝟒 × 𝟐𝟎) ⇒ 𝟖𝟎 ( ] ) − 𝟏𝟐𝟖𝟎
𝟏
𝟎 −𝟑 + 𝟏
𝟎
𝟐 𝟔𝟒
𝑸𝟑 𝟑 𝟐 𝟔𝟒 𝟐 𝟐
= 𝟖𝟎 ( ] ) − 𝟏𝟐𝟖𝟎 ⇒ 𝟖𝟎 × (𝑸𝟑 ] ) ⇒ 𝟏𝟐𝟎 (𝟔𝟒𝟑 − 𝟎𝟑 ) − 𝟏𝟐𝟖𝟎
𝟐 𝟐 𝟎
𝟑 𝟎
⊿𝑷∗ 𝑬𝑷 = 𝟏𝟗𝟐𝟎 − 𝟏𝟐𝟖𝟎 = 𝟔𝟒𝟎
TOPIC 033: DIFFERENTIAL OF CONSUMER SURPLUS USING INTEGRALS
A price-discriminating monopolist faces the weekly market demand curve 𝑷 = 𝟔𝟎 − 𝟐𝑸𝟐 , where
𝒒 is the quantity he sells. The monopolist charges two customers their reservation prices.
______________________________________________________________________________
reservation price of the first customer is $𝟒𝟐. The second customer is charged $𝟏𝟎. How much
surplus does the first customer lose from not being charged the lower price?
Given 𝑷 = 𝟔𝟎 − 𝟐𝑸𝟐 , and 𝑷𝑹 𝑹

𝟏 = 𝟒𝟐 and 𝑷𝟐 = 𝟏𝟎
First Customer Analysis Second Customer Analysis

Demand 𝑷𝑹𝟏 = 𝟔𝟎 − 𝟐(𝑸𝟏 )𝟐
𝑷𝑹𝟐 = 𝟔𝟎 − 𝟐(𝑸𝟐 )
𝟐
𝑸𝒅 𝟒𝟐 = 𝟔𝟎 − 𝟐(𝑸𝟏 )𝟐 𝟏𝟎 = 𝟔𝟎 − 𝟐(𝑸𝟐 )𝟐
𝟔𝟎 − 𝟒𝟐 𝟔𝟎 − 𝟏𝟎
(𝑸𝟏 )𝟐 = ⇒ (𝑸𝟏 )𝟐 = 𝟗 (𝑸𝟐 )𝟐 =
𝟐 𝟐
𝑸𝟏 = 𝟑 (𝑸𝟐 )𝟐 = 𝟐𝟓
Quantity demanded by 1st person at 𝑸𝟐 = 𝟓
reservation price. Quantity demanded by 2nd person at
reservation price.
Actual |𝑶𝑸𝟏 | × |𝑶𝑷𝑹 |𝑶𝑸𝟏 | × |𝑶𝑷𝑹
𝟏| 𝟏|
expenditure
𝟑 × 𝟒𝟐 = 𝟏𝟐𝟔 𝟓 × 𝟏𝟎 = 𝟓𝟎
Consumer 𝑸𝟏 𝑸𝟐
Willingness = ∫ (𝟔𝟎 − 𝟐𝑸𝟐 ) 𝒅𝑸 = ∫ (𝟔𝟎 − 𝟐𝑸𝟐 ) 𝒅𝑸
𝟎 𝟎
to pay 𝟑 𝟓
𝟐
= ∫ (𝟔𝟎 − 𝟐𝑸 ) 𝒅𝑸 = ∫ (𝟔𝟎 − 𝟐𝑸𝟐 ) 𝒅𝑸
𝟎 𝟎
𝟑 𝟑 𝟓
𝑸 𝑸𝟑
= (𝟔𝟎𝑸]𝟑𝟎 − 𝟐 ( )] ) = (𝟔𝟎𝑸]𝟑𝟎 − 𝟐 ( )] )
𝟑 𝟎 𝟑 𝟎
𝟑𝟑 𝟎𝟑 𝟓𝟑 𝟎𝟑
= 𝟔𝟎(𝟑 − 𝟎) − 𝟐 ( − ) = 𝟔𝟎(𝟓 − 𝟎) − 𝟐 ( − )
𝟑 𝟑 𝟑 𝟑
= 𝟔𝟎(𝟑) − 𝟐(𝟗) ⇒ 𝟏𝟖𝟎 −18 𝟏𝟐𝟓
= 𝟏𝟔𝟐 = 𝟔𝟎(𝟓) − 𝟐 ( ) ⇒ 𝟑𝟎𝟎 − 𝟖𝟑. 𝟑𝟑
𝟑
= 𝟐𝟏𝟔. 𝟔𝟔𝟕
Consumers 𝑪𝑺𝟏 = 𝟏𝟔𝟐 − 𝟏𝟐𝟔 𝑪𝑺𝟐 = 𝟐𝟏𝟔. 𝟔𝟔𝟕 − 𝟓𝟎
Surplus 𝑪𝑺𝟏 = 𝟑𝟔 𝑪𝑺𝟐 = 𝟏𝟔𝟔. 𝟔𝟔𝟕
Differential ∆𝑪𝑺 = 𝑪𝑺𝟐 − 𝑪𝑺𝟏 = 𝟏𝟔𝟔. 𝟔𝟔𝟕 − 𝟑𝟔
of Consumer ∆𝑪𝑺 = 𝟏𝟑𝟎. 𝟔𝟔𝟕
Surpluses
______________________________________________________________________________
Interpretation The 1st person loses 130.667 from potential surplus when compared to
2nd person.
TOPIC 034: PRODUCER & SOCIAL SURPLUS USING INTEGRALS
Demand and Supply functions of a good (shirts) are 𝑷 = 𝟔𝟎 − 𝟎. 𝟔𝑸 and 𝑷 = 𝟐𝟎 + 𝟎. 𝟐𝑸. Find
total (social) surplus.
For equilibrium output, let the prices in demand and supply functions be equal. 𝟔𝟎 − 𝟎. 𝟔𝑸 = 𝟐𝟎 +
𝟎. 𝟐𝑸
𝟔𝟎 − 𝟐𝟎 = 𝟎. 𝟔𝑸 + 𝟎. 𝟐𝑸 ⇒ 𝟒𝟎 = 𝟎. 𝟖𝑸
𝑸∗ = 𝟓𝟎
For equilibrium price, substitute 𝑸∗ in demand or supply function (we use supply function): 𝑷∗ =
𝟐𝟎 + 𝟎. 𝟐𝑸∗ ⇒ 𝑷∗ = 𝟐𝟎 + 𝟎. 𝟐(𝟓𝟎)
𝑷∗ = 𝟑𝟎
______________________________________________________________________________
Since both consumer surplus and producer surplus are right triangles, we can use area of right
triangle formula to find both:
Consumer Surplus: ⊿𝑨𝑷𝒐 𝑬𝒐

Producer Surplus: ⊿𝑩𝑷𝒐 𝑬𝒐
𝟏
Area of a ⊿ = 𝟐 × |𝑩𝒂𝒔𝒆| × |𝑨𝒍𝒕𝒊𝒕𝒖𝒕𝒆|
𝟏
Consumer Surplus: Area of ⊿𝑨𝑷𝒐 𝑬𝒐 = × |𝑷𝒐 𝑬𝒐 | × |𝑨𝑷𝒐 |
𝟐
𝟏
= × |𝟓𝟎| × |𝟑𝟎|
𝟐
Consumer Surplus: Area of ⊿𝑨𝑷𝒐 𝑬𝒐 = 𝟕𝟓𝟎
𝟏
Producer Surplus: Area of ⊿𝑩𝑷𝒐 𝑬𝒐 = × |𝑷𝒐 𝑬𝒐 | × |𝑩𝑷𝒐 |
𝟐
𝟏
= × |𝟓𝟎| × |𝟏𝟎|
𝟐
Producer Surplus: Area of ⊿𝑩𝑷𝒐 𝑬𝒐 = 𝟐𝟓𝟎
Finally, we calculate the Social (total) Surplus = Consumer Surplus + Producer Surplus
Geometrically; Social (total) Surplus: Area of ⊿𝑨𝑩𝑬𝒐 = ⊿𝑨𝑷𝒐 𝑬𝒐 + ⊿𝑩𝑷𝒐 𝑬𝒐
⊿𝑨𝑩𝑬𝒐 = 𝟕𝟓𝟎 + 𝟐𝟓𝟎
⊿𝑨𝑩𝑬𝒐 = 𝟏𝟎𝟎𝟎
Hence the Social (total) Surplus is 1000 units.
TOPIC 035: PRODUCER SURPLUS WITH TAX USING INTEGRALS
𝑸𝟑
A firm producing computers has the following cost function 𝑪(𝑸) = − 𝟗𝑸𝟐 + 𝟏𝟔𝟎𝑸 + 𝟐𝟎. The
𝟑
producer has planned to produce 12 computers to meet market demand. What is his total
producer surplus if the market price is 270? What is his loss in producer surplus when, due to
taxation, he reduces production from 12 to 9 machines?
Producer’s surplus at 𝑸 = 𝟏𝟐 is to be calculated and market price is 270.
𝒅 𝒅 𝑸𝟑
𝑴𝑪(𝑸) = 𝑪(𝑸) ⇒ 𝑴𝑪(𝑸) = ( − 𝟗𝑸𝟐 + 𝟏𝟔𝟎𝑸 + 𝟐𝟎)
𝒅𝑸 𝒅𝑸 𝟑
𝑴𝑪(𝑸) = 𝑸𝟐 − 𝟏𝟖𝑸 + 𝟏𝟔𝟎
______________________________________________________________________________
As marginal cost is considered equivalent to supply curve, therefore, we treat 𝑴𝑪(𝑸) as the
supply curve.
Producer surplus from the diagram is the difference of Actual Revenue and Area under the curve:
⊿𝑹𝑷∗ 𝑬 = ▯𝑶𝑷∗ 𝑬𝑸∗ − 𝑻𝒓𝒂𝒑𝒆𝒛𝒐𝒊𝒅(𝑶𝑹𝑬𝑸∗ )
𝟏𝟐
⊿𝑹𝑷∗ 𝑬 = (|𝑶𝑸∗ | × |𝑶𝑷∗ |) − ∫𝟎 𝑴𝑪(𝑸) 𝒅𝑸
𝟏𝟐
⊿𝑹𝑷∗ 𝑬 = (|𝟏𝟐| × |𝟐𝟕𝟎|) − ∫𝟎 (𝑸𝟐 − 𝟏𝟖𝑸 + 𝟏𝟔𝟎) 𝒅𝑸
𝟏𝟐
𝑸𝟑
⊿𝑹𝑷 𝑬 = 𝟑𝟐𝟒𝟎 − ( − 𝟗𝑸𝟐 + 𝟏𝟔𝟎𝑸)]
∗
𝟑 𝟎
𝟑 𝟑
𝟏𝟐 𝟎
= 𝟑𝟐𝟒𝟎 − {( − ) − 𝟗(𝟏𝟐𝟐 − 𝟎𝟐 ) + 𝟏𝟔𝟎(𝟏𝟐 − 𝟎)}
𝟑 𝟑
= 𝟑𝟐𝟒𝟎 − (𝟓𝟕𝟔 − 𝟏𝟐𝟗𝟔 + 𝟏𝟗𝟐𝟎) ⇒ 𝟑𝟐𝟒𝟎 − (𝟏𝟐𝟎𝟎)
⊿𝑹𝑷∗ 𝑬 = 𝟐𝟎𝟒𝟎
Interpretation: 2040 is the producer surplus of the company if it produces 12 units with market
price being 270.
______________________________________________________________________________
Loss in Consumer Surplus due to Tax
Post-tax diagram shows that output is discouraged from 𝑸∗ to 𝑸𝟏 . However, the price is still the
same 𝑷∗ = 𝟐𝟕𝟎.
This causes a decline in the revenue: ▯𝑨𝑸𝟏 𝑸∗ 𝑬 = (|𝑨𝑸𝟏 | × |𝑸𝟏 𝑸∗ |)
Within this rectangle, the triangle ⊿𝑨𝑩𝑬 shows the loss in producer surplus.
Whereas, the area Trapezoid(𝑩𝑸𝟏 𝑸∗ 𝑬) is the unwanted area. Mathematically speaking it is the
𝟏𝟐
area under the curve with lower and upper limits as 𝟗 and 𝟏𝟐, respectively: ∫𝟗 𝑴𝑪(𝑸) 𝒅𝑸 ⇒
𝟏𝟐
∫𝟗 (𝑸𝟐 − 𝟏𝟖𝑸 + 𝟏𝟔𝟎) 𝒅𝑸.
From the diagram, the triangle is the difference between decline in revenue and the area under
the curve: ⊿𝑨𝑩𝑬 = ▯𝑨𝑸𝟏 𝑸∗ 𝑬 − 𝐓𝐫𝐚𝐩𝐞𝐳𝐨𝐢𝐝(𝑩𝑸𝟏 𝑸∗ 𝑬)
𝟏𝟐
⊿𝑨𝑩𝑬 = (|𝑨𝑸𝟏 | × |𝑸𝟏 𝑸∗ |) − ∫ 𝑴𝑪(𝑸) 𝒅𝑸
𝟗
𝟏𝟐
⊿𝑨𝑩𝑬 = 𝟐𝟕𝟎 × (𝟏𝟐 − 𝟗) − ∫ (𝑸𝟐 − 𝟏𝟖𝑸 + 𝟏𝟔𝟎) 𝒅𝑸
𝟗
𝟏𝟐
𝑸𝟑
⊿𝑨𝑩𝑬 = 𝟖𝟏𝟎 − ( − 𝟗𝑸𝟐 + 𝟏𝟔𝟎𝑸)]
𝟑 𝟗
𝟏𝟐𝟑 𝟗𝟑
= 𝟖𝟏𝟎 − {( − ) − 𝟗(𝟏𝟐𝟐 − 𝟗𝟐 ) + 𝟏𝟔𝟎(𝟏𝟐 − 𝟗)}
𝟑 𝟑
= 𝟖𝟏𝟎 − (𝟐𝟒𝟔)
⊿𝑨𝑩𝑬 = 𝟓𝟔𝟒
Interpretation: 𝟓𝟔𝟒 units is the loss in producer surplus due to the imposition of the tax.
______________________________________________________________________________
______________________________________________________________________________
Lesson 10
FIRM’S ANALYSIS USING INTEGRALS
TOPIC 036: PRESENT VALUE OF CASH FLOW USING INTEGRALS

𝒕
Present value of cash flow of a firm is represented by: 𝝅 = ∫𝟎 𝑹(𝒕) 𝒆−𝒓𝒕 𝒅𝒕. Where 𝑹(𝒕) is the
function defining the cash flows over time. 𝒕 is time, which is usually measured in years. 𝒓 is the
discount rate. 𝒆−𝒓𝒕 is the exponential-decay expression that depicts the discounting phenomenon.
What is the present value of a continuous revenue flow lasting for 𝒚 years at the constant rate of
𝑫 dollars per year and discounted at the rate of 𝒓 per year?
𝒕
As already known, that: 𝝅 = ∫𝟎 𝑹(𝒕) 𝒆−𝒓𝒕 𝒅𝒕
𝒚
𝒚 𝒆−𝒓𝒕
Here 𝒕 = 𝒚, 𝑹(𝒕) = 𝑫. Therefore; 𝝅 = ∫𝟎 𝑫 𝒆−𝒓𝒕 𝒅𝒕 ⇒ 𝑫 { 𝒅 }]
(−𝒓𝒕)
𝒅𝒕 𝟎
𝒚
𝒚
𝒆−𝒓𝒕 𝒆−𝒓𝒕
= 𝑫{ }] ⇒ 𝑫 { }]
𝒅 (−𝒓) 𝟎
(−𝒓) (𝒕)
𝒅𝒕 𝟎
𝑫 −𝒓𝒕 𝒚
= − (𝒆 )]
𝒓 𝟎
𝑫 −𝒓(𝒚) 𝑫
= − (𝒆 − 𝒆−𝒓(𝟎) ) ⇒ − (𝒆−𝒓𝒚 − 𝒆𝟎 )
𝒓 𝒓
𝑫 −𝒓𝒚 𝑫
𝝅 = − (𝒆 − 𝟏) ⇒ (𝟏 − 𝒆−𝒓𝒚 )
𝒓 𝒓
General formula for present value of revenue flows over 𝒚 periods of time. Formula shows that
the present value (𝝅) depends on 𝑫, 𝒓 and 𝒚.
Numerical: 𝑫 = 𝟑𝟎𝟎𝟎, 𝒓 = 𝟎. 𝟎𝟔, and 𝒚 = 𝟐.
𝑫
Substituting the values in general formula 𝝅 = 𝒓 (𝟏 − 𝒆−𝒓𝒚 )
𝟑𝟎𝟎𝟎
𝝅= {𝟏 − 𝒆−(𝟎.𝟎𝟔)(𝟐) }
𝟎. 𝟎𝟔
𝟑𝟎𝟎𝟎
= (𝟏 − 𝒆−𝟎.𝟏𝟐 ) ⇒ 𝟓𝟎𝟎𝟎𝟎(𝟏 − 𝟎. 𝟖𝟖𝟔𝟗)
𝟎. 𝟎𝟔
𝝅 = 𝟓𝟎𝟎𝟎𝟎(𝟎. 𝟏𝟏𝟑𝟏) = $𝟓𝟔𝟓𝟓 𝑎𝑝𝑝𝑟𝑜𝑥.
Interpretation: Present value of revenue in constant dollar flows of $3000 over 2 year, with
discount rate of 𝟔% is approximately equal to $5655.
TOPIC 037: PRESENT VALUE OF PERPETUAL FLOW USING INTEGRALS
If a cash flow were to persist forever – a situation exemplified by the interest from a perpetual
bond or the revenue from an indestructible capital asset such as land – the present value of the
∞
flow would be: 𝝅 = ∫𝟎 𝑹(𝒕) 𝒆−𝒓𝒕 𝒅𝒕.
It’s an example of improper integral.
Present value of a perpetual income stream flowing at constant rate of 𝑫 dollars per year and
discounted at the rate of 𝒓 per year?
∞
As already known, that: 𝝅𝒑𝒆𝒓𝒑𝒕𝒖𝒂𝒍 = ∫𝟎 𝑹(𝒕) 𝒆−𝒓𝒕 𝒅𝒕
∞ 𝒚
Here 𝑹(𝒕) = 𝑫. Therefore; 𝝅𝒑𝒆𝒓𝒑𝒕𝒖𝒂𝒍 = ∫𝟎 𝑫 𝒆−𝒓(∞) 𝒅𝒕 ⇒ 𝐥𝐢𝐦 {∫𝟎 𝑫 𝒆−𝒓(𝒕) 𝒅𝒕}
𝒚→∞
______________________________________________________________________________
𝒚
𝒚
𝒆−𝒓𝒕 𝒆−𝒓𝒕
= 𝐥𝐢𝐦 [𝑫 { }] ] ⇒ 𝐥𝐢𝐦 [𝑫 { }] ]
𝒚→∞ 𝒅 𝒚→∞ (−𝒓) 𝟎
(−𝒓) (𝒕)
𝒅𝒕 𝟎
𝑫 −𝒓𝒕 𝒚 𝑫
= 𝐥𝐢𝐦 {− (𝒆 )] } ⇒ 𝐥𝐢𝐦 [− {𝒆−𝒓(𝒚) − 𝒆−𝒓(𝟎) }]
𝒚→∞ 𝒓 𝟎 𝒚→∞ 𝒓
𝑫 −𝒓𝒚
= 𝐥𝐢𝐦 {− (𝒆 − 𝒆𝟎 )}
𝒚→∞ 𝒓
𝑫 𝑫
𝝅𝒑𝒆𝒓𝒑𝒕𝒖𝒂𝒍 = 𝐥𝐢𝐦 {− (𝒆−𝒓𝒚 − 𝟏)} ⇒ 𝐥𝐢𝐦 { (𝟏 − 𝒆−𝒓𝒚 )}
𝒚→∞ 𝒓 𝒚→∞ 𝒓
𝑫 −𝒓(∞)
𝑫 −∞
= {𝟏 − 𝒆 } ⇒ {𝟏 − 𝒆 }
𝒓 𝒓
𝑫
= 𝒓 {𝟏 − 𝟎} [∵ 𝒆−∞ = 𝟎]
𝑫
𝝅𝒑𝒆𝒓𝒑𝒕𝒖𝒂𝒍 =
𝒓
General formula for present value of perpetual income stream flowing over time. Formula shows
that this present value (𝝅𝒑𝒆𝒓𝒑𝒕𝒖𝒂𝒍 ) depends on 𝑫 and 𝒓.
Numerical: 𝑫 = 𝑃𝐾𝑅 𝟏𝟒𝟓𝟎 and 𝒓 = 𝟓%.
𝑫
Substituting the values in general formula 𝝅𝒑𝒆𝒓𝒑𝒕𝒖𝒂𝒍 = 𝒓
𝟏𝟒𝟓𝟎 𝟏𝟒𝟓𝟎
𝝅𝒑𝒆𝒓𝒑𝒕𝒖𝒂𝒍 = ⇒
𝟓% 𝟎. 𝟎𝟓
𝝅𝒑𝒆𝒓𝒑𝒕𝒖𝒂𝒍 = 𝑃𝐾𝑅 𝟐𝟗𝟎𝟎𝟎
Interpretation: Present value of the perpetual income stream flowing at constant rate, of 𝟏𝟒𝟓𝟎
dollars per year and discounted at the rate of 𝟓% per year, is PKR 29,000.
TOPIC 038: WINE STORAGE PROBLEM WITH STORAGE COSTS USING INTEGRALS
Wine-storage problem of Sec. 10.6, we assumed zero storage cost. Permit the storage cost.
Let the purchase cost of the case of wine be an amount 𝑪, incurred at the present time. Its (future)
sale value, which varies with time, may be generally denoted as 𝑽(𝒕) – its present value being
𝑽(𝒕)𝒆𝒓𝒕 . Whereas the sale value represents a single future value (there can be only one sale
transaction on this case of wine), the storage cost is a stream. Assuming this cost to be a constant
stream at the rate of 𝒔 dollars per year, the total present value of the storage cost incurred in a
total of £ years can be calculated as follows:
𝑵𝒆𝒕 𝑷𝒓𝒆𝒔𝒆𝒏𝒕 𝒗𝒂𝒍𝒖𝒆 = 𝑷𝒓𝒆𝒔𝒆𝒏𝒕 𝒗𝒂𝒍𝒖𝒆 𝒐𝒇 𝑺𝒂𝒍𝒆 − 𝑷𝒓𝒆𝒔𝒆𝒏𝒕 𝒗𝒂𝒍𝒖𝒆 𝒐𝒇 𝑪𝒐𝒔𝒕𝒔

Instead of Present Value we use Net Present Value, as now there is ‘net/difference’ of Sales and
Costs.
𝑵𝒆𝒕 𝑷𝒓𝒆𝒔𝒆𝒏𝒕 𝑽𝒂𝒍𝒖𝒆 = (𝑵𝑷𝑽)𝒘𝒊𝒏𝒆
= 𝑷𝒓𝒆𝒔𝒆𝒏𝒕 𝒗𝒂𝒍𝒖𝒆 𝒐𝒇 𝒔𝒂𝒍𝒆
− 𝑷𝒓𝒆𝒔𝒆𝒏𝒕 𝒗𝒂𝒍𝒖𝒆 𝒐𝒇 𝒔𝒕𝒐𝒓𝒂𝒈𝒆 𝒄𝒐𝒔𝒕 − 𝑷𝒖𝒓𝒄𝒉𝒂𝒔𝒆 𝒄𝒐𝒔𝒕
𝒓𝒕 )
⏟ × (𝒆
𝑷𝒓𝒆𝒔𝒆𝒏𝒕 𝒗𝒂𝒍𝒖𝒆 𝒐𝒇 𝒔𝒂𝒍𝒆 = 𝑽(𝒕) ⏟ = 𝑽(𝒕)𝒆𝒓𝒕
𝑺𝒂𝒍𝒆 𝑫𝒊𝒔𝒄𝒐𝒖𝒏𝒕𝒊𝒏𝒈
𝒗𝒂𝒍𝒖𝒆 𝒆𝒙𝒑𝒓𝒆𝒔𝒔𝒊𝒐𝒏
______________________________________________________________________________
𝒕
𝑷𝒓𝒆𝒔𝒆𝒏𝒕 𝒗𝒂𝒍𝒖𝒆 𝒐𝒇 𝒔𝒕𝒐𝒓𝒂𝒈𝒆 𝒄𝒐𝒔𝒕 = ∫ ⏟
𝒔 × ⏟−𝒓𝒕 )
(𝒆 𝒅𝒕
𝟎 𝑪𝒐𝒏𝒔𝒕𝒂𝒏𝒕 𝑫𝒊𝒔𝒄𝒐𝒖𝒏𝒕𝒊𝒏𝒈
{𝒄𝒐𝒔𝒕 𝒔𝒕𝒓𝒆𝒂𝒎 𝒆𝒙𝒑𝒓𝒆𝒔𝒔𝒊𝒐𝒏 }
Caveat: Both ‘Present value of sale’ and ‘Present value of storage cost’ are discounted as they
occur in future. While ‘Present value of sale’ is not integrated as it occurs once when the wine will
be sold. Whereas, ‘Present value of storage cost’ is integrated as it occurs repeatedly on yearly
basis.
Now, simplifying the expression for ‘present value of storage cost’:
𝒕 𝒕
= ∫ 𝒔𝒆−𝒓𝒕 𝒅𝒕 ⇒= 𝒔 ∫ 𝒆−𝒓𝒕 𝒅𝒕
𝟎 𝟎
𝒕
𝒆−𝒓𝒕
𝒔{ }]
𝒅
(−𝒓𝒕)
𝒅𝒕 𝟎
𝒕
−𝒓𝒕 𝒕
𝒆 𝒆−𝒓𝒕
= 𝒔{ }] ⇒ 𝒔 { }]
𝒅 (−𝒓) 𝟎
(−𝒓) (𝒕)
𝒅𝒕 𝟎
𝒔 −𝒓𝒕 𝒕
= − (𝒆 )]
𝒓 𝟎
𝒔 −𝒓(𝒕) −𝒓(𝟎)
𝒔 −𝒓𝒚 𝒔
= − (𝒆 −𝒆 ) ⇒ − (𝒆 − 𝒆𝟎 ) ⇒ − (𝒆−𝒓𝒚 − 𝟏)
𝒓 𝒓 𝒓
𝒔
𝑷𝒓𝒆𝒔𝒆𝒏𝒕 𝒗𝒂𝒍𝒖𝒆 𝒐𝒇 𝒔𝒕𝒐𝒓𝒂𝒈𝒆 𝒄𝒐𝒔𝒕 = (𝟏 − 𝒆−𝒓𝒕 )
𝒓
Substituting value in the formula for Net Present Value of Wine.
(𝑵𝑷𝑽)𝒘𝒊𝒏𝒆 = 𝑷𝒓𝒆𝒔𝒆𝒏𝒕 𝒗𝒂𝒍𝒖𝒆 𝒐𝒇 𝒔𝒂𝒍𝒆 − 𝑷𝒓𝒆𝒔𝒆𝒏𝒕 𝒗𝒂𝒍𝒖𝒆 𝒐𝒇 𝒔𝒕𝒐𝒓𝒂𝒈𝒆 𝒄𝒐𝒔𝒕 − 𝑷𝒖𝒓𝒄𝒉𝒂𝒔𝒆 𝒄𝒐𝒔𝒕
𝒔
(𝑵𝑷𝑽)𝒘𝒊𝒏𝒆 = 𝑽(𝒕)𝒆𝒓𝒕 − (𝟏 − 𝒆−𝒓𝒕 ) − 𝑪
𝒓
𝒔 𝒔 −𝒓𝒕 𝒔 𝒔
= 𝑽(𝒕)𝒆 − + 𝒆 − 𝑪 ⇒ 𝑽(𝒕)𝒆𝒓𝒕 + 𝒆−𝒓𝒕 − − 𝑪
𝒓𝒕
𝒓 𝒓 𝒓 𝒓
𝒔 −𝒓𝒕 𝒔
(𝑵𝑷𝑽)𝒘𝒊𝒏𝒆 = 𝑵(𝒕) = {𝑽(𝒕) + } 𝒆 − − 𝑪
⏟
(𝒕)
𝒓 𝒓
𝑵𝑷𝑽 𝒊𝒔 𝒇𝒖𝒏𝒄𝒕𝒊𝒐𝒏 𝒐𝒇
Since the objective is to maximize the NPV of the wine, there is need to optimize:
Differentiating w.r.t. 𝒕:
𝒅 𝒔 𝒔 𝒅 𝒔 𝒅 𝒔 𝒅
𝑵′ (𝒕) = [{𝑽(𝒕) + } 𝒆−𝒓𝒕 − − 𝑪] ⇒ [{𝑽(𝒕) + } 𝒆−𝒓𝒕 ] − ( ) − (𝑪)
𝒅𝒕 𝒓 𝒓 𝒅𝒕 𝒓 𝒅𝒕 𝒓 𝒅𝒕
𝒅 𝒅
Product Rule requires: {𝒇(𝒙)} = {𝒈(𝒙) × 𝒉(𝒙)} = 𝒈′ (𝒙) × 𝒉(𝒙) + 𝒉′ (𝒙) × 𝒈(𝒙)
𝒅𝒕 𝒅𝒕
𝒔 −𝒓𝒕
Here 𝒈(𝒙) = {𝑽(𝒕) + 𝒓
} and 𝒉(𝒙) = 𝒆
𝒅 𝒔 𝒅 𝒔 𝒅
𝑵′ (𝒕) = {𝑽(𝒕) + } × (𝒆
⏟ −𝒓𝒕 )
− ( ) − (𝑪)
𝒅𝒕 ⏟ 𝒓 𝒉(𝒙)
𝒅𝒕 𝒓 𝒅𝒕
[ 𝒈(𝒙) ]
′ (𝒕)
𝒅 𝒔 −𝒓𝒕
𝒅 −𝒓𝒕 𝒔
𝑵 = {𝑽(𝒕) + } × 𝒆 + (𝒆 ) × {𝑽(𝒕) + } − 𝟎 − 𝟎
𝒅𝒕 𝒓 𝒅𝒕 𝒓
𝒅 𝒔
𝑵′ (𝒕) = {𝑽′ (𝒕) + 𝟎} × 𝒆−𝒓𝒕 + [{ (−𝒓𝒕)} (𝒆−𝒓𝒕 )] × {𝑽(𝒕) + }
𝒅𝒕 𝒓
______________________________________________________________________________
𝒔
𝑵′ (𝒕) = {𝑽′ (𝒕)} × 𝒆−𝒓𝒕 + {(−𝒓)(𝒆−𝒓𝒕 )} × {𝑽(𝒕) + }
𝒓
′ (𝒕) ′ (𝒕)𝒆−𝒓𝒕 −𝒓𝒕
𝒔
𝑵 =𝑽 − 𝒓𝒆 × {𝑽(𝒕) + }
𝒓
′ (𝒕) ′ (𝒕)𝒆−𝒓𝒕 −𝒓𝒕
𝒔 −𝒓𝒕
𝑵 =𝑽 − 𝑽(𝒕)𝒓𝒆 − 𝒓𝒆
𝒓
′ (𝒕) ′ (𝒕)
𝑵 = {𝑽 − 𝒓𝑽(𝒕) − 𝒔}𝒆−𝒓𝒕
First order condition requires 𝒇′ (𝒙) = 𝟎:
𝑵′ (𝒕) = {𝑽′ (𝒕) − 𝒓𝑽(𝒕) − 𝒔}𝒆−𝒓𝒕 = 𝟎
{𝑽′ (𝒕) − 𝒓𝑽(𝒕) − 𝒔} = 𝟎 ⇒ 𝑽′ (𝒕) = 𝒓𝑽(𝒕) + 𝒔
Interpretation: This is the condition for maximizing the net present value of the wine-sale
(𝑵𝑷𝑽𝒘𝒊𝒏𝒆 ). The critical value of variable time (𝒕∗ ) will lead to maximization of net present value
of wine-sale. Moreover, the condition is replica of the usual condition for profit maximization 𝑴𝑪 =
𝑴𝑹.
𝑽′ (𝒕) = ⏟
i.e. ⏟ 𝒓𝑽(𝒕) + 𝒔
𝑴𝑹 𝑴𝑪
TOPIC 039: BREAKEVEN ANALYSIS OF A FIRM USING INTEGRALS
A firm expects its annual income, (𝒀), and expenditure (𝑬) to change according to the equations
𝒀(𝒕) = 𝟐𝟓𝟎 + 𝟎. 𝟓𝒕 and 𝑬(𝒕) = 𝒕𝟐 – 𝟏𝟕𝒕 + 𝟐𝟖𝟕. 𝟓 (𝒕 is time in years).
(a) Sketch the graphs of income and expenditure on the same diagram over the interval 𝟎 ≤
𝒕 ≤ 𝟐𝟎. Confirm the break-even points algebraically.
b) Use integration to calculate the net savings (total income – total expenditure) over the
interval 𝟐. 𝟓 ≤ 𝒕 ≤ 𝟏𝟓.
______________________________________________________________________________
From diagram, it clear that:

Net Savings = Area under the Revenue Curve – Area under the Expenditure Curve
𝟏𝟓 𝟏𝟓
𝑵𝒆𝒕 𝑺𝒂𝒗𝒊𝒏𝒈𝒔 = ∫ (𝟐𝟓𝟎 + 𝟎. 𝟓𝒕) 𝒅𝒕– ∫ (𝒕𝟐 – 𝟏𝟕𝒕 + 𝟐𝟖𝟕. 𝟓) 𝒅𝒕
𝟐.𝟓 𝟐.𝟓
𝟏𝟓 𝟏𝟓
𝒕𝟐 𝒕𝟑 𝟏𝟕𝒕𝟐
= (𝟐𝟓𝟎𝒕 + 𝟎. 𝟓 × )] − ( − + 𝟐𝟖𝟕. 𝟓𝒕)]
𝟐 𝟐.𝟓 𝟑 𝟐 𝟐.𝟓
𝟏𝟓
𝟏𝟓𝒕𝟑
= (𝟐𝟓𝟎𝒕 + 𝟎. 𝟐𝟓𝒕 )]𝟐.𝟓 − ( − 𝟖. 𝟓𝒕𝟐 + 𝟐𝟖𝟕. 𝟓𝒕)]
𝟐
𝟑 𝟐.𝟓
= [{𝟐𝟓𝟎(𝟏𝟓) − 𝟐𝟓𝟎(𝟐. 𝟓)} + 𝟎. 𝟐𝟓{(𝟏𝟓) − (𝟐. 𝟓)𝟐 }]
𝟐
(𝟏𝟓)𝟑 (𝟐. 𝟓)𝟑

−{ − − 𝟖. 𝟓{(𝟏𝟓)𝟐 − (𝟐. 𝟓)𝟐 } + 𝟐𝟖𝟕. 𝟓(𝟏𝟓 − 𝟐. 𝟓)}
𝟑 𝟑
𝑵𝒆𝒕 𝑺𝒂𝒗𝒊𝒏𝒈𝒔 = 𝟑𝟐𝟓. 𝟓
Interpretation: The annual net savings of the firm, if time is considered from 2.5 years till
15 years, will be 𝟑𝟐𝟓. 𝟓 units:
TOPIC 040: SCALE OF PRODUCTION ANALYSIS USING INTEGRALS
Assume that the total quantity of a good produced by a firm using capital (𝑲) and labor (𝑳) is
given by the Cobb–Douglas production function 𝒇(𝑲, 𝑳) = 𝑲𝟎.𝟕 𝑳𝟎.𝟑 . What will be the total increase
in output if the firm increases the quantities of capital and labor to 10 units each from the present
levels of 5 units each?
Visualizing the production function, we get a 3D graph:
______________________________________________________________________________
If implies that we have (3D) volume under the curve instead of (2D) area under the curve.
Therefore, developing an integral expression w.r.t. both capital and labor.
𝟏𝟎
𝟏𝟎
𝑸(𝑲, 𝑳) = ∫ {∫ (𝑲𝟎.𝟕 𝑳𝟎.𝟑 )𝒅𝑲} 𝒅𝑳
𝟓
𝟓
It is called a double-definite integral as integrand is integrated w.r.t 𝑲 and 𝑳, respectively.
Moreover, it has lower and upper limits of 𝟓 and 𝟏𝟎, respectively, as 𝑲 and 𝑳 are doubled from 𝟓
to 𝟏𝟎.
Solution: Integrating w.r.t to 𝑲.
𝟏𝟎
𝟏𝟎
𝟎.𝟑
𝑸(𝑲, 𝑳) = ∫ [∫ {𝑲𝟎.𝟕 (𝑳) } 𝒅𝑲] 𝒅𝑳
𝟓
𝟓
𝟏𝟎
𝟏𝟎
𝑲𝟎.𝟕+𝟏 𝟎.𝟑
𝑸(𝑲, 𝑳) = ∫ [{ × (𝑳) }] ] 𝒅𝑳
𝟎. 𝟕 + 𝟏 𝟓
𝟓
𝟏𝟎 𝟏𝟎
𝟎.𝟑 𝑲𝟏.𝟕
𝑸(𝑲, 𝑳) = ∫ [(𝑳) { }] ] 𝒅𝑳
𝟏. 𝟕 𝟓
𝟓
𝟏𝟎
𝟎.𝟑
(𝑳)
𝑸(𝑳) = ∫ { (𝟑𝟒. 𝟔𝟗)} 𝒅𝑳
𝟏. 𝟕
𝟓
𝟏𝟎
𝟎.𝟑
𝑸(𝑳) = ∫ {(𝑳) 𝟐𝟎. 𝟒𝟏} 𝒅𝑳
𝟓
𝟏𝟎
𝟎.𝟑
𝑸(𝑳) = ∫ {𝟐𝟎. 𝟒𝟏(𝑳) } 𝒅𝑳
𝟓
Integrating w.r.t 𝑳:
𝟏𝟎
𝑸(𝑳) = ∫ {𝟐𝟎. 𝟒𝟏(𝑳)𝟎.𝟑 }𝒅𝑳
𝟓
______________________________________________________________________________
𝟏𝟎
𝑳𝟎.𝟑+𝟏
𝑸(𝑳) = 𝟐𝟎. 𝟒𝟏 ( )]
𝟎. 𝟑 + 𝟏 𝟓
𝟏𝟎
𝑳𝟏.𝟑
𝑸(𝑳) = 𝟐𝟎. 𝟒𝟏 ( )]
𝟏. 𝟑 𝟓
𝟏𝟎
𝑸(𝑳) = 𝟏𝟓. 𝟕 {(𝑳𝟏.𝟑 )]𝟓 }
𝑸 = 𝟏𝟓. 𝟕{(𝟏𝟎)𝟏.𝟑 − (𝟓)𝟏.𝟑 }
𝑸 = 𝟏𝟖𝟔. 𝟎𝟑 units
Interpretation: Therefore, the increase in output when the firm increases the quantities of capital
and labor from 𝟓 units each to 𝟏𝟎 units each will be 𝟏𝟖𝟔. 𝟎𝟑 units.
______________________________________________________________________________
Lesson 11
FIRM’S ANALYSIS USING INTEGRALS (CONTINUED 1)
TOPIC 041: POLLUTION IN CATTLE FATTENING ANALYSIS USING INTEGRALS
Cattle Fattening involves environmental costs.
Here an intensive cattle-fattening unit is adding pollutant (in hundreds of gallons) to a river at the
rate of 𝒓(𝒕) = √𝒕, where 𝒕 is the number of months that the unit has been in operation.
(a) Graph the rate of pollution for t = 0 to 12 months. Comment.
(b) Calculate the volume of pollutant in the river during the first six months.
Here 𝒕 = 𝟔, therefore diagram becomes:
______________________________________________________________________________
⊿𝑶𝑩𝑪 = ▯𝐎𝑨𝑩𝑪 − 𝑨𝒓𝒆𝒂 𝒖𝒏𝒅𝒆𝒓 𝒕𝒉𝒆 𝒄𝒖𝒓𝒗𝒆

𝟔
⊿𝑶𝑩𝑪 = (|𝑶𝑨| × |𝑶𝑪|) − ∫ 𝒓(𝒕) 𝒅𝒕
𝟎
𝟔
⊿𝑶𝑩𝑪 = (|𝟐. 𝟒𝟓| × |𝟔|) − ∫ √𝒕 𝒅𝒕
𝟎
𝟑 𝟔
𝒕𝟐
⊿𝑶𝑩𝑪 = 𝟏𝟒. 𝟕 − ( )]
𝟑
𝟐 𝟎
𝟐 𝟑 𝟑
⊿𝑶𝑩𝑪 = 𝟏𝟒. 𝟕 − {(𝟔)𝟐 − (𝟎)𝟐 }
𝟑
𝟐
= 𝟏𝟒. 𝟕 − (𝟏𝟒. 𝟕)
𝟑
= 𝟏𝟒. 𝟕 − 𝟗. 𝟖
⊿𝑶𝑩𝑪 = 𝟒. 𝟗
As the units of pollutant is in hundreds of gallons, the pollutant in first 6 months will be:
⊿𝑶𝑩𝑪 = 𝟒. 𝟗 × 𝟏𝟎𝟎
⊿𝑶𝑩𝑪 = 𝟒𝟗𝟎 gallons
TOPIC 042: CROP FARMING ANALYSIS USING INTEGRALS
It is estimated that 𝒕 days from now a farmer’s crop will be increasing at the rate of
𝟎. 𝟑𝒕𝟐 + 𝟎. 𝟔𝒕 + 𝟏 bushels per day. By how much will the value of the crop increase during the next
𝟓 days if the market price remains fixed at $3 per bushel?
𝒅𝑪(𝒕)
= 𝟎. 𝟑𝒕𝟐 + 𝟎. 𝟔𝒕 + 𝟏
𝒅𝒕
Moreover, it is given that 𝑷 = $𝟑.
Let 𝑪(𝒕) represent the rate of increase in crop bushels per day.
𝒕 𝒕
𝒅𝑪(𝒕)
∫ = ∫(𝟎. 𝟑𝒕𝟐 + 𝟎. 𝟔𝒕 + 𝟏) 𝒅𝒕
𝒅𝒕
𝟎 𝟎
______________________________________________________________________________
𝟓 𝟓
𝒅𝑪(𝒕)
∫ = ∫(𝟎. 𝟑𝒕𝟐 + 𝟎. 𝟔𝒕 + 𝟏) 𝒅𝒕
𝒅𝒕
𝟎 𝟎
𝟓
𝒕𝟑 𝒕𝟐
𝑪(𝒕) = (𝟎. 𝟑 × + 𝟎. 𝟔 + 𝒕)]
𝟑 𝟐 𝟎
𝟓
𝑪(𝒕) = (𝟎. 𝟏𝒕𝟑 + 𝟎. 𝟑𝒕𝟐 + 𝒕)]𝟎
𝑪(𝒕) = [𝟎. 𝟏{(𝟓)𝟑 − (𝟎)𝟑 } + 𝟎. 𝟑{(𝟓)𝟐 − (𝟎)𝟐 } + (𝟓 − 𝟎)]
𝑪(𝒕) = 𝟐𝟓
The crop after 𝟓 months will be 𝟐𝟓 bushels.
It is the crop function over time, however, if the increase in value of crop is required, Value of
crop = crop × price
𝑽𝑪 (𝒕) = 𝑪(𝒕) × 𝑷
𝟓
𝑽𝑪 (𝒕) = (𝟎. 𝟏𝒕𝟑 + 𝟎. 𝟑𝒕𝟐 + 𝒕)] × $𝟑
𝟎
𝟑 𝟐 𝟓
𝑽𝑪 (𝒕) = (𝟎. 𝟑𝒕 + 𝟎. 𝟗𝒕 + 𝟑𝒕)]𝟎
Applying the lower and upper limits in 𝑽𝑪 (𝒕), we get:
𝑽𝑪 (𝒕) = [𝟎. 𝟑{(𝟓)𝟑 − (𝟎)𝟑 } + 𝟎. 𝟗{(𝟓)𝟐 − (𝟎)𝟐 } + 𝟑(𝟓 − 𝟎)]
𝑽𝑪 (𝒕) = $𝟕𝟓
Interpretation: The value of crop will be $𝟕𝟓, after 𝟓 months, if the price is $3 per bushel.
TOPIC 043: MOBILE PRODUCTION ANALYSIS USING INTEGRALS
Bejax Corporation has set up a production line to manufacture a new type of mobiles. The rate of
𝒅𝑷 𝒕
production of the mobiles is = 𝟏𝟓𝟎𝟎 (𝟐 − ) units/month. How many telephones are
𝒅𝒕 𝟐𝒕+𝟓
produced during the third month?
Given the rate of change of production over time:
𝒅𝑷 𝒕
= 𝟏𝟓𝟎𝟎 (𝟐 − )
𝒅𝒕 𝟐𝒕 + 𝟓
Integrating w.r.t 𝒕 to get the production as a function of time (𝒕).
𝒅𝑷 𝒕
∫ 𝒅𝒕 = ∫ {𝟏𝟓𝟎𝟎 (𝟐 − )} 𝒅𝒕
𝒅𝒕 𝟐𝒕 + 𝟓
𝒕
𝑷(𝒕) = 𝟏𝟓𝟎𝟎 (∫ 𝟐 𝒅𝒕 − ∫ 𝒅𝒕)
𝟐𝒕 + 𝟓
𝒕
𝑷(𝒕) = 𝟏𝟓𝟎𝟎 (𝟐𝒕 − ∫ 𝒅𝒕 )
𝟐𝒕 + 𝟓
𝒕
∫ 𝟐𝒕+𝟓 𝒅𝒕 requires integration by substitution:
𝒖−𝟓 𝒅𝒖 𝒅𝒖
Let 𝒖 = 𝟐𝒕 + 𝟓 ⇒ 𝟐
= 𝒕, 𝒅𝒕
=𝟐⇒ 𝟐
= 𝒅𝒕
Substituting, we get:
(𝒖 − 𝟓)
{ 𝟐 } 𝒅𝒖 (𝒖 − 𝟓)
∫ ⇒∫ 𝒅𝒖
𝒖 𝟐 𝟒𝒖
𝟏 𝟓
= ∫ (𝟏 − ) 𝒅𝒖
𝟒 𝒖
______________________________________________________________________________
𝟏 𝟏
= (∫ 𝟏 𝒅𝒖 − 𝟓 ∫ 𝒅𝒖)
𝟒 𝒖
𝟏
= (𝒖 − 𝟓 𝐥𝐧|𝒖|) + 𝑪𝒖
𝟒
𝟏
= {(𝟐𝒕 + 𝟓) − 𝟓 𝐥𝐧|𝟐𝒕 + 𝟓|} + 𝑪(𝟐𝒕+𝟓)
𝟒
𝟏
= (𝟐𝒕 + 𝟓 − 𝟓 𝐥𝐧|𝟐𝒕 + 𝟓|) + 𝒄
𝟒
[∵ 𝑪(𝟐𝒕+𝟓) = 𝒄′ ]
𝒕 𝟏
∫ 𝒅𝒕 = {(𝟐𝒕 + 𝟓) − 𝟓 𝐥𝐧|𝟐𝒕 + 𝟓|} + 𝒄′
𝟐𝒕 + 𝟓 𝟒
𝒕
𝑷(𝒕) = 𝟏𝟓𝟎𝟎 (𝟐𝒕 − ∫ 𝒅𝒕 )
𝟐𝒕 + 𝟓
𝟏
𝑷(𝒕) = 𝟏𝟓𝟎𝟎 {𝟐𝒕 − {(𝟐𝒕 + 𝟓) − 𝟓 𝐥𝐧|𝟐𝒕 + 𝟓|} }
𝟒
𝟐𝒕 + 𝟓 𝟓
𝑷(𝒕) = 𝟏𝟓𝟎𝟎 {𝟐𝒕 − ( − 𝐥𝐧|𝟐𝒕 + 𝟓|)}
𝟒 𝟒
𝟐𝒕 + 𝟓 𝟓
𝑷(𝒕) = 𝟏𝟓𝟎𝟎 (𝟐𝒕 − + 𝐥𝐧|𝟐𝒕 + 𝟓|)
𝟒 𝟒
𝟑
Recalling the given information regarding time: “during the 3rd month”. Mathematically; ∫𝟐 𝑷(𝒕)𝒅𝒕
which implies 𝒕𝒍𝒐𝒘𝒆𝒓 = 𝟐 and 𝒕𝒖𝒑𝒑𝒆𝒓 = 𝟑.
𝟐(𝟑)+𝟓 𝟐(𝟐)+𝟓 𝟓
= 𝟏𝟓𝟎𝟎 [𝟐(𝟑) − 𝟐(𝟐) − { 𝟒
− 𝟒 } + 𝟒 {𝐥𝐧|𝟐(𝟑) + 𝟓| − 𝐥𝐧|𝟐(𝟐) + 𝟓|}]
𝟏𝟏 𝟗 𝟓
= 𝟏𝟓𝟎𝟎 [𝟐 − ( − ) + (𝐥𝐧|𝟏𝟏| − 𝐥𝐧|𝟗|)]
𝟒 𝟒 𝟒
𝟏 𝟓 𝟏𝟏
= 𝟏𝟓𝟎𝟎 {𝟐 − + (𝐥𝐧 | |)}
𝟐 𝟒 𝟗
𝟑 𝟓 𝟏𝟏
= 𝟏𝟓𝟎𝟎 { + (𝐥𝐧 | |)}
𝟐 𝟒 𝟗
𝑷(𝒕) = 𝟐𝟔𝟐𝟔. 𝟐𝟓𝟕𝟓
Interpretation: Approximately, 𝟐𝟔𝟐𝟔 telephones will be produced during the third month by Bejax
Corporation.
TOPIC 044: FIRM'S AVERAGE PROFIT ANALYSIS USING INTEGRALS
Firm’s profit function is 𝝅 = 𝒇(𝑸) and the level of output (𝑸), varies between 𝒂 and 𝒃 units, then
𝟏 𝒃
the firm’s average profit is 𝝅 = 𝒃−𝒂 ∫𝒂 𝒇(𝑸) 𝒅𝑸. Calculate a firm’s average profit when 𝑻𝑹 =
𝟏𝟎𝟎(𝟏 − 𝒆−𝟎.𝟏𝑸 ), 𝑻𝑪 = 𝟎. 𝟏𝑸𝟐 + 𝟐𝑸 + 𝟏 and output varies from 3 to 8 units.
Solution: 𝝅 = 𝒇(𝑸) = 𝑻𝑹(𝑸) − 𝑻𝑪(𝑸)
𝝅 = 𝒇(𝑸) = {𝟏𝟎𝟎(𝟏 − 𝒆−𝟎.𝟏𝑸 )} − {𝟎. 𝟏𝑸𝟐 + 𝟐𝑸 + 𝟏}
𝝅 = 𝒇(𝑸) = 𝟏𝟎𝟎 − 𝟏𝟎𝟎𝒆−𝟎.𝟏𝑸 − 𝟎. 𝟏𝑸𝟐 − 𝟐𝑸 − 𝟏
𝝅 = 𝒇(𝑸) = 𝟗𝟗 − 𝟏𝟎𝟎𝒆−𝟎.𝟏𝑸 − 𝟐𝑸 − 𝟎. 𝟏𝑸𝟐
Introducing integral with upper and lower limits on general formula of average profit:
𝟏 𝒃 𝟏 𝟖
𝝅𝒂𝒗𝒈 = 𝒃−𝒂 ∫𝒂 𝒇(𝑸) 𝒅𝑸 ⇒ 𝟖−𝟑 ∫𝟑 (𝟗𝟗 − 𝟏𝟎𝟎𝒆−𝟎.𝟏𝑸 − 𝟐𝑸 − 𝟎. 𝟏𝑸𝟐 ) 𝒅𝑸
𝟏 𝟖
𝝅𝒂𝒗𝒈 = 𝟓 ∫𝟑 (𝟗𝟗 − 𝟏𝟎𝟎𝒆−𝟎.𝟏𝑸 − 𝟐𝑸 − 𝟎. 𝟏𝑸𝟐 ) 𝒅𝑸
______________________________________________________________________________
𝟖
𝟏 𝒆−𝟎.𝟏𝑸
𝝅𝒂𝒗𝒈 (𝑸) = {𝟗𝟗𝑸 − 𝟏𝟎𝟎 ( ) − 𝑸𝟐 − 𝟎. 𝟎𝟑𝑸𝟑 }]
𝟓 −𝟎. 𝟏 𝟑
𝟏 𝟖
𝝅𝒂𝒗𝒈 (𝑸) = {𝟗𝟗𝑸 + 𝟏𝟎𝟎𝟎(𝒆−𝟎.𝟏𝑸 ) − 𝑸𝟐 − 𝟎. 𝟎𝟑𝑸𝟑 }]𝟑
𝟓
𝟏
𝝅𝒂𝒗𝒈 (𝑸) = [{𝟗𝟗(𝟖 − 𝟑)} + 𝟏𝟎𝟎𝟎{𝒆−𝟎.𝟏(𝟖) − 𝒆−𝟎.𝟏(𝟑) } − (𝟖𝟐 − 𝟑𝟐 ) − 𝟎. 𝟎𝟑(𝟖𝟑 − 𝟑𝟑 )]
𝟓
𝟏
𝝅𝒂𝒗𝒈 (𝑸) = {𝟗𝟗(𝟓) + 𝟏𝟎𝟎𝟎(𝒆−𝟎.𝟖 − 𝒆−𝟎.𝟑 ) − (𝟓𝟓) − 𝟎. 𝟎𝟑(𝟒𝟖𝟓)}
𝟓
𝟏
𝝅𝒂𝒗𝒈 (𝑸) = {𝟒𝟗𝟓 + 𝟏𝟎𝟎𝟎(𝟎. 𝟒𝟒𝟗 − 𝟎. 𝟕𝟒𝟏) − 𝟓𝟓 − 𝟏𝟒. 𝟓𝟓}
𝟓
𝝅𝒂𝒗𝒈 (𝑸) = 𝟐𝟔. 𝟒𝟔
Interpretation: Between 𝟑 and 𝟖 units of output, the firm makes 𝟐𝟔. 𝟒𝟔 units of average profit.
TOPIC 045: FIRM'S AVERAGE COST ANALYSIS USING INTEGRALS
The cost of producing 𝒙 units of a new product is 𝑪(𝑸) = 𝟑𝑸𝟐 − 𝟐 hundred dollars. What is the
average cost of producing 𝟑 to 𝟕 units?
Solution: Given the total cost function, 𝑪(𝑸) = 𝟑𝑸𝟐 − 𝟐, we can find the average cost between
the lower and upper limits [𝟑, 𝟕] by general formula for average of a function using definite
integrals:
𝒃
𝟏
𝒇𝒂𝒗𝒈 (𝒙) = ∫ 𝒇(𝒙) 𝒅𝒙
𝒃−𝒂
𝒂
𝟏 𝟕
𝑪𝒂𝒗𝒈 (𝑸) = 𝟕−𝟑 ∫𝟑 (𝟑𝑸𝟐 − 𝟐) 𝒅𝑸
𝟕
𝟏
𝑪𝒂𝒗𝒈(𝑸) = 𝑨𝑪(𝑸) = ∫(𝟑𝑸𝟐 − 𝟐) 𝒅𝑸
𝟒
𝟑
𝟕 𝟕
𝟏
𝑨𝑪(𝑸) = {𝟑 ∫(𝑸𝟐 ) 𝒅𝑸 − 𝟐 ∫(𝟏) 𝒅𝑸}
𝟒
𝟑 𝟑
𝟑 𝟕
𝟏 𝑸
𝑨𝑪(𝑸) = {𝟑 ( ] ) − 𝟐(𝑸]𝟕𝟑 )}
𝟒 𝟑 𝟑
𝟏 𝟕
𝑨𝑪(𝑸) = {(𝑸𝟑 ]𝟑 ) − 𝟐(𝑸]𝟕𝟑 )}
𝟒
𝟏
𝑨𝑪(𝑸) = {(𝟕𝟑 − 𝟑𝟑 ) − 𝟐(𝟕 − 𝟑)}
𝟒
𝟏
𝑨𝑪(𝑸) = {(𝟑𝟏𝟔) − 𝟐(𝟒)}
𝟒
𝑨𝑪(𝑸) = 𝟕𝟕
Interpretation: The firm will incur 𝟕𝟕 units as average cost if we consider production of 𝟑 to 𝟕
units.
______________________________________________________________________________
Lesson 12
TOPIC 046: AVERAGE FOOD PRICE ANALYSIS USING INTEGRALS
Records indicate that 𝒕 months after the beginning of the year, the price of ground beef in local
supermarkets was 𝑷(𝒕) = 𝟎. 𝟎𝟗𝒕𝟐 − 𝟎. 𝟐𝒕 + 𝟒 dollars per pound. What was the average price of
ground beef during the first 𝟑 months of the year?
Solution: Given the price function, 𝑷(𝒕) = 𝟎. 𝟎𝟗𝒕𝟐 − 𝟎. 𝟐𝒕 + 𝟒, we can find the average price of
food between the lower and upper limits [𝟎, 𝟑] by general formula for average of a function using
definite integrals:
𝒃
𝟏
𝒃−𝒂
𝒂
Here averaging of price is over time (𝒕) instead of output.
𝟑
𝟏
𝑷𝒂𝒗𝒈 (𝒕) = ∫(𝟎. 𝟎𝟗𝒕𝟐 − 𝟎. 𝟐𝒕 + 𝟒) 𝒅𝒕
𝟑−𝟎
𝟎
𝟑
𝟏 𝒕𝟑 𝒕𝟐
𝑷𝒂𝒗𝒈 (𝒕) = × (𝟎. 𝟎𝟗 − 𝟎. 𝟐 + 𝟒𝒕)]
𝟑 𝟑 𝟐 𝟎
𝟏 𝟑 𝟐 𝟑
𝑷𝒂𝒗𝒈 (𝒕) = × (𝟎. 𝟎𝟑𝑸 − 𝟎. 𝟏𝒕 + 𝟒𝒕)]𝟎
𝟑
𝟏
𝑷𝒂𝒗𝒈 (𝒕) = × [𝟎. 𝟎𝟑{(𝟑)𝟑 − (𝟎)𝟑 } − 𝟎. 𝟏{(𝟑)𝟐 − (𝟎)𝟐 } + 𝟒(𝟑 − 𝟎)]
𝟑
𝟏
𝑷𝒂𝒗𝒈 (𝒕) = × [𝟎. 𝟎𝟑(𝟐𝟕) − 𝟎. 𝟏(𝟗) + 𝟒(𝟑)]
𝟑
𝟏
𝑷𝒂𝒗𝒈 (𝒕) = × [𝟎. 𝟎𝟑(𝟐𝟕) − 𝟎. 𝟏(𝟗) + 𝟒(𝟑)]
𝟑
𝟏
𝑷𝒂𝒗𝒈 (𝒕) = × (𝟏𝟏. 𝟗𝟏)
𝟑
𝑷𝒂𝒗𝒈 (𝒕) = 𝟑. 𝟗𝟕
Interpretation: The average price of beef during first 𝟑 months in the local supermarket will be
𝟑. 𝟗𝟕 dollar per pound.
TOPIC 047: AVERAGE PRODUCTION ANALYSIS USING INTEGRALS
A company determines that if 𝑳 worker-hours of labor are employed, then Q units of a particular
𝟐
commodity will be produced, where 𝑸(𝑳) = 𝟓𝟎𝟎𝑳𝟑. What is the average production as labor varies
from 𝟏, 𝟎𝟎𝟎 to 𝟐, 𝟎𝟎𝟎 worker-hours?
𝟐
Solution: Given the production function, 𝑸(𝑳) = 𝟓𝟎𝟎𝑳𝟑 , we can find the average production
between the lower and upper limits [𝟏𝟎𝟎𝟎, 𝟐𝟎𝟎𝟎] by general formula for average of a function
using definite integrals:
𝒃
𝟏
𝒃−𝒂
𝒂
______________________________________________________________________________
𝟐𝟎𝟎𝟎
𝟏 𝟐
𝑸𝒂𝒗𝒈 (𝑳) = ∫ (𝟓𝟎𝟎𝑳𝟑 ) 𝒅𝑳
𝟐𝟎𝟎𝟎 − 𝟏𝟎𝟎𝟎
𝟏𝟎𝟎𝟎
𝟐 𝟐𝟎𝟎𝟎
𝟏 𝑳𝟑+𝟏
𝑸𝒂𝒗𝒈 (𝑳) = 𝑨𝑷𝑳 = × 𝟓𝟎𝟎 ( )]
𝟏𝟎𝟎𝟎 𝟐
+𝟏
𝟑 𝟏𝟎𝟎𝟎
𝟓 𝟐𝟎𝟎𝟎
𝟏 𝑳𝟑
𝑨𝑷𝑳 = × 𝟓𝟎𝟎 ( )]
𝟏𝟎𝟎𝟎 𝟓
𝟑 𝟏𝟎𝟎𝟎
𝟏 𝟑 𝟓 𝟐𝟎𝟎𝟎
𝑨𝑷𝑳 = × 𝟓𝟎𝟎 × (𝑳𝟑 )]
𝟏𝟎𝟎𝟎 𝟓 𝟏𝟎𝟎𝟎
𝟑 𝟓 𝟐𝟎𝟎𝟎
𝑨𝑷𝑳 = (𝑳𝟑 )]
𝟏𝟎 𝟏𝟎𝟎𝟎
𝟑 𝟓 𝟓
𝑨𝑷𝑳 = {(𝟐𝟎𝟎𝟎)𝟑 − (𝟏𝟎𝟎𝟎)𝟑 }
𝟏𝟎
𝟑
𝑨𝑷𝑳 = {𝟐𝟏𝟕𝟒𝟖𝟎. 𝟐𝟏}
𝟏𝟎
𝑨𝑷𝑳 = 𝟔𝟓𝟐𝟒𝟒. 𝟎𝟔
Interpretation: The average product of labor between 𝟏𝟎𝟎𝟎 and 𝟐𝟎𝟎𝟎, in the company will be
𝟔𝟓𝟐𝟒𝟒. 𝟎𝟔.
TOPIC 048: AVERAGE INVENTORY ANALYSIS USING INTEGRALS
An inventory of 60,000 kilograms of a certain commodity is used at a constant rate and is

exhausted after 1 year. If 𝒚(𝒕) = −𝟔𝟎, 𝟎𝟎𝟎(𝒕 − 𝟏), then what is the average inventory for the year?
Solution: Given the inventory function, 𝒚(𝒕) = −𝟔𝟎, 𝟎𝟎𝟎(𝒕 − 𝟏), we can find the average
inventory function between the lower and upper limits [𝟎, 𝟏] by general formula for average of a
function using definite integrals:
𝒃
𝟏
𝒃−𝒂
𝒂
𝟏
𝟏
𝑰𝒂𝒗𝒈 (𝒕) = ∫{−𝟔𝟎𝟎𝟎𝟎(𝒕 − 𝟏)} 𝒅𝒕
𝟏−𝟎
𝟎
𝟏
𝑰𝒂𝒗𝒈 (𝒕) = 𝑨𝑰𝒕 = ∫{−𝟔𝟎𝟎𝟎𝟎(𝒕 − 𝟏)} 𝒅𝒕

𝟎
𝟏
𝑨𝑰𝒕 = −𝟔𝟎𝟎𝟎𝟎 ∫(𝒕 − 𝟏) 𝒅𝒕

𝟎
𝟏 𝟏
𝑨𝑰𝒕 = −𝟔𝟎𝟎𝟎𝟎 (∫(𝒕) 𝒅𝒕 − ∫(𝟏) 𝒅𝒕)

𝟎 𝟎
𝟐 𝟏
𝒕
𝑨𝑰𝒕 = −𝟔𝟎𝟎𝟎𝟎 ( ] − 𝒕]𝟏𝟎 )
𝟐 𝟎
______________________________________________________________________________
𝟏𝟐 − 𝟎𝟐
𝑨𝑰𝒕 = −𝟔𝟎𝟎𝟎𝟎 {( ) − (𝟏 − 𝟎)}
𝟐
𝟏
𝑨𝑰𝒕 = −𝟔𝟎𝟎𝟎𝟎 { − 𝟏}
𝟐
𝟏
𝑨𝑰𝒕 = −𝟔𝟎𝟎𝟎𝟎 {− }
𝟐
𝑨𝑰𝒕 = 𝟑𝟎𝟎𝟎𝟎
Interpretation: The average inventory of the commodity during the year will be 𝟑𝟎𝟎𝟎𝟎 kilograms.
TOPIC 049: NET SAVINGS ANALYSIS USING INTEGRALS
A hospital trust decides to install a new rooftop solar power system which will save money on its
energy bills. It estimates that t years after installation the savings flow (in $1000s per year) will
steadily increase according to 𝑺(𝒕) = 𝟎. 𝟑𝒕 + 𝟒𝟓. 𝟔. The flow of money needed to repair and
maintain the system in subsequent years is modelled by 𝑪(𝒕) = 𝟎. 𝟓𝒕𝟐 + 𝟐𝒕:
(a) Draw a sketch graph of the savings and cost functions on the same diagram.
(b) Solve the equation 𝑺(𝒕) = 𝑪(𝒕) to determine the period during which the savings exceed
the maintenance costs, and shade the area which determines the net savings during this
time.
Solution: Visualizing two functions i.e. Savings 𝑺(𝒕) and Cost 𝑪(𝒕).
The intersection point of the two functions 𝑺(𝒕) and 𝑪(𝒕) shall exist at a certain 𝒕. It can be found
by equating the two functions, i.e. 𝑺(𝒕) = 𝑪(𝒕).
𝟎. 𝟓𝒕𝟐 + 𝟐𝒕 = 𝟎. 𝟑𝒕 + 𝟒𝟓. 𝟔
𝟎. 𝟓𝒕𝟐 + 𝟐𝒕 − 𝟎. 𝟑𝒕 − 𝟒𝟓. 𝟔 = 𝟎
𝟎. 𝟓𝒕𝟐 + 𝟏. 𝟕𝒕 − 𝟒𝟓. 𝟔 = 𝟎 ⇒ 𝒕𝟐 + 𝟑. 𝟒𝒕 − 𝟗𝟏. 𝟐 = 𝟎
𝒕 + 𝟑. 𝟒𝒕 − 𝟗𝟏. 𝟐 = 𝟎 is comparable with 𝒂𝒙𝟐 + 𝒃𝒙 + 𝒄 = 𝟎
𝟐
−𝒃±√𝒃𝟐 −𝟒𝒂𝒄
This implies: 𝒙 = 𝒕, 𝒂 = 𝟏, 𝒃 = 𝟑. 𝟒 and 𝒄 = −𝟗𝟏. 𝟐. and using 𝒙 = 𝟐𝒂
−(𝟑. 𝟒) ± √(𝟑. 𝟒)𝟐 − 𝟒(𝟏)(−𝟗𝟏. 𝟐)
𝒕=
𝟐(𝟏)
______________________________________________________________________________
−𝟑. 𝟒 ± √𝟏𝟏. 𝟓𝟔 + 𝟑𝟔𝟒. 𝟖

𝒕=
𝟐
−𝟑. 𝟒 ± √𝟑𝟕𝟔. 𝟑𝟔
𝒕=
𝟐
−𝟑. 𝟒 ± 𝟏𝟗. 𝟒 −𝟑. 𝟒 + 𝟏𝟗. 𝟒 −𝟑. 𝟒 − 𝟏𝟗. 𝟒
𝒕= ⇒ ,
𝟐 𝟐 𝟐
𝒕 = 𝟖, −𝟏𝟏. 𝟒
𝒕∗ = 𝟖 [∵ 𝒕 ≮ 𝟎]
Putting 𝒕 = 𝟖 in 𝑺(𝒕) or 𝑪(𝒕), we get their common value 𝑪∗ (𝒕∗ ) = 𝑺∗ (𝒕∗ ):
∗
𝟎. 𝟓(𝒕∗ )𝟐 + 𝟐𝒕∗ = 𝟎. 𝟑𝒕∗ + 𝟒𝟓. 𝟔

𝟎. 𝟓(𝟖)𝟐 + 𝟐(𝟖) = 𝟎. 𝟑(𝟖) + 𝟒𝟓. 𝟔
𝟑𝟐 + 𝟏𝟔 = 𝟐. 𝟒 + 𝟒𝟓. 𝟔
𝟒𝟖 = 𝟒𝟖 = 𝑪∗ (𝒕∗ ) = 𝑺∗ (𝒕∗ )
(c) Use integration to find the exact value of net savings

Net Savings: Net savings is the shaded area, which is the difference of the two integrals:
i.e. Net Savings = Area under Savings curve – Area under Savings curve
______________________________________________________________________________
Mathematically:
𝟖 𝟖
⊿𝑶𝑬𝑨 = ∫ 𝑺(𝒕) 𝒅𝒕 − ∫ 𝑪(𝒕) 𝒅𝒕
𝟎 𝟎
𝟖 𝟖
⊿𝑶𝑬𝑨 = ∫ (𝟎. 𝟑𝒕 + 𝟒𝟓. 𝟔) 𝒅𝒕 − ∫ (𝟎. 𝟓𝒕𝟐 + 𝟐𝒕) 𝒅𝒕
𝟎 𝟎
𝟖 𝟖 𝟖 𝟖
⊿𝑶𝑬𝑨 = 𝟎. 𝟑 ∫ (𝒕) 𝒅𝒕 + 𝟒𝟓. 𝟔 ∫ (𝟏) 𝒅𝒕 − {𝟎. 𝟓 ∫ (𝒕𝟐 ) 𝒅𝒕 + 𝟐 ∫ (𝒕) 𝒅𝒕 }
𝟎 𝟎 𝟎 𝟎
𝟖 𝟖 𝟖 𝟖
⊿𝑶𝑬𝑨 = 𝟎. 𝟑 ∫ (𝒕) 𝒅𝒕 + 𝟒𝟓. 𝟔 ∫ (𝟏) 𝒅𝒕 − {𝟎. 𝟓 ∫ (𝒕𝟐 ) 𝒅𝒕 + 𝟐 ∫ (𝒕) 𝒅𝒕 }
𝟎 𝟎 𝟎 𝟎
𝟖 𝟖 𝟖
𝒕𝟐 𝟖 𝒕𝟑 𝒕𝟐
⊿𝑶𝑬𝑨 = 𝟎. 𝟑 ( ] ) + 𝟒𝟓. 𝟔(𝒕]𝟎 ) − {𝟎. 𝟓 ( ] ) + 𝟐 ( ] ) }
𝟐 𝟎 𝟑 𝟎 𝟐 𝟎
(𝟖)𝟐 (𝟎)𝟐 (𝟖)𝟑 (𝟎)𝟑 (𝟖)𝟐 (𝟎)𝟐
⊿𝑶𝑬𝑨 = 𝟎. 𝟑 { 𝟐
− 𝟐
} + 𝟒𝟓. 𝟔(𝟖 − 𝟎) − [𝟎. 𝟓 { 𝟑 − 𝟑 } + 𝟐 { 𝟐 − 𝟐 }]
⊿𝑶𝑬𝑨 = 𝟎. 𝟑(𝟑𝟐) + 𝟒𝟓. 𝟔(𝟖) − {𝟎. 𝟓(𝟏𝟕𝟎. 𝟔𝟕) + 𝟐(𝟑𝟐)}
⊿𝑶𝑬𝑨 = 𝟐𝟐𝟓. 𝟎𝟔𝟓
Since given data is in $𝟏𝟎𝟎𝟎𝒔 per year, then answer shall be 𝟐𝟐𝟓. 𝟎𝟔𝟓 × 𝟏𝟎𝟎𝟎 = $𝟐𝟐𝟓𝟎𝟔𝟓.
Interpretation: Net savings in installing a new rooftop solar power system, during 𝟖 years, shall
be $𝟐𝟐𝟓𝟎𝟔𝟓.
TOPIC 050: NET EXCESS PROFIT ANALYSIS USING INTEGRALS
The area between curves can sometimes be used as a way of measuring the amount of a quantity
that has been accumulated during a particular procedure. For instance, suppose that 𝒕 years from
now, two investment plans will be generating profit 𝑷𝟏 (𝒕) and 𝑷𝟐 (𝒕), respectively, and that their
respective rates of profitability, 𝑷′𝟏 (𝒕) and 𝑷′𝟐 (𝒕), are expected to satisfy 𝑷′𝟐 (𝒕) ≥ 𝑷′𝟏 (𝒕) for the next
𝑵 years; that is, over the time interval 𝟎 ≤ 𝒕 ≤ 𝑵. Then 𝑬(𝒕) = 𝑷𝟐 (𝒕) − 𝑷𝟏 (𝒕) represents the
excess profit of plan 𝟐 over plan 𝟏 at time 𝒕, and the net excess profit 𝑵𝑬 = 𝑬(𝑵) − 𝑬(𝟎) over
the time interval 𝟎 ≤ 𝒕 ≤ 𝑵 is given by the definite integral:
______________________________________________________________________________
𝑵𝑬 = 𝑵𝑬 = 𝑬(𝑵) − 𝑬(𝟎) = ∫ 𝑬′ (𝒕) 𝒅𝒕

𝟎
𝑬(𝒕) = 𝑷𝟐 (𝒕) − 𝑷𝟏 (𝒕) ⇒ 𝑬′ (𝒕) = 𝑷′𝟐 (𝒕) − 𝑷′𝟏 (𝒕)
𝑵
𝑵𝑬 = ∫{𝑷′𝟐 (𝒕) − 𝑷′𝟏 (𝒕)} 𝒅𝒕

𝟎
Graphically:
Numerical Example: Suppose that t years from now, one investment will be generating profit at
the rate of 𝑷′𝟏 (𝒕) = 𝟓𝟎 + 𝒕𝟐 hundred dollars per year, while a second investment will be generating
profit at the rate of 𝑷′𝟐 (𝒕) = 𝟐𝟎𝟎 + 𝟓𝒕 hundred dollars per year.
a. For how many years does the rate of profitability of the second investment exceed that of the
first?
From the diagram, before the intersection point one of the rates of profit is greater than other.
Therefore, we need to find the point of intersection.
Mathematically speaking; 𝑷′𝟏 (𝒕) = 𝑷′𝟐 (𝒕)
𝟓𝟎 + 𝒕𝟐 = 𝟐𝟎𝟎 + 𝟓𝒕 ⇒ 𝒕𝟐 + 𝟓𝟎 − 𝟐𝟎𝟎 − 𝟓𝒕 = 𝟎
𝒕𝟐 − 𝟓𝒕 − 𝟏𝟓𝟎 = 𝟎
Factorization:
𝒕𝟐 − 𝟏𝟓𝒕 + 𝟏𝟎𝒕 − 𝟏𝟓𝟎 = 𝟎
𝒕(𝒕 − 𝟏𝟓) + 𝟏𝟎(𝒕 − 𝟏𝟓) = 𝟎
(𝒕 + 𝟏𝟎)(𝒕 − 𝟏𝟓) = 𝟎
(𝒕 + 𝟏𝟎) = 𝟎 OR (𝒕 − 𝟏𝟓) = 𝟎
𝒕 = −𝟏𝟎 OR 𝒕 = 𝟏𝟓
Negative value of time (𝒕 < 𝟎) is meaningless in current analysis.
______________________________________________________________________________
b. Compute the net excess profit for the time period determined in part (a). Interpret the net excess
profit as an area.
Given the formula of Net Excess Profit:
𝑵
𝑵𝑬 = ∫{𝑷′𝟐 (𝒕) − 𝑷′𝟏 (𝒕)} 𝒅𝒕

𝟎
Substituting values of 𝑷′𝟐 (𝒕), 𝑷′𝟏 (𝒕) and 𝑵, we get:
𝟏𝟓
𝑵𝑬 = ∫ {(𝟐𝟎𝟎 + 𝟓𝒕) − (𝟓𝟎 + 𝒕𝟐 )} 𝒅𝒕

𝟎
𝟏𝟓 𝟏𝟓
𝟓(𝒕)𝟐 (𝒕)𝟑
𝑵𝑬 = {𝟐𝟎𝟎(𝒕) + }] − {𝟓𝟎(𝒕) + }]
𝟐 𝟎
𝟑 𝟎
𝟓(𝟏𝟓 − 𝟎)𝟐 (𝟏𝟓 − 𝟎)𝟑
𝑵𝑬 = {𝟐𝟎𝟎(𝟏𝟓 − 𝟎) + } − {𝟓𝟎(𝟏𝟓 − 𝟎) + }
𝟐 𝟑
𝟓(𝟐𝟐𝟓) 𝟑𝟑𝟕𝟓
𝑵𝑬 = {𝟐𝟎𝟎(𝟏𝟓) + } − {𝟓𝟎(𝟏𝟓) + }
𝟐 𝟑
𝑵𝑬 = {𝟑𝟎𝟎𝟎 + 𝟓𝟔𝟐. 𝟓} − {𝟕𝟓𝟎 + 𝟏𝟏𝟐𝟓}
𝑵𝑬 = 𝟑𝟎𝟎𝟎 + 𝟓𝟔𝟐. 𝟓 − 𝟕𝟓𝟎 − 𝟏𝟏𝟐𝟓
𝑵𝑬 = 𝟏𝟔𝟖𝟕. 𝟓 hundred dollars OR 𝑵𝑬 = $𝟏𝟔𝟖, 𝟕𝟓𝟎
Interpretation: Thus, the net excess profit (𝑬𝑵) is $𝟏𝟔𝟖, 𝟕𝟓𝟎. The graphs of the rate of
profitability functions 𝑷′𝟏 (𝒕) and 𝑷′𝟐 (𝒕) are shown in Figure. The net excess profit can be
interpreted as the area of the (shaded) region between the rate of profitability curves over the
interval 𝟎 ≤ 𝒕 ≤ 𝑵.
______________________________________________________________________________
Lesson 13
TOPIC 051: GINI INDEX USING INTEGRALS
Area also plays an important role in the study of Lorentz curves which is used by economists to
measure income inequality. It’s a graph of the function 𝑳(𝒙):
The ratio of area 𝑹𝟏 to the area of the region 𝑹𝟐 under the complete equality line 𝒚 = 𝒙 over 𝟎 ≤
𝒙 ≤ 𝟏 is used as a measure of the inequality in the distribution of wealth in the society.
This ratio, called the Gini index, denoted GI (also called the index of income inequality).
𝑎𝑟𝑒𝑎 𝑏⁄𝑤 𝒚=𝑳(𝒙) & 𝒚=𝒙
More precisely, 𝑮𝑰 = 𝑎𝑟𝑒𝑎 𝑢𝑛𝑑𝑒𝑟 𝒚=𝒙 𝑜𝑣𝑒𝑟 𝟎≤𝒙≤𝟏
𝟏
∫𝟎 [𝒙 − 𝑳(𝒙)] 𝒅𝒙
𝑮𝑰 = 𝟏
∫𝟎 (𝒙) 𝒅𝒙
______________________________________________________________________________
𝟏
∫𝟎 [𝒙 − 𝑳(𝒙)] 𝒅𝒙
𝑮𝑰 = 𝟏
𝒙𝟐
( 𝟐 )]
𝟎
𝟏
∫ [𝒙 − 𝑳(𝒙)] 𝒅𝒙
𝑮𝑰 = 𝟎
(𝟏)𝟐 − (𝟎)𝟐
{ 𝟐 }
𝟏
∫𝟎 [𝒙 − 𝑳(𝒙)] 𝒅𝒙
𝑮𝑰 =
𝟏
𝟐
𝟏
𝑮𝑰 = 𝟐 ∫ {𝒙 − 𝑳(𝒙)} 𝒅𝒙
𝟎
Numerical Example: A governmental agency determines that the Lorentz curves for the
distribution of income for dentists and contractors in a certain state are given by the functions
𝑳𝟏 (𝒙) = 𝒙𝟏.𝟕 and 𝑳𝟐 (𝒙) = 𝟎. 𝟖𝒙𝟐 + 𝟎. 𝟐𝒙 respectively. For which profession is the distribution of
income more fairly distributed?
Solution: Gini Index, for both dentists and contractors, needs to be calculated and compared.
𝑮𝑰𝒅𝒆𝒏𝒕𝒊𝒔𝒕𝒔 ⋚ 𝑮𝑰𝒄𝒐𝒏𝒕𝒓𝒂𝒄𝒕𝒐𝒓𝒔
Here 𝑳𝟏 (𝒙): Lorenz curve for the distribution of income for dentists.
and 𝑳𝟐 (𝒙): Lorenz curve for the distribution of income for contractors.
Firstly, calculating Gini Index for dentists:
𝟏
𝑮𝑰𝒅𝒆𝒏𝒕𝒊𝒔𝒕𝒔 = 𝟐 ∫ {𝒙 − 𝑳𝟏 (𝒙)} 𝒅𝒙
𝟎
𝟏
𝑮𝑰𝒅𝒆𝒏𝒕𝒊𝒔𝒕𝒔 = 𝟐 ∫ {𝒙 − (𝒙)𝟏.𝟕 } 𝒅𝒙
𝟎
𝟏 𝟏
𝒙𝟐 𝒙𝟐.𝟕
𝑮𝑰𝒅𝒆𝒏𝒕𝒊𝒔𝒕𝒔 = 𝟐 {( )] − ( )] }
𝟐 𝟎 𝟐. 𝟕 𝟎
(𝟏)𝟐 − (𝟎)𝟐 (𝟏)𝟐.𝟕 − (𝟎)𝟐.𝟕
𝑮𝑰𝒅𝒆𝒏𝒕𝒊𝒔𝒕𝒔 = 𝟐 [{ }−{ }]
𝟐 𝟐. 𝟕
𝟏 𝟏
𝑮𝑰𝒅𝒆𝒏𝒕𝒊𝒔𝒕𝒔 = 𝟐 [{ } − { }]
𝟐 𝟐. 𝟕
𝑮𝑰𝒅𝒆𝒏𝒕𝒊𝒔𝒕𝒔 = 𝟎. 𝟐𝟓𝟗𝟑
Secondly, calculating Gini Index for contractors:
𝟏
𝑮𝑰𝒄𝒐𝒏𝒕𝒓𝒂𝒄𝒕𝒐𝒓𝒔 = 𝟐 ∫ {𝒙 − (𝟎. 𝟖𝒙𝟐 + 𝟎. 𝟐𝒙)} 𝒅𝒙
𝟎
𝟏
𝑮𝑰𝒄𝒐𝒏𝒕𝒓𝒂𝒄𝒕𝒐𝒓𝒔 = 𝟐 ∫ {𝒙 − 𝟎. 𝟖𝒙𝟐 − 𝟎. 𝟐𝒙} 𝒅𝒙
𝟎
𝟏
𝑮𝑰𝒄𝒐𝒏𝒕𝒓𝒂𝒄𝒕𝒐𝒓𝒔 = 𝟐 ∫ {𝟎. 𝟖𝒙 − 𝟎. 𝟖𝒙𝟐 } 𝒅𝒙
𝟎
𝟏
𝑮𝑰𝒄𝒐𝒏𝒕𝒓𝒂𝒄𝒕𝒐𝒓𝒔 = 𝟐 × 𝟎. 𝟖 ∫ (𝒙 − 𝒙𝟐 ) 𝒅𝒙
𝟎
𝟏 𝟏
𝑮𝑰𝒄𝒐𝒏𝒕𝒓𝒂𝒄𝒕𝒐𝒓𝒔 = 𝟏. 𝟔 {∫ (𝒙) 𝒅𝒙 − ∫ (𝒙𝟐 ) 𝒅𝒙}
𝟎 𝟎
______________________________________________________________________________
𝟏 𝟏
𝒙𝟐 𝒙𝟑
𝑮𝑰𝒄𝒐𝒏𝒕𝒓𝒂𝒄𝒕𝒐𝒓𝒔 = 𝟏. 𝟔 {( )] − ( )] }
𝟐 𝟎 𝟑 𝟎
𝟐
(𝟏) − (𝟎) 𝟐 (𝟏)𝟑 − (𝟎)𝟑
𝑮𝑰𝒄𝒐𝒏𝒕𝒓𝒂𝒄𝒕𝒐𝒓𝒔 = 𝟏. 𝟔 [{ }−{ }]
𝟐 𝟑
𝟏 𝟏
𝑮𝑰𝒄𝒐𝒏𝒕𝒓𝒂𝒄𝒕𝒐𝒓𝒔 = 𝟏. 𝟔 [{ } − { }]
𝟐 𝟑
𝑮𝑰𝒄𝒐𝒏𝒕𝒓𝒂𝒄𝒕𝒐𝒓𝒔 = 𝟐. 𝟔𝟔𝟕
Interpretation: As 𝑮𝑰𝒅𝒆𝒏𝒕𝒊𝒔𝒕𝒔 = 𝟎. 𝟐𝟓𝟗𝟑 and 𝑮𝑰𝒄𝒐𝒏𝒕𝒓𝒂𝒄𝒕𝒐𝒓𝒔 = 𝟐. 𝟔𝟔𝟕, and it is evident that

𝑮𝑰𝒅𝒆𝒏𝒕𝒊𝒔𝒕𝒔 < 𝑮𝑰𝒄𝒐𝒏𝒕𝒓𝒂𝒄𝒕𝒐𝒓𝒔. It is straightforward to suggest that income inequality is higher among
contractors as measured through Gini Index.
TOPIC 052: SALES ANALYSIS USING INTEGRALS
A manufacturer determines that 𝒕 months after introducing a new product, the company’s sales
will be 𝑺(𝒕) thousand dollars, where
𝟕𝟓𝟎𝒕
𝑺(𝒕) =
√𝟒𝒕𝟐 + 𝟐𝟓
What are the average monthly sales of the company over the first 𝟔 months after the introduction
of the new product?
Solution:
Given sales function:
𝟕𝟓𝟎𝒕
𝑺(𝒕) =
√𝟒𝒕𝟐 + 𝟐𝟓
Time span is first 𝟔 month: [𝟎, 𝟔].
Average monthly sales are to be calculated, so using the integral for the average function:
𝒃
𝟏
𝒃−𝒂
𝒂
Substituting the values in the general formula above.
𝟔
𝟏
𝑺𝒂𝒗𝒈 (𝒕) = ∫ 𝑺(𝒕) 𝒅𝒕
𝟔−𝟎
𝟎
𝟔
𝟏 𝟕𝟓𝟎𝒕
𝑺𝒂𝒗𝒈 (𝒕) = ∫( ) 𝒅𝒕
𝟔 √𝟒𝒕𝟐 + 𝟐𝟓
𝟎
𝟔
𝟏 𝟕𝟓𝟎𝒕
𝑺𝒂𝒗𝒈 (𝒕) = ∫( ) 𝒅𝒕
𝟔 √𝟒𝒕𝟐 + 𝟐𝟓
𝟎
𝟔 𝒕 𝒅𝒕 𝟏
Let 𝑿 = 𝟕𝟓𝟎 ∫𝟎 ( ) then 𝑺𝒂𝒗𝒈 (𝒕) = 𝟔 (𝑿)
√𝟒𝒕𝟐 +𝟐𝟓
𝟔 𝒕 𝒅𝒕
𝑿 = 𝟕𝟓𝟎 ∫𝟎 ( ) requires integration by substitution:
√𝟒𝒕𝟐 +𝟐𝟓
Derivative Limits Conversion

Let 𝒖 = 𝟒𝒕𝟐 + 𝟐𝟓 If Lower Limit: 𝒕 = 𝟎 ⇒ 𝒖 = 𝟒(𝟎)𝟐 + 𝟐𝟓
Then 𝒖 = 𝟐𝟓
______________________________________________________________________________
𝒅𝒖 𝒅𝒖 If Upper Limit: 𝒕 = 𝟔 ⇒ 𝒖 = 𝟒(𝟔)𝟐 + 𝟐𝟓

= 𝟖𝒕 ⇒ = 𝒕 𝒅𝒕
𝒅𝒕 𝟖 Then 𝒖 = 𝟏𝟔𝟗
𝟔 𝒕 𝒅𝒕
Substituting the tabulated values in 𝐗 = 𝟕𝟓𝟎 ∫𝟎 ( ), we get:
√𝟒𝒕𝟐 +𝟐𝟓
𝒅𝒖
𝟏𝟔𝟗 𝟏𝟔𝟗
𝟖 𝟕𝟓𝟎 𝒅𝒖
𝑿 = 𝟕𝟓𝟎 ∫ ( )⇒ ∫ ( 𝟏)
√𝒖 𝟖
𝟐𝟓 𝒖𝟐 𝟐𝟓
𝟏𝟔𝟗 𝟏 𝟏𝟔𝟗
𝟏 𝒖−𝟐+𝟏
𝑿 = 𝟗𝟑. 𝟕𝟓 ∫ (𝒖−𝟐 ) 𝒅𝒖 ⇒ 𝟗𝟑. 𝟕𝟓 ( )]
𝟏
𝟐𝟓 − +𝟏
𝟐 𝟐𝟓
𝟏 𝟏𝟔𝟗
𝟏𝟔𝟗
𝒖𝟐 𝟏
𝑿 = 𝟗𝟑. 𝟕𝟓 ( )] ⇒ 𝟗𝟑. 𝟕𝟓 × 𝟐 (𝒖𝟐 )]
𝟏 𝟐𝟓
𝟐 𝟐𝟓
𝟏 𝟏𝟔𝟗
𝑿 = 𝟏𝟖𝟕. 𝟓 {(𝒖)𝟐 }]
𝟐𝟓
𝟏 𝟏
= 𝟏𝟖𝟕. 𝟓 {(𝟏𝟔𝟗)𝟐 − (𝟐𝟓)𝟐 }
𝑿 = 𝟏𝟓𝟎𝟎
𝟏
Putting value of 𝑿 back in 𝑺𝒂𝒗𝒈 (𝒕) = 𝟔
(𝑿)
𝟏
𝑺𝒂𝒗𝒈 (𝒕) =
(𝟏𝟓𝟎𝟎)
𝟔
𝑺𝒂𝒗𝒈 (𝒕) = 𝟐𝟓𝟎 thousand dollars
Interpretation: Thus, for the 𝟔-month period immediately after the introduction of the new
product, the company’s sales average $𝟐𝟓𝟎, 𝟎𝟎𝟎 per month.
TOPIC 053: EFFICIENCY ANALYSIS USING INTEGRALS
After 𝒕 months on the job, a postal clerk can sort 𝑸(𝒕) = 𝟕𝟎𝟎 − 𝟒𝟎𝟎𝒆−𝟎.𝟓𝒕 letters per hour. What
is the average rate at which the clerk sorts mail during the first 3 months on the job?
Solution: Given the sorting-efficiency function, 𝑸(𝒕) = 𝟕𝟎𝟎 − 𝟒𝟎𝟎𝒆−𝟎.𝟓𝒕 , we can find the average
sorting-efficiency function between the lower and upper limits [𝟎, 𝟑] by general formula for
average of a function using definite integrals:
𝒃
𝟏
𝒃−𝒂
𝒂
𝟑
𝟏
𝑸𝒂𝒗𝒈 (𝒕) = ∫(𝟕𝟎𝟎 − 𝟒𝟎𝟎𝒆−𝟎.𝟓𝒕 ) 𝒅𝒕
𝟑−𝟎
𝟎
𝟑 𝟑
𝟏
𝑸𝒂𝒗𝒈 (𝒕) = {𝟕𝟎𝟎 ∫(𝟏) 𝒅𝒕 − 𝟒𝟎𝟎 ∫(𝒆−𝟎.𝟓𝒕 ) 𝒅𝒕}
𝟑
𝟎 𝟎
𝟑
𝟏 𝒆−𝟎.𝟓𝒕
𝑸𝒂𝒗𝒈 (𝒕) = {𝟕𝟎𝟎(𝒕]𝟑𝟎 ) − 𝟒𝟎𝟎 ( )] }
𝟑 −𝟎. 𝟓 𝟎
______________________________________________________________________________
𝟏 𝟑
𝑸𝒂𝒗𝒈 (𝒕) = {𝟕𝟎𝟎(𝒕]𝟑𝟎 ) + 𝟖𝟎𝟎(𝒆−𝟎.𝟓𝒕 )]𝟎 }
𝟑
𝟏
𝑸𝒂𝒗𝒈 (𝒕) = [𝟕𝟎𝟎(𝟑 − 𝟎) + 𝟖𝟎𝟎{𝒆−𝟎.𝟓(𝟑) − 𝒆−𝟎.𝟓(𝟎) }]
𝟑
𝟏
𝑸𝒂𝒗𝒈 (𝒕) = {𝟕𝟎𝟎(𝟑) + 𝟖𝟎𝟎(𝒆−𝟏.𝟓 − 𝒆𝟎 )}
𝟑
𝟏
𝑸𝒂𝒗𝒈 (𝒕) = {𝟕𝟎𝟎(𝟑) + 𝟖𝟎𝟎(𝟎. 𝟐𝟐𝟑 − 𝟏)}
𝟑
𝟏
𝑸𝒂𝒗𝒈 (𝒕) = {𝟕𝟎𝟎(𝟑) + 𝟖𝟎𝟎(−𝟎. 𝟕𝟕𝟕)}
𝟑
𝟏
𝑸𝒂𝒗𝒈 (𝒕) = (𝟐𝟏𝟎𝟎 − 𝟔𝟐𝟏. 𝟔)
𝟑
𝟏
𝑸𝒂𝒗𝒈 (𝒕) = (𝟏𝟒𝟕𝟖. 𝟒)
𝟑
𝑸𝒂𝒗𝒈 (𝒕) = 𝟒𝟗𝟐. 𝟖𝟑
Interpretation: The average efficiency of the clerk in sorting mail during the first 3 months is
𝟒𝟗𝟐. 𝟖𝟑 letter per hour.
TOPIC 054: ADVERTISEMENT ANALYSIS USING INTEGRALS
An advertising agency begins a campaign to promote a new product and determines that 𝒕 days
later, the number of people 𝑵(𝒕) who have heard about the product is changing at a rate given by
𝟎.𝟎𝟒𝒕
𝑵′ (𝒕) = 𝟓𝒕𝟐 − 𝟐 people per day. How many people learn about the product during the first
𝒕 +𝟑
week? During the second week?
𝟎.𝟎𝟒𝒕
Solution: Given the rate at which people hear about the new product: 𝑵′ (𝒕) = 𝟓𝒕𝟐 − 𝟐 .
𝒕 +𝟑
Analysis for first week implies that lower limit is 𝟎 and upper limit is 𝟕; i.e. [𝟎, 𝟕]
Using this information, we can express the situation mathematically:
𝟕 𝟕
𝒅 𝟎. 𝟎𝟒𝒕
∫ {𝑵′ (𝒕)} = ∫ (𝟓𝒕𝟐 − 𝟐 ) 𝒅𝒕
𝒅𝒕 𝒕 +𝟑
𝟎 𝟎
𝟕 𝟕
𝟐
𝒕
𝑵(𝒕) = 𝟓 ∫(𝒕 ) 𝒅𝒕 − 𝟎. 𝟎𝟒 ∫ ( ) 𝒅𝒕
𝒕𝟐 +𝟑
𝟎 𝟎
𝒃 𝑓′ (𝑥) 2
In attempt of applying: ∫𝒂 { 𝑓(𝑥) } 𝑑𝑥 = ln|𝑓(𝑥)| , introducing (2)
𝟕 𝟕
𝟐
𝟐 𝒕
𝑵(𝒕) = 𝟓 ∫(𝒕 ) 𝒅𝒕 − 𝟎. 𝟎𝟒 × ( ) ∫ ( 𝟐 ) 𝒅𝒕
𝟐 𝒕 +𝟑
𝟎 𝟎
𝟕 𝟕
𝟐𝒕
𝑵(𝒕) = 𝟓 ∫(𝒕𝟐 ) 𝒅𝒕 − 𝟎. 𝟎𝟐 ∫ ( ) 𝒅𝒕
𝒕𝟐 +𝟑
𝟎 𝟎
𝟕
𝒕𝟑 𝟕
𝑵(𝒕) = 𝟓 ( )] − 𝟎. 𝟎𝟐 ( 𝒍𝒏|𝒕𝟐 + 𝟑| )]
𝟑 𝟎 𝟎
𝟓
𝑵(𝒕) = {(𝟕)𝟑 − (𝟎)𝟑 } − 𝟎. 𝟎𝟐{𝒍𝒏|(𝟕)𝟐 + 𝟑| − 𝒍𝒏|(𝟎)𝟐 + 𝟑|}
𝟑
𝟓
𝑵(𝒕) = (𝟑𝟒𝟑) − 𝟎. 𝟎𝟐{𝟐. 𝟖𝟓}
𝟑
______________________________________________________________________________
𝑵(𝒕) = 𝟓𝟕𝟐 approximately
Interpretation: During the first week (seven days), approximately 𝟓𝟕𝟐 people will learn about the
new product.
If we want to know about the same number during next week [𝟕, 𝟏𝟒], the mathematical formation
will be:
𝟏𝟒 𝟏𝟒
𝒅 𝟎. 𝟎𝟒𝒕
∫ {𝑵′ (𝒕)} = ∫ (𝟓𝒕𝟐 − 𝟐 ) 𝒅𝒕
𝒅𝒕 𝒕 +𝟑
𝟕 𝟕
{𝑵(𝒕)}𝟐𝒏𝒅 𝒘𝒆𝒆𝒌 = 𝟒𝟎𝟎𝟐 approximately
Percentage increase in the number of people getting aware during first and second week:
[{𝑵(𝒕)}𝟐𝒏𝒅 𝒘𝒆𝒆𝒌 − {𝑵(𝒕)}𝟏𝒔𝒕 𝒘𝒆𝒆𝒌 ]
%∆{𝑵(𝒕)} = × 𝟏𝟎𝟎
{𝑵(𝒕)}𝟏𝒔𝒕 𝒘𝒆𝒆𝒌
𝟒𝟎𝟎𝟐 − 𝟓𝟕𝟐
= × 𝟏𝟎𝟎
𝟓𝟕𝟐
𝟑𝟒𝟑𝟎
= × 𝟏𝟎𝟎
𝟓𝟕𝟐
%∆{𝑵(𝒕)} = 𝟔𝟎𝟎%
Interpretation: It is noticeable that there is a rapid increase in the number of people getting aware
of the product during 2nd week. The advertisement campaign appears a success.
TOPIC 055: PROFIT OVER USEFUL LIFE OF MACHINE USING INTEGRALS
Suppose that 𝒕 years after being put into use, a machine has generated total revenue 𝑹(𝒕) and
that the total cost of operating and servicing the machine up to this time is 𝑪(𝒕). Then the total
profit generated by the machine up to time 𝒕 is 𝑷(𝒕) = 𝑹(𝒕) − 𝑪(𝒕). Profit declines when costs
accumulate at a higher rate than revenue; that is, when 𝑪′ (𝒕) > 𝑹′ (𝒕). Thus, a manager may
consider disposing of the machine at the time 𝑻 when 𝑪′ (𝑻) = 𝑹′ (𝑻), and for this reason, the time
period 𝟎 ≤ 𝒕 ≤ 𝑻 is called the useful life of the machine. The net profit over the useful life of the
machine provides the manager with a measure of its value.
Numerical Instance: Suppose that when it is 𝒕 years old, a particular industrial machine is
generating revenue at the rate 𝑹′ (𝒕) = 𝟓, 𝟎𝟎𝟎 − 𝟐𝟎𝒕𝟐 dollars per year and that operating and
______________________________________________________________________________
servicing costs related to the machine are accumulating at the rate 𝑪′ (𝒕) = 𝟐, 𝟎𝟎𝟎 + 𝟏𝟎𝒕𝟐 dollars
per year.
a) What is the useful life of this machine?
b) Compute the net profit generated by the machine over its period of useful life.
Solution: Visualizing two functions i.e. Revenue 𝑹(𝒕) and Cost 𝑪(𝒕).
The intersection point of the two functions 𝑹′(𝒕) and 𝑪′(𝒕) shall exist at a certain point on f x-axis
(𝒕∗ ). It can be found by equating the two functions, i.e. 𝑹′ (𝒕) = 𝑪′ (𝒕).
𝟓𝟎𝟎𝟎 − 𝟐𝟎𝒕𝟐 = 𝟐𝟎𝟎𝟎 + 𝟏𝟎𝒕𝟐
𝟐𝟎𝒕𝟐 + 𝟏𝟎𝒕𝟐 − 𝟓𝟎𝟎𝟎 + 𝟐𝟎𝟎𝟎 = 𝟎
𝟑𝟎𝒕𝟐 − 𝟑𝟎𝟎𝟎 = 𝟎
𝒕𝟐 − 𝟏𝟎𝟎 = 𝟎
𝒕𝟐 = 𝟏𝟎𝟎
𝒕 = 𝟏𝟎 [∵ 𝒕 ≰ 𝟎]
Interpretation: Useful life of machine is till 10 years. This information gives us lower and upper
limit; i.e. [𝟎, 𝟏𝟎]
To find the profit over the same interval, we evaluate the profit function {𝝅(𝒕) = 𝑹(𝒕) − 𝑪(𝒕)} over
[𝟎, 𝟏𝟎].
𝟏𝟎 𝟏𝟎
∫ 𝝅(𝒕) 𝒅𝒕 = ∫ {𝑹(𝒕) − 𝑪(𝒕)} 𝒅𝒕

𝟎 𝟎
𝟏𝟎 𝟏𝟎
∫ 𝝅(𝒕) 𝒅𝒕 = ∫ {(𝟓𝟎𝟎𝟎 − 𝟐𝟎𝒕𝟐 ) − (𝟐𝟎𝟎𝟎 + 𝟏𝟎𝒕𝟐 )} 𝒅𝒕

𝟎 𝟎
𝟏𝟎 𝟏𝟎
∫ 𝝅(𝒕) 𝒅𝒕 = ∫ (𝟓𝟎𝟎𝟎 − 𝟐𝟎𝒕𝟐 − 𝟐𝟎𝟎𝟎 − 𝟏𝟎𝒕𝟐 ) 𝒅𝒕

𝟎 𝟎
𝟏𝟎 𝟏𝟎
∫ 𝝅(𝒕) 𝒅𝒕 = ∫ (𝟑𝟎𝟎𝟎 − 𝟑𝟎𝒕𝟐 ) 𝒅𝒕

𝟎 𝟎
𝟏𝟎 𝟏𝟎 𝟏𝟎
∫ 𝝅(𝒕) 𝒅𝒕 = 𝟑𝟎𝟎𝟎 ∫ (𝟏) 𝒅𝒕 − 𝟑𝟎 ∫ (𝒕𝟐 ) 𝒅𝒕

𝟎 𝟎 𝟎
______________________________________________________________________________
𝟏𝟎 𝟏𝟎
𝒕𝟑
∫ 𝝅(𝒕) 𝒅𝒕 = 𝟑𝟎𝟎𝟎{(𝒕)]𝟏𝟎
𝟎 }− 𝟑𝟎 ( )]
𝟑 𝟎
𝟎
𝟏𝟎
𝟏𝟎
∫ 𝝅(𝒕) 𝒅𝒕 = 𝟑𝟎𝟎𝟎{(𝒕)]𝟏𝟎 𝟑
𝟎 } − 𝟏𝟎(𝒕 )]𝟎
𝟎
𝟏𝟎
∫ 𝝅(𝒕) 𝒅𝒕 = 𝟑𝟎𝟎𝟎(𝟏𝟎 − 𝟎) − 𝟏𝟎{(𝟏𝟎)𝟑 − (𝟎)𝟑 }

𝟎
𝟏𝟎
∫ 𝝅(𝒕) 𝒅𝒕 = 𝟑𝟎𝟎𝟎𝟎 − 𝟏𝟎𝟎𝟎𝟎

𝟎
𝟏𝟎
∫ 𝝅(𝒕) 𝒅𝒕 = 𝟐𝟎𝟎𝟎𝟎
𝟎
Interpretation: Profit generated by the machine over its useful life (10 years) is $20,000.
______________________________________________________________________________
Lesson 14
TOPIC 056: FUND-RAISING NET EARNINGS ANALYSIS USING INTEGRALS
It is estimated that 𝒕 weeks from now, contributions in response to a fundraising campaign will be
coming in at the rate of 𝑹(𝒕) = 𝟔, 𝟓𝟑𝟕𝒆−𝟎.𝟑𝒕 dollars per week, while campaign expenses are
expected to accumulate at the constant rate of $𝟓𝟗𝟑 per week.
For how many weeks does the rate of revenue exceed the rate of cost?
What net earnings will be generated by the campaign during the period of time determined?
Solution: Visualizing two functions i.e. Rate of Revenue 𝑹′ (𝒕) and Rate of Cost 𝑪′ (𝒕).
To find 𝒕∗ mathematically, equate 𝑹′ (𝒕) and 𝑪′ (𝒕).

𝑹′ (𝒕) = 𝑪′ (𝒕)
𝟔𝟓𝟑𝟕𝒆−𝟎.𝟑𝒕 = 𝟓𝟗𝟑
𝟓𝟗𝟑
𝒆−𝟎.𝟑𝒕 =
𝟔𝟓𝟑𝟕
Introducing natural logarithm to eliminate exponential term (𝒆).
𝐥𝐧 𝒆−𝟎.𝟑𝒕 = 𝐥𝐧|0.0907|
−𝟎. 𝟑𝒕 = −𝟐. 𝟒
𝟐. 𝟒
𝒕=
𝟎. 𝟑
𝒕∗ = 𝟖
______________________________________________________________________________
Interpretation: For 𝟖 weeks, the rate of revenue exceeds the rate of cost. While the difference
between the net earnings during these first 8 weeks can be found as follows:
Net Earnings = Area under Revenue Curve – Area under Revenue Curve
Mathematically speaking,
{𝑵𝑬(𝒕) = 𝑹(𝒕) − 𝑪(𝒕)} over [𝟎, 𝟖].
𝟖
⊿𝑨𝑩𝑪 = ∫{𝑹(𝒕) − 𝑪(𝒕)} 𝒅𝒕

𝟎
𝟖
⊿𝑨𝑩𝑪 = ∫{(𝟔𝟓𝟑𝟕𝒆−𝟎.𝟑𝒕 ) − (𝟓𝟗𝟑)} 𝒅𝒕

𝟎
𝟖 𝟖
−𝟎.𝟑𝒕
⊿𝑨𝑩𝑪 = 𝟔𝟓𝟑𝟕 ∫(𝒆 ) 𝒅𝒕 − ∫(𝟓𝟗𝟑) 𝒅𝒕
𝟎 𝟎
𝟖
𝒆−𝟎.𝟑𝒕 𝑒 𝑓(𝑥)
⊿𝑨𝑩𝑪 = 𝟔𝟓𝟑𝟕 ( −𝟎.𝟑 ] ) − 𝟓𝟗𝟑𝒕]𝟖𝟎 [∵ ∫ 𝑒 𝑓(𝑥) 𝑑𝑥 = 𝑓′ (𝑥)]
𝟎
______________________________________________________________________________
𝟔𝟓𝟑𝟕 −𝟎.𝟑(𝟖)
⊿𝑨𝑩𝑪 = − {𝒆 − 𝒆−𝟎.𝟑(𝟎) } − 𝟓𝟗𝟑(𝟖 − 𝟎)
𝟎. 𝟑
𝟔𝟓𝟑𝟕 −𝟐.𝟒
⊿𝑨𝑩𝑪 = − {𝒆 − 𝒆𝟎 } − 𝟓𝟗𝟑(𝟖)
𝟎. 𝟑
⊿𝑨𝑩𝑪 = −𝟐𝟏𝟕𝟗𝟎{𝟎. 𝟎𝟗𝟎𝟕 − 𝟏} − 𝟒𝟕𝟒𝟒
⊿𝑨𝑩𝑪 = 𝟏𝟓𝟎𝟔𝟗. 𝟔𝟒𝟕
Interpretation: Net earnings in fund raising during first 𝟖 weeks shall be $𝟏𝟓𝟎𝟔𝟗. 𝟔𝟒𝟕.
TOPIC 057: RETIREMENT ANNUITY ANALYSIS USING INTEGRALS
Suppose money is being transferred continuously into an account over a time period 𝟎 ≤ 𝒕 ≤ 𝑻 at
a rate given by the function 𝒇(𝒕) and that the account earns interest at an annual rate r
compounded continuously. Then the future value FV of the income stream over the term 𝑻 is
given by the definite integral
𝑻
𝑭𝑽 = ∫ 𝒇(𝒕) 𝒆𝒓(𝑻−𝒕) 𝒅𝒕
𝟎
𝑻
𝑭𝑽 = 𝒆𝒓𝑻 ∫ 𝒇(𝒕) 𝒆𝒓𝑻 ∙ 𝒆−𝒓𝒕 𝒅𝒕

𝟎
𝑻
𝑭𝑽 = 𝒆𝒓𝑻 ∫ 𝒇(𝒕) 𝒆−𝒓𝒕 𝒅𝒕

𝟎
Economic Application: At age 25, Tom starts making annual deposits of $2,500 into an
Individual Retirement Account (IRA) that pays interest at an annual rate of 5% compounded
continuously. Assuming that his payments are made as a continuous income flow, how much
money will be in his account if he retires at age 60? At age 65?
Tom is undergoing a process of continuous compounding.
Annual rate of compounding = 𝟓% = 𝟎. 𝟎𝟓
Annual deposits in the account = $𝟐, 𝟓𝟎𝟎
Upper limit of time (T) is (60 − 25 =)𝟑𝟓
Lower limit of time is = 𝟎
Resorting to formula of continuous compounding:
𝑻
𝒓𝑻
𝑭𝑽 = 𝒆 ∫ 𝒇(𝒕) 𝒆−𝒓𝒕 𝒅𝒕
𝟎
𝟑𝟓
𝑭𝑽 = 𝒆(𝟎.𝟎𝟓)(𝟑𝟓) ∫ (𝟐𝟓𝟎𝟎) 𝒆−(𝟎.𝟎𝟓)𝒕 𝒅𝒕

𝟎
𝟑𝟓
𝑭𝑽 = 𝒆𝟏.𝟕𝟓 × 𝟐𝟓𝟎𝟎 ∫ (𝒆−𝟎.𝟎𝟓𝒕 ) 𝒅𝒕

𝟎
𝟑𝟓
𝟏.𝟕𝟓
𝒆−𝟎.𝟎𝟓𝒕
𝑭𝑽 = 𝒆 × 𝟐𝟓𝟎𝟎 {( )] }
−𝟎. 𝟎𝟓 𝟎
𝟐𝟓𝟎𝟎 −𝟎.𝟎𝟓𝒕 𝟑𝟓
𝑭𝑽 = 𝟓. 𝟕𝟓𝟒𝟔 × (𝒆 ]𝟎 )
−𝟎. 𝟎𝟓
𝑭𝑽 = −𝟐𝟖𝟕𝟕𝟑𝟎{𝒆−𝟎.𝟎𝟓(𝟑𝟓) − 𝒆−𝟎.𝟎𝟓(𝟎) }
𝑭𝑽 = −𝟐𝟖𝟕𝟕𝟑𝟎(𝒆−𝟏.𝟕𝟓 − 𝒆𝟎 )
______________________________________________________________________________
𝑭𝑽 = 𝟐𝟑𝟕𝟕𝟑𝟎. 𝟎𝟐
Interpretation: Tom, at age 𝟐𝟓, starts making annual deposits of $𝟐, 𝟓𝟎𝟎 into an Individual
Retirement Account (IRA) that pays interest at an annual rate of 5% compounded continuously.
At the age 𝟔𝟎, when he retires, $𝟐𝟑𝟕𝟕𝟑𝟎. 𝟎𝟐 will be in his account.
TOPIC 058: WAGE DIFFERENTIAL USING INTEGRALS
Assume that an individual has a choice of either going for work with only high school education
or going for work after obtaining an undergraduate degree in economics, business, or finance. If
the individual decides to go for work with only high school education, the salary (in dollars) over
the years will be given by the function 𝒈(𝒕) = 𝟖𝟎𝟎𝟎𝒆𝟎.𝟎𝟓𝒕 ; and if the individual decides to go for
work after obtaining a degree, the salary over the years will be given by the function 𝒇(𝒕) =
𝟏𝟐𝟎𝟎𝟎𝒆(𝟎.𝟎𝟓+𝟎.𝟎𝟐)𝒕 . If the individual works for 25 years, what will be the difference in total salary
earned over the entire period of work?
Solution:
Visualizing the two possible time paths of salary i.e. 𝒈(𝒕) and 𝒇(𝒕).
Finding the initial condition of the time paths, we introduce 𝒕 = 𝟎 in 𝒈(𝒙) and 𝒇(𝒙), respectively.
Initial condition with High School Initial condition with Undergraduate
Education Education
(𝟎.𝟎𝟓)𝒕
𝒈(𝒙) = 𝟖𝟎𝟎𝟎𝒆 𝒇(𝒙) = 𝟏𝟐𝟎𝟎𝟎𝒆(𝟎.𝟎𝟓+𝟎.𝟎𝟐)𝒕
𝒈(𝟎) = 𝟖𝟎𝟎𝟎𝒆(𝟎.𝟎𝟓)(𝟎) 𝒇(𝟎) = 𝟏𝟐𝟎𝟎𝟎𝒆(𝟎.𝟎𝟓+𝟎.𝟎𝟐)(𝟎)
𝟎
𝒈(𝟎) = 𝟖𝟎𝟎𝟎(𝒆 ) 𝒇(𝟎) = 𝟏𝟐𝟎𝟎𝟎(𝒆𝟎 )
𝒈(𝟎) = 𝟖𝟎𝟎𝟎(𝟏) 𝒇(𝟎) = 𝟏𝟐𝟎𝟎𝟎(𝟏)
𝒈(𝟎) = 𝟖𝟎𝟎𝟎 𝒇(𝟎) = 𝟏𝟐𝟎𝟎𝟎
______________________________________________________________________________
Interpretation
Starting job with High School Education Starting job with Undergraduate Education
will get the individual lower starting-wage will get the individual higher starting-wage
i.e. $𝟖, 𝟎𝟎𝟎 i.e. $𝟏𝟐, 𝟎𝟎𝟎
Determination of Time Path
Since the job is done for 𝟐𝟓 years, the lower and upper limits will be 𝟎 and 𝟐𝟓.
Applying the limits on both of the functions
𝟐𝟓 𝟐𝟓 𝟐𝟓 𝟐𝟓
(𝟎.𝟎𝟓)𝒕
∫ 𝒈(𝒕) 𝒅𝒕 = ∫ 𝟖𝟎𝟎𝟎𝒆 𝒅𝒕 ∫ 𝒇(𝒕) 𝒅𝒕 = ∫ 𝟏𝟐𝟎𝟎𝟎𝒆(𝟎.𝟎𝟓+𝟎.𝟎𝟐)𝒕 𝒅𝒕
𝟎 𝟎 𝟎 𝟎
𝟐𝟓 𝟐𝟓
(𝟎.𝟎𝟓)𝒕
= 𝟖𝟎𝟎𝟎 ∫ 𝒆 𝒅𝒕 = 𝟏𝟐𝟎𝟎𝟎 ∫ 𝒆(𝟎.𝟎𝟓+𝟎.𝟎𝟐)𝒕 𝒅𝒕
𝟎 𝟎
(𝟎.𝟎𝟓)𝒕 𝟐𝟓 𝟐𝟓
𝒆 𝒆(𝟎.𝟎𝟕)𝒕
= 𝟖𝟎𝟎𝟎 { ] } = 𝟏𝟐𝟎𝟎𝟎 { ] }
𝟎. 𝟎𝟓 𝟎 𝟎. 𝟎𝟕 𝟎
𝟖𝟎𝟎𝟎 (𝟎.𝟎𝟓)𝒕 𝟐𝟓 𝟏𝟐𝟎𝟎𝟎 (𝟎.𝟎𝟕)𝒕 𝟐𝟓
= {𝒆 ]𝟎 } = {𝒆 ]𝟎 }
𝟎. 𝟎𝟓 𝟎. 𝟎𝟕
(𝟎.𝟎𝟓)(𝟐𝟓)
= 𝟏𝟔𝟎𝟎𝟎𝟎{𝒆 − 𝒆(𝟎.𝟎𝟓)(𝟎) } = 𝟏𝟕𝟏𝟒𝟐𝟖. 𝟓𝟕{𝒆 (𝟎.𝟎𝟕)(𝟐𝟓)
− 𝒆(𝟎.𝟎𝟕)(𝟎) }
= 𝟏𝟔𝟎𝟎𝟎𝟎(𝒆𝟏.𝟐𝟓 − 𝒆𝟎 ) = 𝟏𝟕𝟏𝟒𝟐𝟖. 𝟓𝟕{𝒆𝟏.𝟕𝟓 − 𝒆𝟎 }
= 𝟏𝟔𝟎𝟎𝟎𝟎(𝟑. 𝟒𝟗 − 𝟏) = 𝟏𝟕𝟏𝟒𝟐𝟖. 𝟓𝟕{𝟓. 𝟕𝟓 − 𝟏}
= 𝟏𝟔𝟎𝟎𝟎𝟎(𝟐. 𝟒𝟗) = 𝟏𝟕𝟏𝟒𝟐𝟖. 𝟓𝟕{𝟒. 𝟕𝟓}
𝟐𝟓 𝟐𝟓
∫ 𝒈(𝒕) 𝒅𝒕 = 𝟑𝟗𝟖𝟒𝟎𝟎 ∫ 𝒇(𝒕) 𝒅𝒕 = 𝟖𝟏𝟒𝟐𝟖𝟓. 𝟕𝟏

𝟎 𝟎
Wage Differential = Wage Time Path with Undergraduate Education - Wage Time Path
with High School Education
______________________________________________________________________________
𝟐𝟓 𝟐𝟓
𝑾𝑫𝑮𝑯 = ∫ 𝒇(𝒕) 𝒅𝒕 − ∫ 𝒈(𝒕) 𝒅𝒕

𝟎 𝟎
𝑾𝑫𝑮𝑯 = 𝟖𝟏𝟒𝟐𝟖𝟓. 𝟕𝟏 − 𝟑𝟗𝟖𝟒𝟎𝟎
𝑾𝑫𝑮𝑯 = 𝟒𝟏𝟓𝟖𝟖𝟓. 𝟕𝟏
Interpretation: Obtaining Undergraduate Education before starting job, instead of starting
after High School Education, earns the individual additional 𝟒𝟏𝟓𝟖𝟖𝟓. 𝟕𝟏 dollars.
TOPIC 059: DEPLETION OF ENERGY RESOURCES ANALYSIS USING INTEGRALS
Oil is being pumped from an oil field 𝒕 years after its opening at the rate of 𝑷′ (𝒕) = 𝟏. 𝟑𝒆𝟎.𝟎𝟒𝒕 billion
barrels per year. The field has a reserve of 𝟐𝟎 billion barrels, and the price of oil holds steady at
$𝟏𝟏𝟐 per barrel.
a) Find 𝑷(𝒕), the amount of oil pumped from the field at time 𝒕. How much oil is pumped from the
field during the first 3 years of operation?
b) For how many years 𝑻 does the field operate before it runs dry?
Solution: Visualizing the rate of oil pumping 𝑷′ (𝒕) = 𝟏. 𝟑𝒆𝟎.𝟎𝟒𝒕 :
To find how much oil is pumped during first 𝟑 years, we apply lower and upper limits [𝟎, 𝟑].
𝟑 𝟑
∫ 𝑷′ (𝒕) 𝒅𝒕 = ∫(𝟏. 𝟑𝒆𝟎.𝟎𝟒𝒕 ) 𝒅𝒕

𝟎 𝟎
𝟑 𝟑
∫ 𝑷′ (𝒕) 𝒅𝒕 = 𝟏. 𝟑 ∫(𝒆𝟎.𝟎𝟒𝒕 ) 𝒅𝒕
𝟎 𝟎
𝟑
𝒆𝟎.𝟎𝟒𝒕
= 𝟏. 𝟑 {( )] }
𝟎. 𝟎𝟒 𝟎
𝟏. 𝟑 𝟎.𝟎𝟒(𝒕) 𝟑
= {𝒆 ]𝟎 }
𝟎. 𝟎𝟒
= 𝟑𝟐. 𝟓{𝒆𝟎.𝟎𝟒(𝟑) − 𝒆𝟎.𝟎𝟒(𝟎) }
______________________________________________________________________________
= 𝟑𝟐. 𝟓(𝒆𝟎.𝟏𝟐 − 𝒆𝟎 )
= 𝟑𝟐. 𝟓(𝟏. 𝟏𝟐𝟕 − 𝟏)
= 𝟑𝟐. 𝟓(𝟎. 𝟏𝟐𝟕)
𝟑
∫ 𝑷′ (𝒕) 𝒅𝒕 = 𝟒. 𝟏𝟑
𝟎
Interpretation: The amount extracted from the oil field during the first year is 𝟒. 𝟏𝟑 billion barrels.
To find the time when the field is used up all of its reservoir (𝟐𝟎 billion barrels) it holds, we
need 𝑷(𝒕) function.
𝑷(𝒕) = ∫ 𝑷′ (𝒕) 𝒅𝒕 = ∫(𝟏. 𝟑𝒆𝟎.𝟎𝟒𝒕 ) 𝒅𝒕
𝑷(𝒕) = 𝟏. 𝟑 ∫(𝒆𝟎.𝟎𝟒𝒕 ) 𝒅𝒕
𝒆𝟎.𝟎𝟒𝒕
𝑷(𝒕) = 𝟏. 𝟑 { + 𝒄′ }
𝟎. 𝟎𝟒
𝟏. 𝟑 𝟎.𝟎𝟒𝒕
𝑷(𝒕) = {𝒆 + 𝟏. 𝟑𝒄′ }
𝟎. 𝟎𝟒
𝑷(𝒕) = 𝟑𝟐. 𝟓𝒆𝟎.𝟎𝟒𝒕 + 𝟑𝟐. 𝟓𝒄′ [∵ 𝟏. 𝟑𝒄′ = 𝒄]
𝟎.𝟎𝟒𝒕 [∵ 𝟑𝟐. 𝟓𝒄′ = 𝒄]
𝑷(𝒕) = 𝟑𝟐. 𝟓𝒆 +𝒄
Substituting the initial condition (𝒕 = 𝟎), for getting rid of arbitrary constant 𝒄.
𝑷(𝟎) = 𝟑𝟐. 𝟓𝒆𝟎.𝟎𝟒(𝟎) + 𝒄
Since at the beginning of pumping process, no oil is pumped out, 𝑷(𝟎) = 𝟎
𝟎 = 𝟑𝟐. 𝟓𝒆𝟎 + 𝒄
𝟎 = 𝟑𝟐. 𝟓 + 𝒄
𝒄 = −𝟑𝟐. 𝟓
Substituting the value of 𝒄 in general time path 𝑷(𝒕) = 𝟑𝟐. 𝟓𝒆𝟎.𝟎𝟒𝒕 + 𝒄:
𝑷(𝒕) = 𝟑𝟐. 𝟓𝒆𝟎.𝟎𝟒𝒕 + (−𝟑𝟐. 𝟓)
𝑷(𝒕) = 𝟑𝟐. 𝟓𝒆𝟎.𝟎𝟒𝒕 − 𝟑𝟐. 𝟓
Now substituting the 20 billion barrels in place of oil pumping function 𝑷(𝒕).
∗
𝟐𝟎 = 𝟑𝟐. 𝟓𝒆𝟎.𝟎𝟒𝒕 − 𝟑𝟐. 𝟓
∗
𝟐𝟎 + 𝟑𝟐. 𝟓 = 𝟑𝟐. 𝟓𝒆𝟎.𝟎𝟒𝒕
∗
𝟑𝟐. 𝟓𝒆𝟎.𝟎𝟒𝒕 = 𝟓𝟐. 𝟓
∗ 𝟓𝟐. 𝟓
𝟑𝟐. 𝟓𝒆𝟎.𝟎𝟒𝒕 =
∗
𝟑𝟐. 𝟓
𝒆𝟎.𝟎𝟒𝒕 = 𝟏. 𝟔𝟏𝟓
∗
𝒍𝒏(𝒆𝟎.𝟎𝟒𝒕 ) = 𝒍𝒏(𝟏. 𝟔𝟏𝟓)
𝟎. 𝟎𝟒𝒕∗ = 𝟎. 𝟒𝟕𝟗
𝟎. 𝟒𝟕𝟗
𝒕∗ =
𝟎. 𝟎𝟒
𝒕∗ = 𝟏𝟏. 𝟗𝟕𝟓 ≅ 𝟏𝟐 𝒚𝒆𝒂𝒓𝒔
Interpretation: It will take approximately 𝟏𝟐 years to pump out the oil field till no more oil is left.
TOPIC 060: PRESENT VALUE OF GOLD MINE USING INTEGRALS

______________________________________________________________________________
Given the function of income earned from a gold mine is 𝑷(𝒕) = 𝟔𝟎𝟎𝒆−𝟎.𝟓𝒕. If money is discounted
at 𝟏𝟎% a year within a period of 𝟐𝟎 years, what is the present value of the gold mine?
𝒕
As already known, that: 𝝅 = ∫𝟎 𝑹(𝒕) 𝒆−𝒓𝒕 𝒅𝒕
Here 𝒕 = 𝟐𝟎, 𝑹(𝒕) = 𝑷(𝒕), 𝒓 = 𝟏𝟎% = 𝟎. 𝟏, 𝝅 = 𝑷𝑽𝑮𝑴 .
Therefore, substitution of values gives:
𝟐𝟎
𝑷𝑽𝑮𝑴 = ∫ (𝟔𝟎𝟎𝒆−𝟎.𝟓𝒕 ) 𝒆−𝟎.𝟏𝒕 𝒅𝒕
𝟎
𝟐𝟎
𝑷𝑽𝑮𝑴 = 𝟔𝟎𝟎 ∫ 𝒆−𝟎.𝟓𝒕 𝒆−𝟎.𝟏𝒕 𝒅𝒕
𝟎
𝟐𝟎
𝑷𝑽𝑮𝑴 = 𝟔𝟎𝟎 ∫ 𝒆−𝟎.𝟔𝒕 𝒅𝒕
𝟎
𝟐𝟎
𝒆−𝟎.𝟔𝒕
𝑷𝑽𝑮𝑴 = 𝟔𝟎𝟎 {( )] }
−𝟎. 𝟔 𝟎
𝟔𝟎𝟎 𝟐𝟎
𝑷𝑽𝑮𝑴 = {(𝒆−𝟎.𝟔𝒕 )]𝟎 }
−𝟎. 𝟔
𝑷𝑽𝑮𝑴 = −𝟏𝟎𝟎𝟎(𝒆−𝟎.𝟔(𝟐𝟎) − 𝒆−𝟎.𝟔(𝟎) )
𝑷𝑽𝑮𝑴 = −𝟏𝟎𝟎𝟎(𝒆−𝟏𝟐 − 𝒆𝟎 )
𝑷𝑽𝑮𝑴 = −𝟏𝟎𝟎𝟎{(𝟔. 𝟏𝟒 × 𝟏𝟎−𝟔 ) − 𝟏}
𝑷𝑽𝑮𝑴 = −𝟏𝟎𝟎𝟎(−𝟎. 𝟗𝟗𝟗𝟗𝟗𝟑𝟖𝟔)
𝑷𝑽𝑮𝑴 = 𝟗𝟗𝟗. 𝟗𝟗 ≅ 𝟏𝟎𝟎𝟎
Interpretation: The present value of the gold mine that mined for 𝟐𝟎 years, is $𝟏𝟎𝟎𝟎 thousands
of dollars i.e. ($𝟏𝟎, 𝟎𝟎, 𝟎𝟎𝟎).
______________________________________________________________________________
Lesson 15
TOPIC 061: LOTTERY PAYOUT ANALYSIS USING INTEGRALS
The winner of a state lottery is offered a choice of either receiving $𝟏𝟎 million now as a lump sum
or of receiving 𝑨 dollars a year for the next 𝟔 years as a continuous income stream. If the prevailing
annual interest rate is 𝟓% compounded continuously and the two payouts are worth the same,
what is 𝑨?
Solution: Winner of the lottery has two options:
Option – I Option – II
Receiving $𝟏𝟎 million now as a lump sum. ′𝑨′ dollars a year for the next 𝟔 years as a
continuous income stream (annual interest
rate is 𝟓%)
The unknown in the two options is ′𝑨′ : i.e. how much should be given to the individual for next
6 years so that he remains indifferent between two options.
Mathematically, we can use equation to represent the indifference of the individual between
two options. i.e.
$𝟏𝟎 million now = ‘𝑨’ annually for 6 years
𝟔
𝟏𝟎 = ∫𝟎 𝑨 𝒆−𝟎.𝟎𝟓𝒕 𝒅𝒕
𝟔
𝟏𝟎 = 𝑨 ∫𝟎 𝒆−𝟎.𝟎𝟓𝒕 𝒅𝒕
𝟔
𝒆−𝟎.𝟎𝟓𝒕
𝟏𝟎 = 𝑨 ( )]
−𝟎.𝟎𝟓 𝟎
𝑨 𝟔
𝟏𝟎 = −𝟎.𝟎𝟓 (𝒆−𝟎.𝟎𝟓𝒕 )]𝟎
−𝟎. 𝟎𝟓 × 𝟏𝟎 = 𝑨{𝒆−𝟎.𝟎𝟓(𝟔) − 𝒆−𝟎.𝟎𝟓(𝟎) }
−𝟎. 𝟓 = 𝑨(𝒆−𝟎.𝟑 − 𝒆𝟎 )
−𝟎. 𝟓 = 𝑨(𝟎. 𝟕𝟒𝟏 − 𝟏)
−𝟎. 𝟓 = 𝑨(−𝟎. 𝟐𝟓𝟗)
−𝟎. 𝟓
𝑨=
−𝟎. 𝟐𝟓𝟗
𝑨 = $𝟏. 𝟗𝟑𝟏 million
Interpretation: Present value analysis shows the lottery winner can be kept indifferent
between Option – I and Option – II, he can be given $𝟏. 𝟗𝟑𝟏 for next 6 years.
TOPIC 062: BALANCE OF TRADE ANALYSIS USING INTEGRALS
The government of a certain country estimates that 𝒕 years from now, imports will be increasing
at the rate 𝑰𝒕 and exports at the rate 𝑬𝒕 , both in billions of dollars per year, where 𝑰′ (𝒕) = 𝟏𝟐. 𝟓𝒆𝟎.𝟐𝒕
and 𝑬′ (𝒕) = 𝟏. 𝟕𝒕 + 𝟑.
The trade deficit is 𝑫(𝒕) = 𝑰(𝒕) − 𝑬(𝒕). By how much will the trade deficit for this country change
over the next 5 years? Will it increase or decrease during this time period?
Solution: Visualizing the given functions:
______________________________________________________________________________
We are given the rate of import and export functions i.e. 𝑰′ (𝒕) and 𝑬′ (𝒕). While in the trade deficit
𝑫(𝒕) formula needs the original functions i.e. 𝑰(𝒕) and 𝑬(𝒕).
Therefore, we integrate the two functions, respectively, w.r.t time (𝒕).

Moreover, we need to apply the lower and upper limits [𝟎, 𝟓]. As deficit is to be quantified for five
years.
𝑰′ (𝒕) = 𝟏𝟐. 𝟓𝒆𝟎.𝟐𝒕 𝑬′ (𝒕) = 𝟏. 𝟕𝒕 + 𝟑
𝟓 𝟓 𝟓 𝟓
∫ 𝑰′ (𝒕) 𝒅𝒕 = ∫ 𝟏𝟐. 𝟓𝒆𝟎.𝟐𝒕 𝒅𝒕 ∫ 𝑬′ (𝒕) 𝒅𝒕 = ∫ (𝟏. 𝟕𝒕 + 𝟑) 𝒅𝒕
𝟎 𝟎 𝟎 𝟎
𝟓 𝟓 𝟓 𝟓 𝟓
∫ 𝑰 ′ (𝒕)
𝒅𝒕 = 𝟏𝟐. 𝟓 ∫ 𝒆 𝟎.𝟐𝒕
𝒅𝒕 ∫𝟎 𝑬′ (𝒕) 𝒅𝒕 = 𝟏. 𝟕 ∫𝟎 (𝒕) 𝒅𝒕 + 𝟑 ∫𝟎 (𝟏) 𝒅𝒕
𝟓 𝟓
𝟎 𝟎 𝒕𝟐
𝟓 𝒆𝟎.𝟐𝒕
𝟓 ∫ 𝑬 ′ (𝒕)
𝒅𝒕 = 𝟏. 𝟕 {( )] } + 𝟑(𝒕]𝟓𝟎 )
∫𝟎 𝑰′ (𝒕) 𝒅𝒕 = 𝟏𝟐. 𝟓 {( 𝟎.𝟐 )] } 𝟎 𝟐 𝟎
𝟎
𝟓 ′ (𝟓)𝟐 −(𝟎)𝟐
𝟓 𝒆𝟎.𝟐(𝟓) −𝒆𝟎.𝟐(𝟎) ∫𝟎 𝑬 (𝒕) 𝒅𝒕 = 𝟏. 𝟕 ( 𝟐 ) + 𝟑(𝟓 − 𝟎)
∫𝟎 𝑰′ (𝒕) 𝒅𝒕 = 𝟏𝟐. 𝟓 { 𝟎.𝟐
}
𝟓 𝟐𝟓
𝟓 ∫𝟎 𝑬′ (𝒕) 𝒅𝒕 = 𝟏. 𝟕 { 𝟐 } + 𝟑(𝟓)
∫𝟎 𝑰′ (𝒕) 𝒅𝒕 = 𝟔𝟐. 𝟓(𝒆𝟏 − 𝒆𝟎 ) 𝟓
𝟓 ∫𝟎 𝑬′ (𝒕) 𝒅𝒕 = 𝟏. 𝟕(𝟏𝟐. 𝟓) + 𝟏𝟓
∫𝟎 𝑰′ (𝒕) 𝒅𝒕 = 𝟔𝟐. 𝟓(𝟐. 𝟕𝟏𝟖 − 𝟏) 𝟓
𝟓 ∫𝟎 𝑬′ (𝒕) 𝒅𝒕 = 𝟐𝟏. 𝟐𝟓 + 𝟏𝟓
∫𝟎 𝑰′ (𝒕) 𝒅𝒕 = 𝟔𝟐. 𝟓(𝟏. 𝟕𝟏𝟖) 𝟓
𝟓 𝑬(𝒕) = ∫𝟎 𝑬′ (𝒕) 𝒅𝒕 = 𝟑𝟔. 𝟐𝟓
𝑰(𝒕) = ∫𝟎 𝑰′ (𝒕) 𝒅𝒕 = 𝟏𝟎𝟕. 𝟑𝟕𝟓
______________________________________________________________________________
With both 𝑰(𝒕) and 𝑬(𝒕) available, we can find the value of deficit 𝑫(𝒕):
𝑫(𝒕) = 𝑰(𝒕) − 𝑬(𝒕)
𝟓 𝟓
𝑫(𝒕) = ∫ 𝑰′ (𝒕) 𝒅𝒕 − ∫ 𝑬′ (𝒕) 𝒅𝒕
𝟎 𝟎
= 𝟏𝟎𝟕. 𝟑𝟕𝟓 − 𝟑𝟔. 𝟐𝟓
𝑫(𝒕) = 𝟕𝟏. 𝟏𝟐𝟓
Interpretation: During the next 𝟓 years, the trade deficit 𝑫(𝒕) shall be 𝟕𝟏. 𝟏𝟐𝟓 billion dollars.
TOPIC 063: TICKET SALES ANALYSIS USING INTEGRALS
The promoters of a county fair estimate that 𝒕 hours after the gates open at 9:00 A.M. visitors will
be entering the fair at the rate of −𝟒(𝒕 + 𝟐)𝟑 + 𝟓𝟒(𝒕 + 𝟐)𝟐 people per hour. How many people will
enter the fair between 10:00 A.M. and noon?
Solution: Primarily one can assume that ‘number of tickets sold = visitors entering’.
Therefore, Rate of Ticket Sales 𝑻𝑺′ (𝒕) function is same as that of Rate of visitors entering 𝑽𝑬′ (𝒕)
function, i.e.
𝑽𝑬′ (𝒕) = 𝑻𝑺′ (𝒕) = −𝟒(𝒕 + 𝟐)𝟑 + 𝟓𝟒(𝒕 + 𝟐)𝟐
Visualizing the function:
______________________________________________________________________________
As mentioned, the Ticket Sales are to be found between 10:00 A.M. till noon.
Since county fair starts from 9:00 A.M., therefore, at 9:00 A.M., 𝒕 = 𝟎
Accordingly, 10:00 A.M., therefore, at 10:00 A.M., 𝒕 = 𝟏
At noon, 12:00 P.M., therefore, at 12:00 A.M., 𝒕 = 𝟑
Summary Table:
______________________________________________________________________________
9:00 A.M. ⇒ 𝑡 = 0
10:00 A.M. ⇒ 𝒕 = 𝟏
11:00 A.M. ⇒ 𝑡 = 2
12:00 A.M. ⇒ 𝒕 = 𝟑
Therefore, the lower and upper limits of the function are [𝟏, 𝟑].
𝟑 𝟑
∫ 𝑻𝑺′ (𝒕) 𝒅𝒕 = ∫ {−𝟒(𝒕 + 𝟐)𝟑 + 𝟓𝟒(𝒕 + 𝟐)𝟐 } 𝒅𝒕
𝟏 𝟏
𝟑 𝟑 𝟑
′ (𝒕)
∫ 𝑻𝑺 𝒅𝒕 = ∫ {−𝟒(𝒕 + 𝟐) } 𝒅𝒕 + ∫ {𝟓𝟒(𝒕 + 𝟐)𝟐 } 𝒅𝒕
𝟑
𝟏 𝟏 𝟏
𝟑 𝟑 𝟑
∫ 𝑻𝑺′ (𝒕) 𝒅𝒕 = −𝟒 ∫ (𝒕 + 𝟐)𝟑 𝒅𝒕 + 𝟓𝟒 ∫ (𝒕 + 𝟐)𝟐 𝒅𝒕
𝟏 𝟏 𝟏
𝟑 𝟑 𝟑
(𝒕 + 𝟐)𝟒 (𝒕 + 𝟐)𝟑
∫ 𝑻𝑺′ (𝒕) 𝒅𝒕 = −𝟒 { ] } + 𝟓𝟒 { ] }
𝟏 𝟒 𝟏
𝟑 𝟏
𝟑
𝟒 𝟓𝟒 𝟑
∫ 𝑻𝑺′ (𝒕) 𝒅𝒕 = − {(𝒕 + 𝟐)𝟒 ]𝟑𝟏 } + {(𝒕 + 𝟐)𝟑 ]𝟏 }
𝟏 𝟒 𝟑
𝟑
∫ 𝑻𝑺′ (𝒕) 𝒅𝒕 = −{(𝟑 + 𝟐)𝟒 − (𝟏 + 𝟐)𝟒 } + 𝟏𝟖{(𝟑 + 𝟐)𝟑 − (𝟏 + 𝟐)𝟑 }
𝟏
𝟑
∫ 𝑻𝑺′ (𝒕) 𝒅𝒕 = −{(𝟓)𝟒 − (𝟑)𝟒 } + 𝟏𝟖{(𝟓)𝟑 − (𝟑)𝟑 }
𝟏
𝟑
∫ 𝑻𝑺′ (𝒕) 𝒅𝒕 = −(𝟔𝟐𝟓 − 𝟖𝟏) + 𝟏𝟖(𝟏𝟐𝟓 − 𝟐𝟕)
𝟏
𝟑
∫ 𝑻𝑺′ (𝒕) 𝒅𝒕 = −(𝟓𝟒𝟒) + 𝟏𝟖(𝟗𝟖)
𝟏
𝟑
∫ 𝑻𝑺′ (𝒕) 𝒅𝒕 = −𝟓𝟒𝟒 + 𝟏𝟕𝟔𝟒
𝟏
𝟑
∫ 𝑻𝑺′ (𝒕) 𝒅𝒕 = 𝟏𝟐𝟐𝟎
𝟏
Interpretation: Hence, between 10:00 A.M. till noon (12:00 A.M.), the Ticket Sales will be 1220
in number. In other words, 1220 people will be entering the country fair.
TOPIC 064: DEMAND ANALYSIS OF OIL USING INTEGRALS
It is estimated that 𝒕 years from the beginning of the year 2005, the demand for oil in a certain
country will be changing at the rate of 𝑫′ (𝒕) = (𝟏 + 𝟐𝒕)−𝟏 billion barrels per year. Will more oil be
consumed (demanded) during 2006 or during 2009? How much more?
Solution: Rate of change of demand in certain country for Oil is 𝑫′ (𝒕) = (𝟏 + 𝟐𝒕)−𝟏 .
Analysis of demand for oil starts from year 2005, therefore, during 2005 A.M., 𝒕 = 𝟎 to 𝒕 = 𝟏.
Demand for oil during year 2006, will be; 𝒕 = 𝟏 to 𝒕 = 𝟐.
Summary Table:
Demand for oil during 2005 ⇒ 𝒕 = 𝟎 to 𝒕 = 𝟏
Demand for oil during 2006 ⇒ 𝒕 = 𝟏 to 𝒕 = 𝟐
Demand for oil during 2007 ⇒ 𝒕 = 𝟐 to 𝒕 = 𝟑
Demand for oil during 2008 ⇒ 𝒕 = 𝟑 to 𝒕 = 𝟒
Demand for oil during 2009 ⇒ 𝒕 = 𝟒 to 𝒕 = 𝟓
Therefore, the lower and upper limits of the function, for year 2006 are [𝟏, 𝟐].
______________________________________________________________________________
𝟐 𝟐
∫ 𝑫′ (𝒕) 𝒅𝒕 = ∫ {(𝟏 + 𝟐𝒕)−𝟏 } 𝒅𝒕
𝟏 𝟏
𝟐 𝟐
𝟏
∫ 𝑫′ (𝒕) 𝒅𝒕 = ∫ ( ) 𝒅𝒕
𝟏 𝟏 𝟏 + 𝟐𝒕
𝒇′ (𝒙)
Setting stage for application of: ∫ { 𝒇(𝒙) } 𝒅𝒙 = 𝐥𝐧|𝒇(𝒙)|.
As derivative of (𝟏 + 𝟐𝒕) is 𝟐, we can multiply and divide by 𝟐 .
𝟐
′ (𝒕)
𝟐 𝟐 𝟏
∫ 𝑫 𝒅𝒕 = ∫ ( ) 𝒅𝒕
𝟏 𝟐 𝟏 𝟏 + 𝟐𝒕
𝟐
𝟏 𝟐 𝟐
∫ 𝑫′ (𝒕) 𝒅𝒕 = ∫ ( ) 𝒅𝒕
𝟏 𝟐 𝟏 𝟏 + 𝟐𝒕
𝒇′ (𝒙)
Applying the rule: ∫ {
𝒇(𝒙)
} 𝒅𝒙 = 𝐥𝐧|𝒇(𝒙)|
𝟐
𝟏
∫ 𝑫′ (𝒕) 𝒅𝒕 = {𝒍𝒏|𝟏 + 𝟐𝒕|]𝟐𝟏 }
𝟏 𝟐
𝟐
𝟏
∫ 𝑫′ (𝒕) 𝒅𝒕 = (𝒍𝒏|𝟏 + 𝟐(𝟐)| − 𝒍𝒏|𝟏 + 𝟐(𝟏)|)
𝟏 𝟐
𝟐
𝟏
∫ 𝑫′ (𝒕) 𝒅𝒕 = (𝒍𝒏|𝟓| − 𝒍𝒏|𝟑|)
𝟏 𝟐
𝟐
𝑫(𝟐𝟎𝟎𝟔) = ∫𝟏 𝑫′ (𝒕) 𝒅𝒕 = 𝟎. 𝟐𝟓𝟓𝟒 billion barrels approximately.
Interpretation: The Demand for oil in a certain country during 2006 shall be 𝟎. 𝟐𝟓𝟓𝟒 ×
𝟏, 𝟎𝟎𝟎, 𝟎𝟎𝟎, 𝟎𝟎𝟎 = 𝟐, 𝟓𝟓, 𝟒𝟎𝟎, 𝟎𝟎𝟎 barrels.
Similarly, for the year 2009 same analysis can be done. However, the lower and upper limits will
be [𝟒, 𝟓]:
𝟐 𝟐
∫ 𝑫′ (𝒕) 𝒅𝒕 = ∫ {(𝟏 + 𝟐𝒕)−𝟏 } 𝒅𝒕
𝟏 𝟏
𝟐 𝟐
′ (𝒕)
𝟏
∫ 𝑫 𝒅𝒕 = ∫ ( ) 𝒅𝒕
𝟏 𝟏 𝟏 + 𝟐𝒕
𝒇′ (𝒙)
Setting stage for application of: ∫ { 𝒇(𝒙) } 𝒅𝒙 = 𝐥𝐧|𝒇(𝒙)|.
Borrowing the solution of integral from above, as the function is the same.
𝟓
𝟏 𝟓 𝟐
∫ 𝑫′ (𝒕) 𝒅𝒕 = ∫ ( ) 𝒅𝒕
𝟒 𝟐 𝟒 𝟏 + 𝟐𝒕
𝒇′ (𝒙)
Applying the rule: ∫ { 𝒇(𝒙) } 𝒅𝒙 = 𝐥𝐧|𝒇(𝒙)|
𝟓
𝟏
∫ 𝑫′ (𝒕) 𝒅𝒕 = {𝒍𝒏|𝟏 + 𝟐𝒕|]𝟓𝟒 }
𝟒 𝟐
𝟓
𝟏
∫ 𝑫′ (𝒕) 𝒅𝒕 = (𝒍𝒏|𝟏 + 𝟐(𝟓)| − 𝒍𝒏|𝟏 + 𝟐(𝟒)|)
𝟒 𝟐
𝟓
𝟏
∫ 𝑫′ (𝒕) 𝒅𝒕 = (𝒍𝒏|𝟏𝟏| − 𝒍𝒏|𝟗|)
𝟒 𝟐
𝟓
𝑫(𝟐𝟎𝟎𝟗) = ∫𝟒 𝑫′ (𝒕) 𝒅𝒕 = 𝟎. 𝟏𝟎𝟎𝟑 billion barrels approximately.
Interpretation: The Demand for oil in a certain country during 2006 shall be 𝟎. 𝟏𝟎𝟎𝟑 ×
𝟏, 𝟎𝟎𝟎, 𝟎𝟎𝟎, 𝟎𝟎𝟎 = 𝟏, 𝟎𝟎, 𝟑𝟎𝟎, 𝟎𝟎𝟎 barrels.
Comparative Analysis:
______________________________________________________________________________
𝑫(𝟐𝟎𝟎𝟔) = 𝟐, 𝟓𝟓, 𝟒𝟎𝟎, 𝟎𝟎𝟎 > 𝑫(𝟐𝟎𝟎𝟗) = 𝟏, 𝟎𝟎, 𝟑𝟎𝟎, 𝟎𝟎𝟎

Demand for Oil in the country under consideration is higher during 𝟐𝟎𝟎𝟔 as compared to that in
𝟐𝟎𝟎𝟗.
TOPIC 065: MARGINAL TO TOTAL UTILITY USING INTEGRALS
𝟑𝒙𝟐 +𝟔√𝒙+𝟐𝟎
If the marginal utility of a consumer is 𝑴𝑼𝒙 = 𝒙
, where 𝒙 is the number of units consumed
of a good, find the increase in his total utility when 𝒙 goes up from 𝟏𝟔 to 𝟏𝟎𝟎 units.
𝟑𝒙𝟐 +𝟔√𝒙+𝟐𝟎
Solution: Given the marginal utility function: 𝑴𝑼𝒙 = 𝒙
.
Additional total utility from 16th to 100th unit is to be found.

𝟏𝟎𝟎 𝟏𝟎𝟎 𝟑𝒙𝟐 +𝟔√𝒙+𝟐𝟎
Mathematically, it will be: ∫𝟏𝟔 𝑻𝑼(𝒙) 𝒅𝒙 = ∫𝟏𝟔 ( 𝒙
) 𝒅𝒙
𝟏𝟎𝟎
𝟏𝟎𝟎
𝟑𝒙𝟐 𝟏𝟎𝟎
𝟔√𝒙 𝟏𝟎𝟎
𝟐𝟎
∫ 𝑻𝑼(𝒙) 𝒅𝒙 = ∫ ( ) 𝒅𝒙 + ∫ ( ) 𝒅𝒙 + ∫ ( ) 𝒅𝒙
𝟏𝟔 𝒙 𝟏𝟔 𝒙 𝟏𝟔 𝒙
𝟏𝟔
𝟏𝟎𝟎 𝟏
𝟏𝟎𝟎 𝟏𝟎𝟎 (𝒙)𝟐 𝟏𝟎𝟎
𝟏
∫ 𝑻𝑼(𝒙) 𝒅𝒙 = 𝟑 ∫ (𝒙) 𝒅𝒙 + 𝟔 ∫ { } 𝒅𝒙 + 𝟐𝟎 ∫ ( ) 𝒅𝒙
𝟏𝟔 𝟏𝟔 𝒙 𝟏𝟔 𝒙
𝟏𝟔
𝟏𝟎𝟎
𝟏𝟎𝟎 𝟏𝟎𝟎 𝟏𝟎𝟎
𝟏 𝟏
∫ 𝑻𝑼(𝒙) 𝒅𝒙 = 𝟑 ∫ (𝒙) 𝒅𝒙 + 𝟔 ∫ (𝒙−𝟐 ) 𝒅𝒙 + 𝟐𝟎 ∫ ( ) 𝒅𝒙
𝟏𝟔 𝟏𝟔 𝟏𝟔 𝒙
𝟏𝟔
______________________________________________________________________________
𝟏𝟎𝟎 𝟏 𝟏𝟎𝟎
𝟏𝟎𝟎
𝒙𝟐 𝒙𝟐
∫ 𝑻𝑼(𝒙) 𝒅𝒙 = 𝟑 {( )] } + 𝟔 {( )] } + 𝟐𝟎{(𝒍𝒏|𝒙|)]𝟏𝟎𝟎 𝟏𝟔 }
𝟐 𝟏𝟔 𝟏
𝟏𝟔 𝟐 𝟏𝟔
(𝟏𝟎𝟎)𝟐 (𝟏𝟔)𝟐 𝟏 𝟏
= 𝟑{ − } + 𝟏𝟐 {(𝟏𝟎𝟎)𝟐 − (𝟏𝟔)𝟐 } + 𝟐𝟎(𝒍𝒏|𝟏𝟎𝟎| − 𝒍𝒏|𝟏𝟔|)
𝟐 𝟐
= 𝟑(𝟓𝟎𝟎𝟎 − 𝟏𝟐𝟖) + 𝟏𝟐(𝟏𝟎 − 𝟒) + 𝟐𝟎(𝟒. 𝟔𝟎𝟓𝟐 − 𝟐. 𝟕𝟕𝟐𝟔)
= 𝟑(𝟒𝟖𝟕𝟐) + 𝟏𝟐(𝟔) + 𝟐𝟎(𝟏. 𝟖𝟑𝟐𝟔)
𝟏𝟎𝟎
∫ 𝑻𝑼(𝒙) 𝒅𝒙 = 𝟏𝟒𝟔𝟏𝟔 + 𝟕𝟐 + 𝟑𝟔. 𝟔𝟓𝟐

𝟏𝟔
𝟏𝟎𝟎
∫ 𝑻𝑼(𝒙) 𝒅𝒙 = 𝟏𝟒𝟕𝟐𝟒. 𝟔𝟓𝟐

𝟏𝟔
Interpretation: The total utility of the good from 16th to 100th unit is equal to 𝟏𝟒𝟕𝟐𝟒. 𝟔𝟓𝟐.
______________________________________________________________________________
Lesson 16
TOPIC 066: REAL ESTATE ANALYSIS USING INTEGRALS
In a certain community the fraction of the homes placed on the market that remain unsold for at
least 𝒕 weeks is approximately 𝒇(𝒕) = 𝒆−𝟎.𝟐𝒕 . If 𝟐𝟎𝟎 homes are currently on the market the function
for existing unsold homes will be 𝒈(𝒕) = 𝟐𝟎𝟎𝒆−𝟎.𝟐𝒕 . Now, if additional homes are placed on the
market at the rate of 𝟖 per week, approximately, the function for unsold new houses will be 𝒉(𝒕) =
𝟖𝒆−𝟎.𝟐(𝟏𝟎−𝒕). How many homes will be on the market 10 weeks from now?
Solution: The total number of unsold homes will be sum of existing and new ones, i.e.:
𝐓𝐨𝐭𝐚𝐥 𝐔𝐧𝐬𝐨𝐥𝐝 = 𝐩𝐫𝐞𝐞𝐱𝐢𝐬𝐭𝐢𝐧𝐠 𝐮𝐧𝐬𝐨𝐥𝐝 + 𝐍𝐞𝐰 𝐮𝐧𝐬𝐨𝐥𝐝
After 𝟏𝟎 years, the existing unsold houses: 𝒈(𝟏𝟎) = 𝟐𝟎𝟎𝒆−𝟎.𝟐(𝟏𝟎)
Caveat: 𝒈(𝒕) function didn’t use limits as homes are pre-existing and no integration over time is
needed.
𝟏𝟎 𝟏𝟎
After 𝟏𝟎 years, the new unsold houses: ∫𝟎 𝒉(𝒕) 𝒅𝒕 = ∫𝟎 {𝟖𝒆−𝟎.𝟐(𝟏𝟎−𝒕) } 𝒅𝒕
Caveat: In case of 𝒉(𝒕), integration over time is needed as the new homes are being added on
weekly basis (as a stream).
Resuming the function:
𝐓𝐨𝐭𝐚𝐥 𝐔𝐧𝐬𝐨𝐥𝐝 = 𝐩𝐫𝐞𝐞𝐱𝐢𝐬𝐭𝐢𝐧𝐠 𝐮𝐧𝐬𝐨𝐥𝐝 + 𝐍𝐞𝐰 𝐮𝐧𝐬𝐨𝐥𝐝
𝟏𝟎
𝑼(𝒕) = 𝒈(𝟏𝟎) + ∫ 𝒉(𝒕) 𝒅𝒕

𝟎
𝟏𝟎
𝑼(𝒕) = 𝟐𝟎𝟎𝒆−𝟎.𝟐(𝟏𝟎) + ∫ {𝟖𝒆−𝟎.𝟐(𝟏𝟎−𝒕) } 𝒅𝒕

𝟎
𝟏𝟎
𝑼(𝒕) = 𝟐𝟎𝟎𝒆−𝟐 + 𝟖 ∫ (𝒆−𝟐+𝟎.𝟐𝒕 ) 𝒅𝒕

𝟎
𝟏𝟎
−𝟐
𝒆−𝟐+𝟎.𝟐𝒕
𝑼(𝒕) = 𝟐𝟎𝟎𝒆 + 𝟖 ( )]
𝟎. 𝟐 𝟎
𝟖 𝟏𝟎
𝑼(𝒕) = 𝟐𝟎𝟎𝒆−𝟐 + (𝒆−𝟐+𝟎.𝟐𝒕 )]𝟎
𝟎. 𝟐
𝑼(𝒕) = 𝟐𝟎𝟎𝒆−𝟐 + 𝟒𝟎{𝒆−𝟐+𝟎.𝟐(𝟏𝟎) − 𝒆−𝟐+𝟎.𝟐(𝟎) }
𝑼(𝒕) = 𝟐𝟎𝟎𝒆−𝟐 + 𝟒𝟎(𝒆𝟎 − 𝒆−𝟐 )
𝑼(𝒕) = 𝟐𝟎𝟎(𝟎. 𝟏𝟑𝟓) + 𝟒𝟎(𝟏 − 𝟎. 𝟏𝟑𝟓)
𝑼(𝒕) = 𝟐𝟕 + 𝟒𝟎(𝟎. 𝟖𝟔𝟓)
𝑼(𝒕)
⏟ = 𝟐𝟕
⏟ +⏟𝟑𝟒. 𝟔
𝑻𝒐𝒕𝒂𝒍 𝒑𝒓𝒆𝒆𝒙𝒊𝒔𝒕𝒊𝒏𝒈 𝑵𝒆𝒘
𝒖𝒏𝒔𝒐𝒍𝒅 𝒖𝒏𝒔𝒐𝒍𝒅 𝒖𝒏𝒔𝒐𝒍𝒅
𝑼(𝒕) = 𝟔𝟏. 𝟔
𝑼(𝒕) = 𝟔𝟐 unsold homes approximately.
Interpretation: If 𝟐𝟎𝟎 homes are currently on the market the function for existing unsold homes,
𝟐𝟕 will be left unsold after 𝟏𝟎 weeks. If additional homes are placed on the market at the rate of
8 per week, approximately 𝟑𝟒. 𝟔 homes will be left unsold after 𝟏𝟎 weeks. Consequently,
(𝟐𝟕 + 𝟑𝟒. 𝟔 = 𝟔𝟐) will be left unsold on the whole at the end of 𝟏𝟎 weeks.
______________________________________________________________________________
TOPIC 067: DYNAMICS OF LABOR COST USING INTEGRALS
The marginal labor cost function is given by the equation 𝑴𝑳𝑪 = 𝟑 + 𝟒𝑳.
(a) Find the equation of the total labor cost function if total labor cost is zero when 𝑳 = 𝟎.
(b) Calculate the cost of employing successive labor units from 𝑳 = 𝟏 to 𝑳 = 𝟕.
Solution: Initial condition of Marginal Labor Cost (𝑴𝑳𝑪) can be found at 𝑳 = 𝟎.
𝑴𝑳𝑪(𝑳) = 𝟑 + 𝟒𝑳
𝑴𝑳𝑪(𝟎) = 𝟑 + 𝟒(𝟎)
𝑴𝑳𝑪(𝟎) = 𝟑
Successive labor cost of employing 1 to 7 units of labor is found using the definite integral with
lower and upper limits [𝟏, 𝟕].
𝟕 𝟕
∫ 𝑴𝑳𝑪(𝑳) 𝒅𝑳 = ∫(𝟑 + 𝟒𝑳) 𝒅𝑳

𝟏 𝟏
𝟕 𝟕 𝟕
∫ 𝑴𝑳𝑪(𝑳) 𝒅𝑳 = 𝟑 ∫(𝟏) 𝒅𝑳 + 𝟒 ∫(𝑳) 𝒅𝑳

𝟏 𝟏 𝟏
𝟕 𝟕
𝑳𝟐
∫ 𝑴𝑳𝑪(𝑳) 𝒅𝑳 = 𝟑(𝑳)]𝟕𝟏 + 𝟒 ( )]
𝟐 𝟏
𝟏
𝟕
𝟕
∫ 𝑴𝑳𝑪(𝑳) 𝒅𝑳 = 𝟑(𝑳)]𝟕𝟏 + 𝟐(𝑳𝟐 )]𝟏
𝟏
𝟕
∫ 𝑴𝑳𝑪(𝑳) 𝒅𝑳 = 𝟑(𝟕 − 𝟏) + 𝟐{(𝟕)𝟐 − (𝟏)𝟐 }

𝟏
𝟕
∫ 𝑴𝑳𝑪(𝑳) 𝒅𝑳 = 𝟏𝟖 + 𝟗𝟔
𝟏
𝟕
∫ 𝑴𝑳𝑪(𝑳) 𝒅𝑳 = 𝟏𝟏𝟒
𝟏
______________________________________________________________________________
Interpretation: The Marginal labor cost of higher first 7 labor sums to 𝟏𝟏𝟒.
TOPIC 068: DEPRECIATION ANALYSIS USING INTEGRALS
The resale value of a certain industrial machine decreases over a 𝟏𝟎-year period at a rate that
changes with time. When the machine is 𝒙 years old, the rate at which its value is changing is
𝟐𝟐𝟎(𝒙 − 𝟏𝟎) dollars per year. By how much does the machine depreciate during the second year?
Solution: Let 𝑽′ (𝒙) be the rate of change of value of machine, then:
𝑽′ (𝒙) = 𝟐𝟐𝟎(𝒙 − 𝟏𝟎)
______________________________________________________________________________
𝟐 𝟐
𝑽(𝟐) − 𝑽(𝟏) = ∫ 𝑽′ (𝒙) 𝒅𝒙 = ∫{𝟐𝟐𝟎(𝒙 − 𝟏𝟎)} 𝒅𝒙

𝟏 𝟏
𝟐 𝟐 𝟐
∫ 𝑽′ (𝒙) 𝒅𝒙 = 𝟐𝟐𝟎 {∫ 𝒙 𝒅𝒙 − 𝟏𝟎 ∫(𝟏) 𝒅𝒙}

𝟏 𝟏 𝟏
𝟐 𝟐
𝟐
𝒙
∫ 𝑽′ (𝒙) 𝒅𝒙 = 𝟐𝟐𝟎 [{( )] } − 𝟏𝟎{(𝒙)]𝟐𝟏 }]
𝟐 𝟏
𝟏
𝟐
(𝟐)𝟐 (𝟏)𝟐
∫ 𝑽′ (𝒙) 𝒅𝒙 = 𝟐𝟐𝟎 [{ − } − 𝟏𝟎(𝟐 − 𝟏)]
𝟐 𝟐
𝟏
𝟐
𝟑
∫ 𝑽′ (𝒙) 𝒅𝒙 = 𝟐𝟐𝟎 ( − 𝟏𝟎)
𝟐
𝟏
𝟐
∫ 𝑽′ (𝒙) 𝒅𝒙 = 𝟐𝟐𝟎(−𝟖. 𝟓)
𝟏
𝟐
∫ 𝑽′ (𝒙) 𝒅𝒙 = −𝟏𝟖𝟕𝟎
𝟏
Interpretation: Change in value of machine is −𝟏𝟖𝟕𝟎, which means a decline of $𝟏𝟖𝟕𝟎 in the
value of machine. In other words, the machine depreciates by $𝟏𝟖𝟕𝟎 during the 2nd year.
TOPIC 069: COST EFFECTIVENESS DECISION USING DIFFERENTIAL EQUATIONS
A hospital is considering the installation of two alternative heating systems. The rate of increase
in costs for each system is given by the equations
𝒅𝑪𝟏 𝒅𝑪
𝒅𝒕
= 𝟎. 𝟎𝟏𝒕 and 𝒅𝒕𝟐 = 𝒕𝟎.𝟐𝟓
(a) Derive expressions for the total cost of each system at any time 𝒕 given that cost = £𝟓𝟎𝟎𝟎𝟎 at
𝒕 = 𝟎.
(b) Which system is more cost effective in the next 𝟏𝟎 years.
𝒅𝑪 𝒅𝑪
Solution: Rate of increase in costs for each heating system are 𝒅𝒕𝟏 = 𝟎. 𝟎𝟏𝒕 and 𝒅𝒕𝟐 = 𝒕𝟎.𝟐𝟓 ,
respectively. Lower and upper limit can be [𝟎, 𝟏𝟎]
______________________________________________________________________________
𝒅𝑪𝟏 𝒅𝑪𝟐
= 𝟎. 𝟎𝟏𝒕 = 𝒕𝟎.𝟐𝟓
𝒅𝒕 𝒅𝒕
Definite Integral implies:
𝟏𝟎 𝟏𝟎 𝟏𝟎 𝟏𝟎
𝒅𝑪𝟏 𝒅𝑪𝟐
∫( ) 𝒅𝒕 = ∫ (𝟎. 𝟎𝟏𝒕) 𝒅𝒕 ∫( ) 𝒅𝒕 = ∫ (𝒕𝟎.𝟐𝟓 ) 𝒅𝒕
𝒅𝒕 𝒅𝒕
𝟎 𝟎 𝟎 𝟎
𝟏𝟎 𝟎.𝟐𝟓+𝟏 𝟏𝟎
𝒕
𝑪𝟏 (𝒕) = 𝟎. 𝟎𝟏 ∫ (𝒕) 𝒅𝒕 𝑪𝟐 (𝒕) = {( )] }
𝟎. 𝟐𝟓 + 𝟏 𝟎
𝟎 𝟏𝟎
𝒕𝟐
𝟏𝟎 𝒕𝟏.𝟐𝟓
𝑪𝟏 (𝒕) = 𝟎. 𝟎𝟏 {( )] } 𝑪𝟐 (𝒕) = {( )] }
𝟐 𝟎 𝟏. 𝟐𝟓 𝟎
𝟎. 𝟎𝟏 𝟐 𝟏𝟎 (𝟏𝟎)𝟏.𝟐𝟓 − (𝟎)𝟏.𝟐𝟓
𝑪𝟏 (𝒕) = {(𝒕 )]𝟎 } 𝑪𝟐 (𝒕) =
𝟐 𝟏. 𝟐𝟓
𝑪𝟏 (𝒕) = 𝟎. 𝟎𝟎𝟓{(𝟏𝟎)𝟐 − (𝟎)𝟐 } 𝟏𝟕. 𝟕𝟖𝟐𝟖 − 𝟎
𝑪𝟐 (𝒕) =
𝑪𝟏 (𝒕) = 𝟎. 𝟎𝟎𝟓(𝟏𝟎𝟎) 𝟏. 𝟐𝟓
𝑪𝟏 (𝒕) = 𝟎. 𝟓 𝑪𝟐 (𝒕) = 𝟏𝟒. 𝟐𝟐𝟔
Clearly 𝑪𝟏 (𝒕) < 𝑪𝟐 (𝒕) therefore, 1st heating system is more cost effective. Same is visible in
the graph.
______________________________________________________________________________
Lesson 17
DOMAR GROWTH MODEL
TOPIC 070: DOMAR GROWTH MODEL: FRAMEWORK
PREMISES:
Stimulus: A change in the rate of investment will affect both aggregate demand and the
productive capability of an economy: (∆𝜿 ← ∆𝑰 → ∆𝑨𝑫)
Here 𝜿 is the Productive Capacity and 𝑰 & 𝑨𝑫 are investment and aggregate demand as usual.
Purpose: A Time path of investment 𝑰(𝒕) is needed along which an economy could grow while
maintaining full utilization of its productive capacity (𝜿).
(∆𝑰) Change in Investment causes dual change:
∆𝒀 (Change in Income/Aggregate ∆𝜿 (Change in Productive Capacity)
Demand)
Keynes Investment Multiplier Theory ‘Productive capacity per unit of capital’ is: 𝜿 ≡
𝑲
suggests that: 𝝆.
𝒅𝒀 𝟏
=𝑲= Here 𝜿 is the Productive Capacity and 𝑲 is
𝒅𝑰 𝒔 capital. Numerically speaking Productive
OR
Capacity-Capital ratio:
𝒅𝒀 𝟏
=( ) 𝜿 𝟓
𝒅𝑰 𝒔 𝝆= =𝟓=
Rearranging for 𝒅𝒀: 𝑲 𝟏
𝟏 Each unit of capital is capable of producing 5
𝒅𝒀
⏟ = ( ) × 𝒅𝑰
⏟ units of output.
𝑪𝒉𝒂𝒏𝒈𝒆 𝒊𝒏
⏟ 𝒔 𝑪𝒉𝒂𝒏𝒈𝒆 𝒊𝒏
𝑰𝒏𝒄𝒐𝒎𝒆 𝑴𝒖𝒍𝒕𝒊𝒑𝒍𝒊𝒆𝒓 𝑰𝒏𝒗𝒆𝒔𝒕𝒎𝒆𝒏𝒕
𝝆 is assumed to be constant as technology is
For instance: assumed to be constant.
𝟓𝟎𝟎
⏟ = ⏟
𝟓 × 𝟏𝟎𝟎⏟ Rewriting Productive Capacity-Capital ratio
𝜿
𝑪𝒉𝒂𝒏𝒈𝒆 𝒊𝒏 𝑴𝒖𝒍𝒕𝒊𝒑𝒍𝒊𝒆𝒓 𝑪𝒉𝒂𝒏𝒈𝒆 𝒊𝒏 from, 𝝆 = 𝑲:
𝑰𝒏𝒄𝒐𝒎𝒆 𝑰𝒏𝒗𝒆𝒔𝒕𝒎𝒆𝒏𝒕
Dynamize the changes in national 𝜿 = 𝝆𝑲
income (𝒅𝒀) and investment (𝒅𝑰): For dynamic analysis, we need to dynamize the
𝒅𝒀 𝒅𝑰 equation.
(𝒅𝒀 & 𝒅𝑰) ⇒ ( & ) Differentiating 𝜿 = 𝝆𝑲 w.r.t. 𝒕:
𝒅𝒕 𝒅𝒕 𝒅𝜿 𝒅𝑲
Differentiating w.r.t 𝒕 on both sides =𝝆
𝟏 𝒅𝒕 𝒅𝒕
the equation, 𝒅𝒀 = 𝒔 𝒅𝑰, we get: Since rate of change of capital is investment;
𝒅𝒀 𝟏 𝒅𝑰 𝒅𝑲
= 𝑰. We get:
=( ) 𝒅𝒕
𝒅𝒕 𝒔 𝒅𝒕 𝒅𝜿
Here 𝒔 (Marginal Propensity to Save) = 𝝆𝑰
𝒅𝒕
is assumed to be constant.
Static sense of Equilibrium Condition: At equilibrium, productive capacity is fully
utilized:
⏟
𝒀 = ⏟
𝜿
𝐂𝐮𝐫𝐫𝐞𝐧𝐭 𝐈𝐧𝐜𝐨𝐦𝐞(𝐩𝐫𝐨𝐝𝐮𝐜𝐭𝐢𝐨𝐧) 𝐩𝐨𝐭𝐞𝐧𝐭𝐢𝐚𝐥 𝐩𝐫𝐨𝐝𝐮𝐜𝐭𝐢𝐨𝐧
Dynamic sense of Equilibrium Condition: If economy starts from this (initial &
equilibrium) condition, then respective changes should be balanced, i.e.
𝒅𝒀 𝒅𝜿
=
𝒅𝒕 𝒅𝒕
______________________________________________________________________________
𝒅𝒀 𝒅𝜿
=
⏟
𝒅𝒕 ⏟
𝒅𝒕
𝑹𝒂𝒕𝒆 𝒐𝒇 𝑪𝒉𝒂𝒏𝒈𝒆 𝒐𝒇 𝑰𝒏𝒄𝒐𝒎𝒆 𝑹𝒂𝒕𝒆 𝒐𝒇 𝑪𝒉𝒂𝒏𝒈𝒆 𝒐𝒇 𝑷𝒓𝒐𝒅𝒖𝒄𝒕𝒊𝒗𝒆 𝑪𝒂𝒑𝒂𝒄𝒊𝒕𝒚
𝒐𝒗𝒆𝒓 𝒕𝒊𝒎𝒆 𝒐𝒗𝒆𝒓 𝒕𝒊𝒎𝒆
Above mentioned is the condition for dynamic equilibrium in the economy under Domar
Growth model.
TOPIC 071: DOMAR GROWTH MODEL: SOLUTION
From the framework of Domar Growth Model, productive capacity remains fully utilized over time:
𝒅𝒀 𝒅𝜿
=
𝒅𝒕 𝒅𝒕
𝒅𝒀 𝒅𝜿
=
⏟
𝒅𝒕 ⏟
𝒅𝒕
𝑹𝒂𝒕𝒆 𝒐𝒇 𝑪𝒉𝒂𝒏𝒈𝒆 𝒐𝒇 𝑰𝒏𝒄𝒐𝒎𝒆 𝑹𝒂𝒕𝒆 𝒐𝒇 𝑪𝒉𝒂𝒏𝒈𝒆 𝒐𝒇 𝑷𝒓𝒐𝒅𝒖𝒄𝒕𝒊𝒗𝒆 𝑪𝒂𝒑𝒂𝒄𝒊𝒕𝒚
𝒐𝒗𝒆𝒓 𝒕𝒊𝒎𝒆 𝒐𝒗𝒆𝒓 𝒕𝒊𝒎𝒆
In the framework, we found:
𝒅𝒀 𝟏 𝒅𝑰
=
𝒅𝒕 𝒔 𝒅𝒕
𝒅𝜿
=𝝆𝑰
𝒅𝒕
𝒅𝒀 𝒅𝜿
Substituting the values in 𝒅𝒕
= 𝒅𝒕
, we get:
𝟏 𝒅𝑰
=𝝆𝑰
𝒔 𝒅𝒕
Rearranging for 𝑰 in terms of 𝒕:
𝒅𝑰
= 𝝆𝒔 𝒅𝒕
𝑰
Introducing integral on both sides:
𝒅𝑰
∫ = ∫ 𝝆𝒔 𝒅𝒕
𝑰
𝐥𝐧|𝑰| + 𝒄𝑰 = 𝝆𝒔𝒕 + 𝒄𝒕
𝐥𝐧|𝑰| = 𝝆𝒔𝒕 + 𝒄𝒕 − 𝒄𝑰
𝐥𝐧|𝑰| = 𝝆𝒔𝒕 + 𝒄 [∵ 𝒄𝒕 − 𝒄𝑰 = 𝒄 (𝑦𝑒𝑡 𝑎𝑛𝑜𝑡ℎ𝑒𝑟 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡)]
We need solution as a function of 𝑰 and not of its logarithm (𝐥𝐧|𝑰|).
Therefore, raising both side to natural exponent (𝒆).
𝒆𝐥𝐧|𝑰| = 𝒆𝝆𝒔𝒕+𝒄
𝑰 = 𝒆𝝆𝒔𝒕 𝒆𝒄
𝑰 = 𝒆𝝆𝒔𝒕 𝑨 [∵ 𝒆𝒄 = 𝑨(𝑎𝑛𝑜𝑡ℎ𝑒𝑟 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡)]
Rearranging R.H.S and writing 𝑰 as function is in terms of 𝒕:
𝑰(𝒕) = 𝑨𝒆𝝆𝒔𝒕
Indeterminate/General Time Path of investment due to 𝑨.
Introducing initial condition to definitize arbitrary constant 𝑨, i.e. 𝒕 = 𝟎
𝑰(𝟎) = 𝑨𝒆𝝆𝒔(𝟎)
𝑰(𝟎) = 𝑨
Substituting value of arbitrary constant 𝑨 in general solution, 𝑰(𝒕) = 𝑨𝒆𝝆𝒔𝒕
𝑰(𝒕) = 𝑰(𝟎) 𝒆𝝆𝒔𝒕
______________________________________________________________________________
Here 𝑰(𝟎) is the initial investment. Moreover, marginal propensity to save (𝒔) and productive
capacity-capital ratio (𝝆) are already assumed to be constants.
Therefore, the definite solution of the time path of investment is:
𝑰(𝒕) = 𝑰(𝟎)𝒆𝝆𝒔𝒕
TOPIC 072: DOMAR GROWTH MODEL: NUMERICAL
From the solution of Domar Growth Model, we got the definite solution of time path of
investment: 𝑰(𝒕) = 𝑰(𝟎)𝒆𝝆𝒔𝒕
It’s an exponential function and when compared with standard form 𝒆𝒓𝒕 , we get 𝒓 = 𝝆𝒔
So, the growth of investment depends on the expression 𝝆𝒔.
Assuming values of 𝝆 and 𝒔: 𝝆 = 𝟐 and 𝒔 = 𝟎. 𝟎𝟏 to 𝟎. 𝟓.
Time Path 𝒔 𝝆𝒔 𝑰(𝒕) = 𝑰(𝟎)𝒆𝝆𝒔𝒕 𝑰(𝒕) = 𝒆𝝆𝒔𝒕 [∵ 𝑰(𝟎) = 𝟏]
A 0.01 0.02 𝐼(0)𝑒 0.02𝑡 𝑰(𝒕) = 𝒆𝟎.𝟎𝟐𝒕
0.1𝑡
B 0.05 0.1 𝐼(0)𝑒 𝑰(𝒕) = 𝒆𝟎.𝟏𝒕
C 0.1 0.2 𝐼(0)𝑒 0.2𝑡 𝑰(𝒕) = 𝒆𝟎.𝟐𝒕
D 0.2 0.4 𝐼(0)𝑒 0.4𝑡 𝑰(𝒕) = 𝒆𝟎.𝟒𝒕
0.6𝑡
E 0.3 0.6 𝐼(0)𝑒 𝑰(𝒕) = 𝒆𝟎.𝟔𝒕
F 0.4 0.8 𝐼(0)𝑒 0.8𝑡 𝑰(𝒕) = 𝒆𝟎.𝟖𝒕
G 0.5 1 𝐼(0)𝑒 𝑡 𝑰(𝒕) = 𝒆𝒕
Plotting the numerical form of definite solution of time paths; A, B, C, D, E, F, G:
Observations: The composite graph shows 7-time paths. These time paths are generated due
to alternative values of Marginal Propensity to Save (𝒔). However, all time paths have a common
initial condition 𝑰(𝟎).
It is observable that increase in value of 𝒔 uplifts the time path. Specifically, increase in marginal
propensity to save increases the rate at which the investment takes place. As increased savings
lead to increased investment, the rate of growth of investment also increases.
______________________________________________________________________________
Lesson 18
FIRST ORDER DIFFERENTIAL EQUATIONS
TOPIC 073: FIRST-ORDER DIFFERENTIAL EQUATIONS: INTRODUCTION
A differential equation is an equation which expresses an explicit or implicit relationship between

a function 𝒚 = 𝒇(𝒕) and one or more of its derivatives or differentials. For example:
𝒅𝒚
= 𝟏𝟎𝒙
𝒅𝒙
It is a First-Order differential equation, since the highest derivative is a first-order derivative.
𝒚′ = 𝟏𝟐𝒚 is also a First-Order differential equation as it contains 𝒚′ .
𝒚′′ − 𝟐𝒚′ + 𝟏𝟗 = 𝟎 is a Second-Order differential equation as it contains second order derivative
𝒚′′ .
In addition to order of differential equation, degree can also be noted. It is the power of the highest
order of the derivative in the equation.
Differential Equation Order Degree
𝒅𝒚 1st
Order 1 st
Degree
= 𝟐𝒙 + 𝟔
𝒅𝒕
OR
𝒅𝒚 𝟏
( 𝒅𝒕 ) = 𝟐𝒙 + 𝟔
𝟒 1st Order 4th Degree
𝒅𝒚
( ) − 𝟓𝒕𝟓 = 𝟎
𝒅𝒕
𝒅𝟐 𝒚 𝒅𝒚 𝟑 2nd Order 1st Degree
𝒅𝒕𝟐
+ ( 𝒅𝒕
) + 𝒙𝟐 = 𝟎
OR
𝟏
𝒅𝟐 𝒚 𝒅𝒚 𝟑
( 𝒅𝒕𝟐
) + ( 𝒅𝒕 ) + 𝒙𝟐 = 𝟎
𝟕 𝟓 3rd Order 5th Degree
𝒅𝟐 𝒚 𝒅𝟑 𝒚
( 𝒅𝒕𝟐 ) + ( 𝒅𝒕𝟑 ) = 𝟕𝟓𝒚
OR
𝟕 𝟓
𝒅𝟐 𝒚 𝒅𝟑 𝒚
( 𝟐 ) +( ) = 𝟕𝟓𝒚
𝒅𝒕 𝒅𝒕𝟑
Solution: Solution of such an equation will be 𝒚 = 𝒇(𝒙), where no derivative term would be
present.
A standard form of First-Order Differential Equation:
𝒅𝒚
+ 𝒖(𝒕) ∙ 𝒚 = 𝒘(𝒕)
𝒅𝒕
𝒅𝒚
𝒅𝒕
: is the derivative of dependent variable 𝒚 w.r.t independent variable 𝒕.
𝒖(𝒕): Coefficient of 𝒚. A function in terms of 𝒕.
𝒘(𝒕): Term. A function in terms of 𝒕.
Caveat: Here we assume 𝒖(𝒕) and 𝒘(𝒕) are constant. Therefore, using constant symbols for 𝒖(𝒕)
and 𝒘(𝒕):
𝒖(𝒕) = 𝒂
𝒘(𝒕) = 𝒃
𝒅𝒚
Therefore, the standard form 𝒅𝒕 + 𝒖(𝒕) ∙ 𝒚 = 𝒘(𝒕) becomes:
______________________________________________________________________________
𝒅𝒚
+𝒂∙𝒚= 𝒃
𝒅𝒕
𝒅𝒚
+ ⏟
𝒂 𝒚
⏟ = ⏟
𝒃
𝒅𝒕 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒 𝑡𝑒𝑟𝑚
TOPIC 074: FIRST-ORDER DIFFERENTIAL EQUATIONS: HOMOGENEOUS CASE
Standard Form of First-Order Differential Equation

𝒅𝒚
+𝒂∙𝒚= 𝒃
𝒅𝒕
Depending on the values of 𝒂 and 𝒃, we get 3 possible cases:
Case – I: 𝒂 ≠ 𝟎, 𝒃 = 𝟎 [Homogenous Case]
Case – II: 𝒂 ≠ 𝟎, 𝒃 ≠ 𝟎 [Non-Homogenous Case – I]
Case – III: 𝒂 = 𝟎, 𝒃 ≠ 𝟎 [Non-Homogenous Case – II]
Case – I: Homogenous Case: 𝒂 ≠ 𝟎, 𝒃 = 𝟎

𝒅𝒚
Standard form of First-Order Differential Equation + 𝒂𝒚 = 𝒃, becomes:
𝒅𝒕
𝒅𝒚
+ 𝒂𝒚 = ⏟
𝟎
𝒅𝒕 ℎ𝑜𝑚𝑜𝑔𝑒𝑛𝑒𝑜𝑢𝑠
𝑐𝑎𝑠𝑒
𝒅𝒚
= −𝒂𝒚
𝒅𝒕
𝟏 𝒅𝒚
= −𝒂
𝒚 𝒅𝒕
𝒅𝒚
( )
𝒅𝒕
= −𝒂
𝒚
Integrating w.r.t. 𝒕
𝒅𝒚
( )
𝒅𝒕
∫ 𝒅𝒕 = − ∫ 𝒂 𝒅𝒕
𝒚
(𝒅𝒚)
∫ = − ∫ 𝒂 𝒅𝒕
𝒚
𝟏
∫ ( ) 𝒅𝒚 = − ∫ 𝒂 𝒅𝒕
𝒚
𝒍𝒏|𝒚| + 𝒄𝟏 = −𝒂𝒕 + 𝒄𝟐
𝒍𝒏|𝒚| = −𝒂𝒕 + 𝒄𝟐 − 𝒄𝟏 [∵ 𝒄𝟐 − 𝒄𝟐 = 𝒄]
𝒍𝒏|𝒚| = −𝒂𝒕 + 𝒄
To get rid of natural logarithm, raise exponent on both sides:
𝒆𝒍𝒏|𝒚| = 𝒆−𝒂𝒕+𝒄
𝒚 = 𝒆−𝒂𝒕 𝒆𝒄
𝒚(𝒕) = 𝒆−𝒂𝒕 𝑨 [∵ 𝒆𝒄 = 𝑨] 𝑦𝑒𝑡 𝑎𝑛𝑜𝑡ℎ𝑒𝑟 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
−𝒂𝒕
𝒚(𝒕) = 𝑨𝒆 (General Solution)
Introducing the initial condition to definitize the arbitrary constant; (𝒕 = 𝟎): 𝒚(𝒕) = 𝑨𝒆−𝒂𝒕
becomes
𝒚(𝟎) = 𝑨𝒆−𝒂(𝟎)
𝒚(𝟎) = 𝑨𝒆𝟎
______________________________________________________________________________
𝒚(𝟎) = 𝑨
𝑨 = 𝒚(𝟎)
𝒚(𝒕) = 𝒚(𝟎)𝒆−𝒂𝒕
It is Definite Solution as the arbitrary constant 𝑨 is determined.
TOPIC 075: FIRST-ORDER DIFFERENTIAL EQUATIONS: NONHOMOGENEOUS CASE – I
Case – II: Non-Homogenous Case – I: 𝒂 ≠ 𝟎, 𝒃 ≠ 𝟎. (𝑵𝑯𝑪𝒂≠𝟎 )

𝒅𝒚
Standard form of First-Order Differential Equation 𝒅𝒕 + 𝒂𝒚 = 𝒃, remains:
𝒅𝒚
+ 𝒂𝒚 = ⏟
𝒃
𝒅𝒕 𝑁𝑜𝑛−𝐻𝑜𝑚𝑜𝑔𝑒𝑛𝑜𝑢𝑠
𝑐𝑎𝑠𝑒
As no term is equal to zero (𝒃 ≠ 𝟎) in this case, Non-Homogenous case is also termed as
Complete Equation.
In the Complete Equation of time path, there is equilibrium value and the deviation from it.
As in graph of time path, 𝟓𝟎 is equilibrium value and ±𝟒𝟎𝒆−𝟐𝒕 is the deviation from equilibrium.
Time path convergences towards equilibrium (𝟓𝟎) in either case. i.e. when deviation is positive
+𝟒𝟎𝒆−𝟐𝒕 or negative −𝟒𝟎𝒆−𝟐𝒕 .
Therefore, solution has two components i.e. Complementary Function (𝒚𝒄 = ±𝟒𝟎𝒆−𝟐𝒕 ) and
Particular Solution (𝒚𝒑 = 𝟓𝟎).
Complementary Function (𝒚𝒄 ) is complementary (‫ پورا کرنے واال‬- ‫ )تکمیلی‬to the Equilibrium value
(𝒚𝒑 ) of the time path.
Equilibrium value (𝒚𝒑 ) is called Particular Solution as it is one Particular (‫ )مخصوص‬value of 𝒚 out
of infinite values on whole of the time path.
Case of Positive 𝒚𝒄 (𝒚𝒄 = 𝟒𝟎𝒆−𝟐𝒕 ) Case of Negative 𝒚𝒄 (𝒚𝒄 = −𝟒𝟎𝒆−𝟐𝒕 )
𝒚(𝒕) = + 𝟒𝟎𝒆−𝟐𝒕
⏟ + 𝟓𝟎
⏟ 𝒚(𝒕) = − 𝟒𝟎𝒆−𝟐𝒕
⏟ + 𝟓𝟎
⏟
𝒚𝒄 [𝒕→∞,𝒚𝒄 →𝟎,𝒚(𝒕)→𝒚𝒑 ] 𝒚𝒑 𝒚𝒄 [𝒕→∞,|𝒚𝒄 |→𝟎,𝒚(𝒕)→𝒚𝒑 ] 𝒚𝒑
Mathematically, addition decreases Mathematically, deduction decreases
______________________________________________________________________________
Graphically, converges from above the Graphically, converges from below the
equilibrium. equilibrium.
Complementary Function (𝒚𝒄 ): is equivalent to the general solution of homogenous case.

(𝒚𝒄 )𝑵𝑯𝑪 = {𝒚(𝒕)}𝑯𝑪
(𝒚𝒄 )𝑵𝑯𝑪 = 𝑨𝒆−𝒂𝒕
𝒚(𝒕) = 𝑨𝒆−𝒂𝒕 + 𝒚𝒑
Particular Solution (𝒚𝒑 ): of the Non-Homogenous Case is as follows:
Two possibilities of particular values of 𝒚: constant and non-constant.
1. Can be a constant (say 𝒌) when (𝒂 ≠ 𝟎):
Let,
𝒚=𝒌
𝒅𝒚
=𝟎
𝒅𝒕
𝒅𝒚 𝒅𝒚
Substituting value of 𝒅𝒕 in ( 𝒅𝒕 + 𝒂𝒚 = 𝒃), we get:
𝟎 + 𝒂𝒚 = 𝒃
𝒃
𝒚=
𝒂
𝒃
In proper notion as per the definition is: 𝒚𝒑 = 𝒂
Substituting the value of 𝒚𝒑 in 𝒚(𝒕):
𝒃
𝒚(𝒕) = 𝑨𝒆−𝒂𝒕 + (General Solution of Non-homogenous Case with 𝒂 ≠ 𝟎)
𝒂
𝒃
For Definite Solution: Put 𝒕 = 𝟎 in 𝒚(𝒕) = 𝑨𝒆−𝒂𝒕 +
𝒂
𝒃
𝒚(𝟎) = 𝑨𝒆−𝒂(𝟎) +
𝒂
𝒃
𝒚(𝟎) = 𝑨𝒆𝟎 +
𝒂
𝒃
𝒚(𝟎) = 𝑨 +
𝒂
𝒃
𝑨 = {𝒚(𝟎) − }
𝒂
𝒃
Putting value of 𝑨 in 𝒚(𝒕) = 𝑨𝒆−𝒂𝒕 + 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
(Definite Solution of Non-homogenous Case with 𝒂 ≠ 𝟎).
What if, 𝒂 = 𝟎?
______________________________________________________________________________
TOPIC 076: FIRST-ORDER DIFFERENTIAL EQUATIONS: NONHOMOGENEOUS CASE – II
Case – III: Non-Homogenous Case – II: 𝒂 = 𝟎, 𝒃 ≠ 𝟎. (𝑵𝑯𝑪𝒂=𝟎 )
𝒅𝒚
Standard form of First-Order Differential Equation 𝒅𝒕 + 𝒂𝒚 = 𝒃, at 𝒂 = 𝟎 becomes:
𝒅𝒚
+ (𝟎)𝒚 = 𝒃
𝒅𝒕
𝒅𝒚
=𝒃
𝒅𝒕
Solution 𝒚(𝒕) has two components:
𝒚(𝒕) = 𝒚𝒄 + 𝒚𝒑
Complementary Function (𝒚𝒄 ): is equivalent to the general solution of homogenous case.
(𝒚𝒄 )𝑵𝑯𝑪 = {𝒚(𝒕)}𝑯𝑪
(𝒚𝒄 )𝑵𝑯𝑪 = 𝑨𝒆−𝒂𝒕
However here 𝒂 = 𝟎 therefore 𝒚𝒄 reduces to:
𝒚𝒄 = 𝑨𝒆−(𝟎)𝒕
𝒚𝒄 = 𝑨𝒆𝟎
𝒚𝒄 = 𝑨
𝒚(𝒕) = 𝑨 + 𝒚𝒑
Particular Solution (𝒚𝒑 ): If 𝒚 is considered as a constant 𝒌 as in case of 𝑵𝑯𝑪𝒂≠𝟎, 𝒚 = 𝒌
𝒅𝒚
=𝟎
𝒅𝒕
𝒅𝒚 𝒅𝒚
Substituting value of 𝒅𝒕 in ( 𝒅𝒕 = 𝒃), we get:
𝟎=𝒃
𝟎≠𝒃
So, equality is not established and some other value of 𝒚 should be experimented.
Therefore, we experiment a non-constant (𝒚 = 𝒌𝒕) can be experimented.
𝒚 = 𝒌𝒕
Taking derivative w.r.t 𝒕:
𝒅𝒚 𝒅𝒕
= 𝒌( )
𝒅𝒕 𝒅𝒕
𝒅𝒚
= 𝒌
𝒅𝒕
𝒅𝒚
We already have the 𝑵𝑯𝑪𝒂=𝟎 differential equation 𝒅𝒕
= 𝒃
𝒌 = 𝒃
Therefore, 𝒚 = 𝒌𝒕 can be written as 𝒚 = 𝒃𝒕.
More properly, it is the proper solution: 𝒚𝒑 = 𝒃𝒕.
Substituting the value of 𝒚𝒑 in 𝒚(𝒕) = 𝑨 + 𝒚𝒑 :
𝒚(𝒕) = 𝑨 + 𝒃𝒕
𝒚(𝒕) = 𝑨 + 𝒃𝒕 is the General Solution of 𝑵𝑯𝑪𝒂=𝟎 .
Definite Solution: Put 𝒕 = 𝟎 in
𝒚(𝒕) = 𝑨 + 𝒃𝒕
𝒚(𝟎) = 𝑨 + 𝒃(𝟎)
𝒚(𝟎) = 𝑨 + 𝟎
______________________________________________________________________________
𝒚(𝟎) = 𝑨
𝑨 = 𝒚(𝟎)
Putting value of 𝑨 in General Solution of 𝑵𝑯𝑪𝒂=𝟎 .
𝒚(𝒕) = 𝒚(𝟎) + 𝒃𝒕
Definite Solution of Non-homogenous Case with 𝒂 = 𝟎.
TOPIC 077: VERIFICATION OF SOLUTION OF FIRST-ORDER DIFFERENTIAL EQUATIONS
The validity of all solutions of differential equations can always be checked by differentiation.
Consider the definite solution of First Order Differential Equation where 𝒂 ≠ 𝟎:
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
Differentiating w.r.t 𝒕:
𝒅 𝒅 𝒃 𝒃
{𝒚(𝒕)} = [{𝒚(𝟎) − } 𝒆−𝒂𝒕 + ]
𝒅𝒕 𝒅𝒕 𝒂 𝒂
𝒅𝒚 𝒅 𝒃 −𝒂𝒕 𝒅 𝒃
= [{𝒚(𝟎) − } 𝒆 ] + ( )
𝒅𝒕 𝒅𝒕 𝒂 𝒅𝒕 𝒂
𝒅𝒚 𝒃 𝒅
= {𝒚(𝟎) − } { (−𝒂𝒕)} (𝒆−𝒂𝒕 ) + 𝟎
𝒅𝒕 𝒂 𝒅𝒕
𝒅𝒚 𝒃
= {𝒚(𝟎) − } −𝒂(𝒆−𝒂𝒕 )
𝒅𝒕 𝒂
𝒅𝒚 𝒃
= −𝒂 {𝒚(𝟎) − } 𝒆−𝒂𝒕
𝒅𝒕 𝒂
𝒅𝒚
Recalling standard form of First Order Differential Equation + 𝒂 𝒚 = 𝒃 and substituting boxed
𝒅𝒕
values:
𝒃 𝒃 𝒃
−𝒂 {𝒚(𝟎) − } 𝒆−𝒂𝒕 + 𝒂 {𝒚(𝟎) − } 𝒆−𝒂𝒕 + = 𝒃
𝒂 𝒂 𝒂
𝒃 −𝒂𝒕 𝒃 −𝒂𝒕 𝒃
−𝒂 {𝒚(𝟎) − } 𝒆 + 𝒂 {𝒚(𝟎) − } 𝒆 +𝒂× =𝒃
𝒂 𝒂 𝒂
𝒃
⏟= 𝒃
𝐿.𝐻.𝑆=𝑅.𝐻.𝑆
𝒅𝒚
Numerical Example: Verify if {𝒚(𝒕) = 𝟒(𝟏 − 𝒆−𝒕 )} the definite solution of ( 𝒅𝒕 + 𝒚 = 𝟒) is true:
Definite Solution: 𝒚(𝒕) = 𝟒(𝟏 − 𝒆−𝒕 )
Differentiating w.r.t 𝒕.
𝒅 𝒅
{𝒚(𝒕)} = {𝟒(𝟏 − 𝒆−𝒕 )}
𝒅𝒕 𝒅𝒕
𝒅𝒚 𝒅
= 𝟒 (𝟏 − 𝒆−𝒕 )
𝒅𝒕 𝒅𝒕
𝒅𝒚 𝒅
= 𝟒 [𝟎 − { (−𝒕)} 𝒆−𝒕 ]
𝒅𝒕 𝒅𝒕
𝒅𝒚
= 𝟒{(−𝟏)𝒆−𝒕 }
𝒅𝒕
𝒅𝒚
= −𝟒𝒆−𝒕
𝒅𝒕
𝒅𝒚
Substituting values in the First Order Differential Equation, 𝒅𝒕
+ 𝒚 = 𝟒;
−𝟒𝒆−𝒕 + 𝟒(𝟏 − 𝒆−𝒕 ) = 𝟒
______________________________________________________________________________
−𝟒𝒆−𝒕 + 𝟒(𝟏 − 𝒆−𝒕 ) = 𝟒

−𝟒𝒆−𝒕 + 𝟒 − 𝟒𝒆−𝒕 = 𝟒
𝟒
⏟= 𝟒
𝐿.𝐻.𝑆=𝑅.𝐻.𝑆
TOPIC 078: FIRST-ORDER DIFFERENTIAL EQUATIONS: SUMMARY & NUMERICALS
𝒅𝒚
1. GRAPH of 𝒅𝒕
− 𝟐𝒚 = 𝟎: A HOMOGENOUS CASE:
Solution: Comparing the given differential equation with the standard form
𝒅𝒚
− 𝟐𝒚 = 𝟎
𝒅𝒕
𝒅𝒚
+ 𝒂𝒚 = 𝒃
𝒅𝒕
𝒂 = −𝟐, 𝒃 = 𝟎 (𝐚 𝐇𝐨𝐦𝐨𝐠𝐞𝐧𝐨𝐮𝐬 𝐜𝐚𝐬𝐞)
Formula for General Solution of Homogenous case:
𝒚(𝒕) = 𝑨𝒆−𝒂𝒕
Substituting value of 𝒂:
𝒚(𝒕) = 𝑨𝒆−(−𝟐)𝒕
𝒚(𝒕) = 𝑨𝒆𝟐𝒕 (General Solution)
𝟐𝒕
𝒚(𝒕) = 𝑨𝒆
⏟ +⏟
𝟎 (General Solution)
𝒚𝒄 𝒚𝒑
Interpretation: The graph verifies, that there is an exponential growth term in the solution (𝒆𝟐𝒕 ).
While 𝑨 remains the arbitrary element, which can be determined by introducing the initial
condition.
Definite Solution: requires the introduction of initial condition; 𝒕 = 𝟎:
𝒚(𝟎) = 𝑨𝒆𝟐(𝟎)
𝒚(𝟎) = 𝑨𝒆𝟎
𝒚(𝟎) = 𝑨(𝟏)
𝑨 = 𝒚(𝟎)
Substituting the values of 𝑨 in General Solution: 𝒚(𝒕) = 𝑨𝒆𝟐𝒕
𝒚(𝒕) = 𝒚(𝟎)𝒆𝟐𝒕
______________________________________________________________________________
𝒅𝒚
𝒚(𝒕) = 𝒚(𝟎)𝒆𝟐𝒕 is the definite solution of the differential equation 𝒅𝒕
− 𝟐𝒚 = 𝟎.
𝒅𝒚
+ 𝟒𝒚 = 𝟏𝟐: A NON-HOMOGENOUS CASE: with (𝒂 ≠ 𝟎, 𝒃 ≠ 𝟎)
𝒅𝒚
+ 𝟒𝒚 = 𝟏𝟐
𝒅𝒕
𝒅𝒚
+ 𝒂𝒚 = 𝒃
𝒅𝒕
𝒂 = 𝟒, 𝒃 = 𝟏𝟐 (𝐚 𝐍𝐨𝐧 − 𝐇𝐨𝐦𝐨𝐠𝐞𝐧𝐨𝐮𝐬 𝐜𝐚𝐬𝐞 𝐰𝐢𝐭𝐡 𝒂 ≠ 𝟎)
Formula for General Solution of Non-Homogenous case:
𝒃
𝒚(𝒕) = 𝑨𝒆−𝒂𝒕 +
𝒂
Substituting value of 𝒂:
𝟏𝟐
𝒚(𝒕) = 𝑨𝒆−(𝟒)𝒕 +
𝟒
−𝟒𝒕
⏟ +⏟𝟑 (General Solution)
𝒚𝒄 𝒚𝒑
Interpretation: In the graph it is visible with a red horizontal line that value of particular integral
is 𝟑. However, the complementary function has an exponential decay that leads to its
convergence towards equilibrium (𝒚𝒑 ).
Definite Solution: For its sake, we introduce; 𝒕 = 𝟎:
𝒚(𝒕) = 𝑨𝒆−𝟒𝒕 + 𝟑
𝒚(𝟎) = 𝑨𝒆−𝟒(𝟎) + 𝟑
𝒚(𝟎) = 𝑨𝒆𝟎 + 𝟑
𝒚(𝟎) = 𝑨(𝟏) + 𝟑
𝒚(𝟎) = 𝑨 + 𝟑
𝒚(𝟎) − 𝟑 = 𝑨
𝒚(𝟎) − 𝟑 = 𝑨
𝑨 = {𝒚(𝟎) − 𝟑}
Substituting the value of arbitrary constant in general solution: 𝒚(𝒕) = 𝑨𝒆−𝟒𝒕 + 𝟑
______________________________________________________________________________
𝒚(𝒕) = {𝒚(𝟎) − 𝟑}𝒆−𝟒𝒕 + 𝟑

𝒅𝒚
𝒚(𝒕) = {𝒚(𝟎) − 𝟑}𝒆−𝟒𝒕 + 𝟑 is the definite solution of the differential equation 𝒅𝒕
+ 𝟒𝒚 = 𝟏𝟐.
𝒅𝒚
= 𝟐𝟑: A NON-HOMOGENOUS CASE: with (𝒂 = 𝟎, 𝒃 ≠ 𝟎)
𝒅𝒚
= 𝟐𝟑
𝒅𝒕
Rewriting for comparability
𝒅𝒚
+ (𝟎)𝒚 = 𝟐𝟑
𝒅𝒕
𝒅𝒚
+ 𝒂𝒚 = 𝒃
𝒅𝒕
𝒂 = 𝟎, 𝒃 = 𝟐𝟑 (𝐚 𝐍𝐨𝐧 − 𝐇𝐨𝐦𝐨𝐠𝐞𝐧𝐨𝐮𝐬 𝐜𝐚𝐬𝐞 𝐰𝐢𝐭𝐡 𝒂 = 𝟎)
Formula for General Solution of Non-Homogenous case:
𝒚(𝒕) = 𝑨 + 𝒃𝒕
Substituting value of 𝒃:
𝒚(𝒕) = 𝑨 + 𝟐𝟑𝒕
𝒚(𝒕) = 𝑨
⏟ + 𝟐𝟑𝒕
⏟ (General Solution)
𝒚𝒄 𝒚𝒑
Interpretation: In the graph it is visible with a red horizontal line that value of particular integral
is 𝟑. However, the complementary function has an exponential decay that leads to its
convergence towards equilibrium (𝒚𝒑 ).
Definite Solution: For its sake, we introduce; 𝒕 = 𝟎:
𝒚(𝒕) = 𝑨 + 𝟐𝟑𝒕
𝒚(𝟎) = 𝑨 + 𝟐𝟑(𝟎)
𝒚(𝟎) = 𝑨 + 𝟎
𝒚(𝟎) = 𝑨
𝑨 = 𝒚(𝟎)
Substituting the value of arbitrary constant in general solution: 𝒚(𝒕) = 𝑨 + 𝟐𝟑𝒕
______________________________________________________________________________
𝒚(𝒕) = 𝒚(𝟎) + 𝟐𝟑𝒕

𝒚(𝒕) = 𝒚(𝟎) + 𝟐𝟑𝒕
𝒅𝒚
𝒚(𝒕) = 𝒚(𝟎) + 𝟐𝟑𝒕 is the definite solution of the differential equation 𝒅𝒕 = 𝟐𝟑.
Summary of Homogenous and Non-Homogenous Cases of First Order Differential
Equations:
𝒅𝒚
+ 𝒂𝒚 = 𝒃
𝒅𝒕
Complementary
Constant Term
Coefficient
Particular
Constant
Function
Solution
Solution
Solution
General
Definite
̅
𝒂 ̅
𝒃 𝒚𝒄 𝒚𝒑 𝒚(𝒕)𝒕≠𝟎 𝒚(𝒕)𝒕=𝟎
HC 𝒂≠𝟎 𝒃=𝟎 𝑨𝒆−𝒂𝒕 N.A. 𝑨𝒆−𝒂𝒕 𝒚(𝟎)𝒆−𝒂𝒕
𝒂≠𝟎 𝒃≠𝟎 𝑨𝒆−𝒂𝒕 𝒃 𝒃 𝒃 𝒃
𝑨𝒆−𝒂𝒕 + {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
NHC 𝒂 𝒂 𝒂 𝒂
𝒂=𝟎 𝒃≠𝟎 𝑨 𝒃𝒕 𝑨 + 𝒃𝒕 𝒚(𝟎) + 𝒃𝒕
______________________________________________________________________________
Lesson 19
DYNAMICS OF MARKET PRICE
TOPIC 079: DYNAMICS OF MARKET PRICE: FRAMEWORK
Famous Microeconomic Model: Demand and Supply equations are given.

𝑸𝒅 = 𝜶 − 𝜷𝑷 [𝜶, 𝜷 > 𝟎]
𝑸𝒔 = −𝜸 + 𝜹𝑷 [𝜸, 𝜹 > 𝟎]
Static Sense of Equilibrium: In Static Sense the equilibrium shall be:
𝑸𝒅 = 𝑸𝒔
Substituting values of 𝑸𝒅 and 𝑸𝒔 in equilibrium condition:
𝑸𝒅 = 𝑸𝒔
𝜶 − 𝜷𝑷 = −𝜸 + 𝜹𝑷
𝜶 + 𝜸 = 𝜷𝑷 + 𝜹𝑷
𝜶 + 𝜸 = 𝑷(𝜷 + 𝜹)
𝜶+𝜸
=𝑷
𝜷+𝜹
𝜶+𝜸
𝑷∗ =
𝜷+𝜹
Parametric Interpretation: Since 𝜶, 𝜷, 𝜸 and 𝜹 are positive constants by given parametric
stipulations, the answer of their addition and division shall give rise to yet another positive
constant. i.e. 𝑷∗ = 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕 (> 𝟎).
Dynamic Sense of Equilibrium: In dynamic sense, the price shall follow a time-path 𝑷(𝒕), rather
being limited to a single equilibrium (𝑷∗ ). So, we write the expression of definite time-path for
price.
𝑷(𝒕) = 𝑷𝒄 + 𝑷𝒑
Here, particular solution of price 𝑷𝒑 is the static equilibrium value of price (𝑷∗ ). 𝑷(𝒕) = 𝑷𝒄 + 𝑷∗
𝑷(𝒕) = 𝑷𝒄 + 𝑷∗
Possibilities of Initial Conditions: Time path of price can start {𝑷(𝟎)} either from above, at or
below the equilibrium value of price (𝑷∗ ):
______________________________________________________________________________
Putting all 3 cases summarily: [𝑷(𝟎) ⋚ 𝑷∗ ]

Case – I: [𝑷(𝟎) = 𝑷∗ ] requires no adjustments over time. However, [𝑷(𝟎) ≶ 𝑷∗ ] do.
Case – II: [𝑷(𝟎) > 𝑷∗ ] requires adjustments over time.
Case – III: [𝑷(𝟎) < 𝑷∗ ] requires adjustments over time.
TOPIC 080: DYNAMICS OF MARKET PRICE: THE TIME PATH
Price Adjustment over time: Price adjustments over time can be measured by the rate of change
𝒅𝑷
of price over time ( 𝒅𝒕 ).
Mathematizing the shortage/surplus in static sense:
Shortage Surplus
𝑸𝒅 > 𝑸𝒔 𝑸𝒅 < 𝑸𝒔
(𝑸𝒅 − 𝑸𝒔 ) > 𝟎 {𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒} (𝑸𝒅 − 𝑸𝒔 ) < 𝟎 {𝑛𝑒𝑔𝑎𝑡𝑖𝑣𝑒}
Excess Demand Negative Excess Demand
Greater the shortage/surplus in the market, more responsive will be change in price and vice
versa.
(Price Adjustments) are directly proportional to (Excess demand)
𝒅𝑷
( ) ∝ (𝑸𝒅 − 𝑸𝒔 )
𝒅𝒕
Developing an equation and brining the constant of proportionality (𝒋).
𝒅𝑷
𝒅𝒕
= 𝒋 (𝑸𝒅 − 𝑸𝒔 )
𝒅𝑷
Here 𝒋 > 𝟎, which means that the relationship between 𝒅𝒕 and (𝑸𝒅 − 𝑸𝒔 ) shall be positive.
𝒋 is, suitably, called the ‘adjustment’ coefficient.
𝒅𝑷
= 𝒋
⏟ × (𝑸
⏟ 𝒅 − 𝑸𝒔 )
⏟
𝒅𝒕 𝐴𝑑𝑗𝑢𝑠𝑡𝑚𝑒𝑛𝑡 𝑀𝑎𝑟𝑘𝑒𝑡 𝐶𝑙𝑒𝑎𝑟𝑖𝑛𝑔
𝐼𝑛𝑡𝑒𝑟𝑡𝑒𝑚𝑝𝑜𝑟𝑎𝑙
𝐶𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑆𝑒𝑛𝑠𝑒
𝑆𝑒𝑛𝑠𝑒
______________________________________________________________________________
𝒅𝑷
At equilibrium, no need for adjustment in price over time in equilibrium. i.e. ( = 𝟎)
𝒅𝒕
At equilibrium, there is no excess demand in the market. i.e. (𝑸𝒅 − 𝑸𝒔 = 𝟎)
𝒅𝑷
Putting the two values in = 𝒋 × (𝑸𝒅 − 𝑸𝒔 ):
𝒅𝒕
𝟎 = 𝒋 × (𝟎)
𝟎=𝟎
Therefore, the adjustment equation makes sense for market equilibrium both in dynamic
sense as well as static sense.
Now substituting the values of 𝑸𝒅 and 𝑸𝒔 in Adjustment equation:
𝒅𝑷
= 𝒋 × (𝑸𝒅 − 𝑸𝒔 )
𝒅𝒕
𝒅𝑷
= 𝒋 {(𝜶 − 𝜷𝑷) − (−𝜸 + 𝜹𝑷)}
𝒅𝒕
𝒅𝑷
= 𝒋(𝜶 − 𝜷𝑷 + 𝜸 − 𝜹𝑷)
𝒅𝒕
Gathering terms containing 𝑷.
𝒅𝑷
= 𝒋(𝜶 + 𝜸 − 𝜷𝑷 − 𝜹𝑷)
𝒅𝒕
𝒅𝑷
= 𝒋{(𝜶 + 𝜸) − (𝜷 + 𝜹)𝑷}
𝒅𝒕
𝒅𝑷
= 𝒋(𝜶 + 𝜸) − 𝒋(𝜷 + 𝜹)𝑷
𝒅𝒕
𝒅𝑷
+ 𝒋(𝜷 + 𝜹)𝑷 = 𝒋(𝜶 + 𝜸)
𝒅𝒕
Comparing with the standard form of First Order Differential Equation:
𝒅𝑷
+ 𝒋(𝜷 + 𝜹) 𝑷
⏟=⏟ 𝒋(𝜶 + 𝜸)
⏟ ⏟
𝒅𝒕
𝑎 𝑦 𝑏
𝑑𝑦
𝑑𝑡
𝒅𝒚 𝒅𝑷
Here 𝒅𝒕
= 𝒅𝒕
,𝒚 = 𝑷, 𝒂 = 𝒋(𝜷 + 𝜹), 𝒃 = 𝒋(𝜶 + 𝜸)
Choice of Case of the Differential Equation:
𝒃 = 𝒋(𝜶 + 𝜸) ; as 𝒋, 𝜶 and 𝜸 are positive, the value of 𝒃 shall be some positive constant. i.e.
𝒃 ≠ 𝟎 – Non-Homogenous Case.
𝒂 = 𝒋(𝜷 + 𝜹); as 𝒋, 𝜷 and 𝜹 are positive, the value of 𝑎 shall be some positive constant. i.e.
𝒂 ≠ 𝟎 – Non-Homogenous Case with 𝒂 ≠ 𝟎.
Therefore, using the definite Solution of time-path of 𝑵𝑯𝑪𝒂≠𝟎 :
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
Emphasizing on the values to be substituted:
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
Substituting values of 𝒚, 𝒂 and 𝒃, respectively:
𝒋(𝜶 + 𝜸) − 𝒋(𝜷+𝜹) 𝒕 𝒋(𝜶 + 𝜸)
𝑷(𝒕) = {𝑷(𝟎) − }𝒆 +
𝒋(𝜷 + 𝜹) 𝒋(𝜷 + 𝜹)
𝜶+𝜸 𝜶+𝜸
= {𝑷(𝟎) − 𝜷+𝜹} 𝒆−𝒋(𝜷+𝜹)𝒕 + 𝜷+𝜹
𝑷(𝒕) = {𝑷(𝟎) − 𝑷∗ }𝒆−𝒋(𝜷+𝜹)𝒕 + 𝑷∗ Let {𝒋(𝜷 + 𝜹)} = 𝒌
𝑷(𝒕) = {𝑷(𝟎) − 𝑷∗ }𝒆−𝒌𝒕 + 𝑷∗
______________________________________________________________________________
{𝑷(𝟎) − 𝑷∗ }𝒆−𝒌𝒕 + 𝑷
𝑷(𝒕) = ⏟ ⏟∗
𝑃𝑐 𝑃𝑝
Definite Time-path for the Market Equilibrium in Dynamic Sense.
TOPIC 081: DYNAMICS OF MARKET PRICE: DYNAMIC STABILITY
For the dynamic stability of the time-path of First Order Differential Equation, it should converge
towards equilibrium. We know that equilibrium is particular solution (𝑷𝒑 ). Whereas convergence
happens when deviation from equilibrium fades to zero. In a differential equation, deviation
component is measured by (𝑷𝒄 ). Therefore, for dynamic stability; 𝑷𝒄 → 𝟎 as 𝒕 → ∞.
Using the Definite Time-path for Market Price in Dynamic Sense.
𝑷(𝒕) = {𝑷(𝟎) − 𝑷∗ }𝒆−𝒌𝒕 + 𝑷∗
{𝑷(𝟎) − 𝑷∗ }𝒆−𝒌𝒕 + 𝑷
𝑷(𝒕) = ⏟ ⏟∗
𝑃𝑐 𝑃𝑝
Need for time path of price {𝑷(𝒕)} to converge to equilibrium price (𝑷∗ ) i.e. 𝑷(𝒕) → 𝑷∗ as 𝑷𝒄 →
𝟎 when 𝒕 → ∞.
𝑷𝒄 → 𝟎
{𝑷(𝟎) − 𝑷∗ }𝒆−𝒌𝒕 → 𝟎
𝜶+𝜸
{ >𝟎} {𝒋(𝜷+𝜹)>𝟎}
𝜷+𝜹
⏞ ⏞ 𝒌
⏟ − 𝑷∗
𝑷(𝟎) ⏟−
𝒆 𝒕
→𝟎
𝑷(𝟎)≷𝑷∗ 𝑫𝒆𝒄𝒂𝒚𝒔
⏟ 𝒐𝒗𝒆𝒓 𝒕𝒊𝒎𝒆
{ }
[𝑫𝒆𝒗𝒊𝒂𝒕𝒊𝒐𝒏 (+𝒗𝒆 𝒐𝒓−𝒗𝒆) ]
Therefore,
{𝑷(𝟎) − 𝑷∗ }𝒆−𝒌𝒕 + 𝑷∗
𝑷(𝒕) = ⏟
→𝟎
𝑷(𝒕) → 𝑷∗
Interpretation: It can be said that provided 𝜶, 𝜷, 𝜹, 𝜸 and 𝒋 are positive, regardless of initial price
𝑷(𝟎) is greater or smaller than equilibrium price (𝑷∗ ), the time-path of market price shall converge
towards equilibrium price as time is allowed to pass indefinitely. Hence the time-path is
dynamically stable.
______________________________________________________________________________
𝑷(𝒕) = ±𝟒𝟎𝒆−𝟒𝒕 + 𝟓𝟎
𝑷(𝒕) = 𝟒𝟎𝒆−𝟒𝒕 + 𝟓𝟎 𝑷(𝒕) = −𝟒𝟎𝒆−𝟒𝒕 + 𝟓𝟎
Complementary Complementary
Function Time-Path Function Time-Path
(Deviation) (Deviation)
𝒕 𝑷𝒄 = 𝟒𝟎𝒆−𝟒𝒕 𝑷(𝒕) = 𝟒𝟎𝒆−𝟒𝒕 + 𝟓𝟎 𝑷𝒄 = −𝟒𝟎𝒆−𝟒𝒕 𝑷(𝒕) = −𝟒𝟎𝒆−𝟒𝒕 + 𝟓𝟎
0 40 90 -40 10
0.732625555549 50.7326255555494 -0.7326255555493 49.2673744444506
1 3
0.013418505116 50.0134185051161 -0.0134185051161 49.9865814948839
2 1
0.000245768494 50.0002457684941 -0.0002457684941 49.9997542315059
3 1
0.000004501406 50.0000045014070 -0.0000045014069 49.9999954985930
4 9
0.000000082446 50.0000000824461 -0.0000000824461 49.9999999175539
5 1
0.000000001510 50.0000000015101 -0.0000000015100 49.9999999984899
6 0
0.000000000027 50.0000000000277 -0.0000000000276 49.9999999999723
7 6
0.000000000000 50.0000000000005 -0.0000000000005 49.9999999999995
8 5
… … … … …
Remarks
Positive Time-path slides Negative Time-path rises and
deviation and converges to deviation decays converges to
decays over time equilibrium (𝑷∗ = 𝟓𝟎) over time equilibrium (𝑷∗ = 𝟓𝟎)
TOPIC 082: DYNAMICS OF MARKET PRICE: NUMERICAL
Consider the Market Model:

𝑸𝒔 = 𝟑𝑷 − 𝟒
𝑸𝒅 = −𝟓𝑷 + 𝟐𝟎
______________________________________________________________________________
𝒅𝑷
= 𝟎. 𝟐(𝑸𝒅 − 𝑸𝒔 )
𝒅𝒕
Find expressions for 𝑷(𝒕), 𝑸𝒔 (𝒕) and 𝑸𝒅 (𝒕) when 𝑷(𝟎) = 𝟐. Is this system stable or
unstable?
𝒅𝑷
Solution: Given the adjustment equation = 𝟎. 𝟐(𝑸𝒅 − 𝑸𝒔 ), values of 𝑸𝒅 and 𝑸𝒔 can be
𝒅𝒕
substituted.
𝒅𝑷
= 𝟎. 𝟐 (𝑸𝒅 − 𝑸𝒔 )
𝒅𝒕
𝒅𝑷
= 𝟎. 𝟐 {(−𝟓𝑷 + 𝟐𝟎) − (𝟑𝑷 − 𝟒)}
𝒅𝒕
𝒅𝑷
= 𝟎. 𝟐{(𝟐𝟎 + 𝟒) − (𝟓𝑷 + 𝟑𝑷)}
𝒅𝒕
𝒅𝑷
= 𝟎. 𝟐{(𝟐𝟒) − (𝟖𝑷)}
𝒅𝒕
𝒅𝑷
= 𝟎. 𝟐(𝟐𝟒 − 𝟖𝑷)
𝒅𝒕
𝒅𝑷
= 𝟒. 𝟖 − 𝟏. 𝟔𝑷
𝒅𝒕
𝒅𝑷
+ 𝟏. 𝟔𝑷 = 𝟒. 𝟖
𝒅𝒕
Comparing with the standard form:
𝒅𝑷
+ 𝟏.
⏟ 𝟔 𝑷 = 𝟒.
⏟ 𝟖
⏟
𝒅𝒕 𝑎 𝑏
𝑑𝑦
𝑑𝑡
𝒅𝒚 𝒅𝑷
Here 𝒅𝒕 = 𝒅𝒕
, 𝒚 = 𝑷, 𝒂 = 𝟏. 𝟔, 𝒃 = 𝟒. 𝟖
𝒃 ≠ 𝟎: Non-Homogenous Case.
𝒂 ≠ 𝟎: Non-Homogenous Case with 𝒂 ≠ 𝟎.
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
Substituting values of 𝒚, 𝒂, 𝒃 and 𝒚(𝟎) respectively:
𝟒. 𝟖 𝟒. 𝟖
𝑷(𝒕) = (𝑷(𝟎) − ) 𝒆− 𝟏.𝟔 𝒕 +
𝟏. 𝟔 𝟏. 𝟔
−𝟏.𝟔𝒕
𝑷(𝒕) = (𝟐 − 𝟑)𝒆 +𝟑
−𝟏.𝟔𝒕
𝑷(𝒕) = (−𝟏)𝒆 +𝟑
−𝟏.𝟔𝒕
𝑷(𝒕) = −𝒆 +𝟑
It is the definite Time-Path of market price.
Dynamic Stability: The expression; 𝑷(𝒕) = −𝒆−𝟏.𝟔𝒕 + 𝟑 has an exponential decay term. i.e.
Complementary Function
Time-Path
(Deviation)
𝒕 𝑷𝒄 = −𝒆−𝟏.𝟔𝒕 𝑷(𝒕) = −𝒆−𝟏.𝟔𝒕 + 𝟑
0 -1.0000000 2.00000000
______________________________________________________________________________
1 -0.2018965 2.79810348
2 -0.0407622 2.95923779
3 -0.0082297 2.99177025
4 -0.0016616 2.99833844
5 -0.0003354 2.99966453
6 -0.0000677 2.99993227
7 -0.0000137 2.99998632
8 -0.0000027 2.99999723
9 -0.0000005 2.99999944
1 2.99999988
0 -0.0000001
… … …
Remarks
Negative deviation decays over Time-path rises and converges to equilibrium (𝑷∗ =
time 𝟑)
As 𝒕 → ∞, 𝑷𝒄 → 𝟎 ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ (𝑷∗ )
And when 𝑷𝒄 → 𝟎, (𝑷𝒕 )𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑠
Graphically
Red line shows the equilibrium value of market price (𝑷∗ ), while green line shows the
deviation from equilibrium (𝑷𝒄 ), which decreases over time and leads the time-path to
convergence.
______________________________________________________________________________
Lesson 20
PRICE DYNAMICS USING DIFFERENTIAL EQUATIONS
TOPIC 083: CONSUMER'S RESERVATION PRICE USING DIFFERENTIAL EQUATIONS
Given the direct demand function 𝑸(𝑷), if you know that the own price elasticity of demand is
−(𝟑𝑷+𝟐𝑷𝟐 )
given by 𝝐𝒅 = , where 𝑸 = 𝟓𝟎 and 𝑷 = 𝟏𝟎. What is the highest reservation price
𝑸
consumers are willing to pay?
𝒅𝑸 𝑷
Solution: The general formula of elasticity of demand: 𝝐𝒅 = × , we can rewrite the given
𝒅𝑷 𝑸
(−𝟑𝑷+𝟐𝑷𝟐 )
expression of elasticity of demand 𝝐𝒅 = 𝑸
.
𝒅𝑸 𝑷 −(𝟑𝑷 + 𝟐𝑷𝟐 )
× =
𝒅𝑷 𝑸 𝑸
𝒅𝑸 𝑷 −(𝟑 + 𝟐𝑷)𝑷
× =
𝒅𝑷 𝑸 𝑸
𝒅𝑸
= −𝟑 − 𝟐𝑷
𝒅𝑷
𝒅𝑸
+ 𝟐 𝑷 = −𝟑
𝒅𝑷
Standard form requires dependent variable 𝑸 . Instead, we have the independent variable 𝑸 .
So, we cannot use the usual formula for first order differential equation. An alternative approach
to solving the differential equation is to integrate it, so the derivative gets cancelled with integral
and value of dependent variable (𝑸) comes out as a result.
𝒅𝑸
Keeping derivative term 𝒅𝑷 on L.H.S.
𝒅𝑸
+ 𝟐 𝑷 = −𝟑
𝒅𝑷
𝒅𝑸
+ 𝟐𝑷 = −𝟑
𝒅𝑷
𝒅𝑸
= −𝟐𝑷 − 𝟑
𝒅𝑷
Integrating the result w.r.t. 𝑷:
𝒅𝑸
∫ 𝒅𝑷 = ∫(−𝟐𝑷 − 𝟑) 𝒅𝑷
𝒅𝑷
𝒅𝑸
∫ 𝒅𝑷 = ∫(−𝟐𝑷) 𝒅𝑷 − ∫(𝟑) 𝒅𝑷
𝒅𝑷
𝑸 = −𝟐 ∫(𝑷) 𝒅𝑷 − 𝟑 ∫(𝑷𝟎 ) 𝒅𝑷
𝑷𝟐
𝑸 = −𝟐 ( ) − 𝟑𝑷 + 𝒄
𝟐
𝑸(𝑷) = −𝑷𝟐 − 𝟑𝑷 + 𝒄
General Solution as 𝒄 is still indefinite.
Using the given information that at equilibrium 𝑸 = 𝟓𝟎 if 𝑷 = 𝟏𝟎.
Putting the given information in General Solution.
𝟓𝟎 = −(𝟏𝟎)𝟐 − 𝟑(𝟏𝟎) + 𝒄
______________________________________________________________________________
𝟓𝟎 = −𝟏𝟎𝟎 − 𝟑𝟎 + 𝒄
𝟓𝟎 + 𝟏𝟎𝟎 + 𝟑𝟎 = 𝒄
𝒄 = 𝟏𝟖𝟎
Substituting the value of 𝒄 = 𝟏𝟖𝟎 back into the General Solution; 𝑸(𝑷) = −𝑷𝟐 − 𝟑𝑷 + 𝒄.
𝑸(𝑷) = −𝑷𝟐 − 𝟑𝑷 + 𝟏𝟖𝟎
We get the Definite Solution.
Reservation Price: Price at which no demand for good/service doesn’t start.
Mathematically speaking, reservation price is: 𝑷𝑹 = (𝑷)𝑸=𝟎 . Therefore, putting 𝑸 = 𝟎 in the
Definite Solution; 𝑸 = −𝑷𝟐 − 𝟑𝑷 + 𝟏𝟖𝟎:
𝟎 = −𝑷𝟐 − 𝟑𝑷 + 𝟏𝟖𝟎
−𝑷𝟐 − 𝟑𝑷 + 𝟏𝟖𝟎 = 𝟎
𝑷𝟐 + 𝟑𝑷 − 𝟏𝟖𝟎 = 𝟎
Quadratic Formula is applicable: 𝒙 = 𝑷, 𝒂 = 𝟏, 𝒃 = 𝟑, 𝒄 = −𝟏𝟖𝟎
−𝒃 ± √𝒃𝟐 − 𝟒𝒂𝒄
𝒙=
𝟐𝒂
−𝟑 ± √(𝟑)𝟐 − 𝟒(𝟏)(−𝟏𝟖𝟎)
𝑷=
𝟐(𝟏)
−𝟑 ± √𝟕𝟐𝟗
𝑷=
𝟐(𝟏)
−𝟑 ± 𝟐𝟕
𝑷=
𝟐(𝟏)
Either
−𝟑+𝟐𝟕 −𝟑−𝟐𝟕
𝑷 = 𝟐 OR 𝑷 = 𝟐
𝟐𝟒 −𝟑𝟎
𝑷 = OR 𝑷 =
𝟐 𝟐
𝑷 = 𝟏𝟐 OR 𝑷 = −𝟏𝟓
Dispelling the ‘unrealistic’ negative value of 𝑷, we get the Reservation Price (𝑷𝑹 ).
𝑷𝑹 = 𝟏𝟐
Verification: Since Reservation Price is one at which demand doesn’t starts (𝑸 = 𝟎). Therefore,
putting Reservation Price should make demand equal to zero:
𝑸(𝑷𝑹 ) = 𝟎
𝑸(𝑷𝑹 ) = −𝑷𝑹 𝟐 − 𝟑𝑷𝑹 + 𝟏𝟖𝟎
𝑸(𝟏𝟐) = −(𝟏𝟐)𝟐 − 𝟑(𝟏𝟐) + 𝟏𝟖𝟎
𝑸(𝟏𝟐) = −𝟏𝟒𝟒 − 𝟑𝟔 + 𝟏𝟖𝟎
𝑸(𝟏𝟐) = −𝟏𝟖𝟎 + 𝟏𝟖𝟎
𝑸(𝟏𝟐) = 𝟎
Interpretation: Hence its evident that when price is 𝟏𝟐, demand for the good doesn’t exist.
TOPIC 084: DYNAMICS OF PRICE & MONEY MARKET USING DIFFERENTIAL EQUATIONS
Money supply 𝑴𝒔 is fixed at a level determined by the government.

Transaction demand for money 𝑳 is positively related to prices 𝑷(𝒕) and real output 𝒀:
𝑳 = 𝒇{𝑷(𝒕) × 𝒀}
As given that the relationship is positive – direct proportionality (∝) exist.
𝑳 ∝ {𝑷(𝒕) × 𝒀}
Introducing the constant of proportionality (𝜶) to develop equation.
𝑳 = 𝜶𝑷𝒀 [𝜶 > 𝟎]
______________________________________________________________________________
Constant of Proportionality is positive which will keep the relationship between variables
positive.
𝒅𝑷
If inflation is the increase in the overall price level with time 𝒅𝒕 , it depends on excess money
supply. Excess Money Supply is the difference between Money-Supply and Transaction Demand
for Money. The model then becomes:
𝒅𝑷
= 𝒋(𝑴𝒔 − 𝑳) [𝒋 > 𝟎]
𝒅𝒕
𝒋 > 𝟎 implies that price level rises with excess money-supply.
Time-path of the price level in the economy can be found by substituting values in the given
information.
𝒅𝑷
= 𝒋 (𝑴𝒔 − 𝑳)
𝒅𝒕
𝒅𝑷
= 𝒋 (𝑴𝒔 − 𝜶𝑷𝒀)
𝒅𝒕
𝒅𝑷
= 𝒋(𝑴𝒔 − 𝜶𝑷𝒀)
𝒅𝒕
𝒅𝑷
= 𝒋𝑴𝒔 − 𝜶𝒋𝒀𝑷
𝒅𝒕
𝒅𝑷
+ 𝜶𝒋𝒀𝑷 = 𝒋𝑴𝒔
𝒅𝒕
𝒅𝒚
Comparing with the standard form of the differential equation; 𝒅𝒕 + 𝒂𝒚 = 𝒃
𝒅𝑷
+ 𝜶𝒋𝒀 𝑷⏟ = 𝒋𝑴
⏟ ⏟
𝒅𝒕
⏟𝒔
𝑎 𝑏𝑦
𝑑𝑦
𝑑𝑡
𝒅𝒚 𝒅𝑷
Here 𝒅𝒕 = 𝒅𝒕 , 𝒚 = 𝑷, 𝒂 = 𝜶𝒋𝒀, 𝒃 = 𝒋𝑴𝒔
𝒃 = 𝒋𝑴𝒔 ; as 𝒋 is positive constant, and money-supply (𝑴𝒔 ) is fixed (some positive number)
by the government, the product will be non-zero positive constant. i.e. 𝒃 ≠ 𝟎 – Non-
Homogenous Case.
𝒂 = 𝜶𝒋𝒀; as 𝜶 and 𝒋 are positive constants, and 𝒀 is national income (some positive
number), the value of 𝒂 shall be some positive (non-zero) constant. i.e. 𝒂 ≠ 𝟎 – Non-
Homogenous Case with 𝒂 ≠ 𝟎.
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
𝒋𝑴𝒔 − 𝜶𝒋𝒀 𝒕 𝒋𝑴𝒔
𝑷(𝒕) = {𝑷(𝟎) − }𝒆 +
𝜶𝒋𝒀 𝜶𝒋𝒀
𝒋𝑴𝒔 −(𝜶𝒋𝒀)𝒕 𝒋𝑴𝒔
𝑷(𝒕) = {𝑷(𝟎) − }𝒆 +
𝑴𝒔 −(𝜶𝒋𝒀)𝒕 𝑴𝒔
𝑷(𝒕) = {𝑷(𝟎) − }𝒆 +
𝜶𝒀 𝜶𝒀
Definite Time-path for Market price in Dynamic Sense.
______________________________________________________________________________
Interpretation: Since, 𝑴𝒔 appears in numerator, it has positive relationship with price level. 𝒀 is
in the denominator, and has negative relationship with price level. If money-supply increases in
economy, it will increase the prices in the economy. If output in the economy increases, it shall
decrease the price level due to increased supply of goods/services.
Dynamic Stability: For the dynamic stability of the time-path of First Order Differential Equation,
it should converge towards equilibrium. We know that equilibrium is particular solution (𝑷𝒑 ).
Whereas convergence happens when deviation from equilibrium fades to zero. In a differential
equation, deviation component is measured by (𝑷𝒄 ). Therefore, for dynamic stability; 𝑷𝒄 → 𝟎 as
𝒕 → ∞.
Using the Definite Time-path for Price level in Dynamic Sense.
𝑷(𝒕) = {𝑷(𝟎) − }𝒆 +
𝜶𝒀 𝜶𝒀
𝑷(𝒕) = {𝑷(𝟎) − }𝒆 +
⏟ 𝜶𝒀 ⏟
𝜶𝒀
𝑷𝒄 𝑃𝑝
Need for time path of price {𝑷(𝒕)} to converge to equilibrium price (𝑷∗ ) i.e. 𝑷(𝒕) → 𝑷∗ as 𝑷𝒄 →
𝟎 when 𝒕 → ∞.
𝑷𝒄 → 𝟎
𝑴𝒔 −(𝜶𝒋𝒀)𝒕
{𝑷(𝟎) − }𝒆 →𝟎
𝜶𝒀
{𝑴𝒔 ,𝜶,𝒀>𝟎}
{𝜶,𝒋,𝒀>𝟎}
⏞ 𝑴𝒔 −⏞ 𝜶𝒋𝒀 𝒕
{𝑷(𝟎) − }⏟
𝒆 →𝟎
⏟ 𝜶𝒀
𝑬𝒙𝒑𝒐𝒏𝒆𝒏𝒕𝒊𝒂𝒍
⏟ 𝑷(𝟎)≷𝑷𝒑 𝑫𝒆𝒄𝒂𝒚
[ 𝑫𝒆𝒗𝒊𝒂𝒕𝒊𝒐𝒏 (+𝒗𝒆 𝒐𝒓−𝒗𝒆) 𝒐𝒗𝒆𝒓 𝒕𝒊𝒎𝒆 ]
Therefore,
𝑷(𝒕) = {𝑷(𝟎) − }𝒆 +
⏟ 𝜶𝒀 𝜶𝒀
→𝟎
𝑷(𝒕) = {𝑷(𝟎)
⏟ − 𝑷𝒑 } 𝒆−(𝜶𝒋𝒀)𝒕 + 𝑷𝒑
→𝟎
{𝑷(𝟎) − 𝑷∗ }𝒆−(𝜶𝒋𝒀)𝒕 + 𝑷∗
𝑷(𝒕) = ⏟
→𝟎
𝑷(𝒕) → 𝑷∗
Interpretation: It can be said that provided 𝜶, 𝒋, 𝑴𝒔 and 𝒀 are positive, regardless of initial price
𝑷(𝟎) is greater or smaller than equilibrium price (𝑷∗ ), the time-path of market price shall converge
towards equilibrium price as time is allowed to pass indefinitely. Hence the time-path is
dynamically stable.
TOPIC 085: PRICE DYNAMICS & EQUATION OF EXCHANGE USING DIFFERENTIAL

EQUATIONS
Borrowing the equation of exchange:

𝑴𝒔 × 𝑽 = 𝑷 × 𝒀
𝑴𝒔 is the Money Supply, 𝑽 is the velocity of Money, 𝑷 is general price level in the economy and
𝒀 shows the Aggregate Output.
Rewriting for 𝑴𝒔 :
______________________________________________________________________________
𝑷×𝒀
𝑴𝒔 =
𝑽
Other side of money i.e. demand for money 𝑳 is assumed to have transactive component only
and is considered constant.
The price level shall make adjustments due to differences between (transaction) demand and
supply of the money (𝑳 ≠ 𝑴𝒔 ):
𝒅𝑷
= 𝒋( 𝑴𝒔 − 𝑳
⏟ )
⏟
𝒅𝒕 𝑬𝒙𝒄𝒆𝒔𝒔
𝑷𝒓𝒊𝒄𝒆 𝑨𝒅𝒋𝒖𝒔𝒕𝒎𝒆𝒏𝒕 𝑴𝒐𝒏𝒆𝒚−𝑺𝒖𝒑𝒑𝒍𝒚
𝒐𝒗𝒆𝒓 𝒕𝒊𝒎𝒆
The price adjustments over time are directly proportional to Excess Money- supply.
𝒋 is the adjustment coefficient which is positive (𝒋 > 𝟎) that implies that price level rises with
excess money-supply.
Time-path of the price level in the economy can be found by substituting values in the 𝑴𝒔 .
𝒅𝑷
= 𝒋( 𝑴𝒔 − 𝑳)
𝒅𝒕
𝒅𝑷 𝑷×𝒀
= 𝒋( − 𝑳)
𝒅𝒕 𝑽
𝒅𝑷 𝑷×𝒀
= 𝒋( − 𝑳)
𝒅𝒕 𝑽
𝒅𝑷 𝒀
= 𝒋 ( ) 𝑷 − 𝒋𝑳
𝒅𝒕 𝑽
𝒅𝑷 𝒀
− 𝒋 ( ) 𝑷 = −𝒋𝑳
𝒅𝒕 𝑽
𝒅𝒚 𝒅𝑷 𝒀
Here 𝒅𝒕 = 𝒅𝒕 , 𝒚 = 𝑷, 𝒂 = −𝒋 (𝑽) , 𝒃 = −𝒋𝑳
𝒃 = −𝒋𝑳 ; as 𝒋 is positive constant, and (transaction) demand for money (𝑳) is fixed (some
positive number), the product will be non-zero positive constant. However, 𝒃 = − 𝒋𝑳, so the
final answer will be negative but still not zero i.e. 𝒃 ≠ 𝟎 – Non-Homogenous Case.
𝒀
𝒂 = −𝒋 (𝑽); as 𝒋 is positive constant. 𝒀 is Aggregate Output (some positive number) and 𝑽 is
𝒀
velocity of money (some positive number). The ratio (𝑽) when multiplied by 𝒋 will give a
𝒀
positive answer. However, 𝒂 = − 𝒋 (𝑽), so the final answer will be negative but still not zero.
i.e. 𝒂 ≠ 𝟎 – Non-Homogenous Case with 𝒂 ≠ 𝟎.
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
−𝒋𝑳 𝒀
− (−𝒋 ) 𝒕 −𝒋𝑳
𝑷(𝒕) = 𝑷(𝟎) − 𝒆 𝑽 +
𝒀 𝒀
(−𝒋 𝑽) (−𝒋 𝑽)
{ }
______________________________________________________________________________
−𝒋𝑳 𝒀
−(−𝒋 )𝒕 −𝒋𝑳
𝑷(𝒕) = {𝑷(𝟎) − }𝒆 𝑽 +
𝒀 𝒀
(−𝒋 𝑽) (−𝒋 𝑽)
−𝑳 𝒀
−(− )𝒕 −𝑳
𝑷(𝒕) = {𝑷(𝟎) − }𝒆 𝑽 +
𝒀 𝒀
(− 𝑽) (− 𝑽)
𝑳 𝒀
( )𝒕 𝑳
𝑷(𝒕) = {𝑷(𝟎) − }𝒆 𝑽 +
𝒀 𝒀
(𝑽) (𝑽)
𝒀
Reciprocalizing (𝑽):
𝑽𝑳 (𝒀)𝒕 𝑽𝑳
𝑷(𝒕) = {𝑷(𝟎) − } 𝒆 𝑽 +
𝒀 𝒀
𝑷(𝒕) = {𝑷(𝟎) − } 𝒆 𝑽 +
𝒀 𝒀
Definite Time-path for Market price in Dynamic Sense.

Interpretation: Since, 𝑽 and 𝑳 appear in numerator, they have positive relationship with price
level. 𝒀 is in the denominator, and has negative relationship with price level:
Velocity of Money (𝑽): The faster money circulates in the economy, the higher the prices;
(𝑷 ∝ 𝑽).
Transaction Demand for Money (𝑳): The higher the transaction demand for money, the higher
the price level is; (𝑷 ∝ 𝑳).
Aggregate Output (𝒀): If the economy produces greater aggregate output, this will reduce prices;
𝟏
(𝑷 ∝ ).
𝒀
it should converge towards equilibrium. We know that equilibrium is particular solution (𝑷𝒑 ).
equation, deviation component is measured by (𝑷𝒄 ). Therefore, for dynamic stability; 𝑷𝒄 → 𝟎 as
𝒕 → ∞.
Using the Definite Time-path for Price level in Dynamic Sense:
𝑷(𝒕) = {𝑷(𝟎) − } 𝒆 𝑽 +
⏟ 𝒀 ⏟
𝒀
𝑷𝒄 𝑷𝒑
Need for time path of price {𝑷(𝒕)} to converge to equilibrium price (𝑷∗ ) i.e. 𝑷(𝒕) → 𝑷∗ as 𝑷𝒄 → 𝟎
when 𝒕 → ∞.
𝑷𝒄 → 𝟎
𝑽𝑳 (𝒀)𝒕
{𝑷(𝟎) − } 𝒆 𝑽 → 𝟎
𝒀
However, on hair splitting, we don’t get the desired result.
{𝑽,𝑳,𝒀>𝟎} {𝒀,𝑽>𝟎}
⏞
⏞𝑽𝑳 𝒀
( ) 𝒕
{𝑷(𝟎) − }⏟
𝒆 𝑽
→∞
⏟ 𝒀
⏟ 𝑷(𝟎)≷𝑷𝒑 𝑮𝒓𝒐𝒘𝒕𝒉
[𝑫𝒆𝒗𝒊𝒂𝒕𝒊𝒐𝒏 (+𝒗𝒆 𝒐𝒓−𝒗𝒆) 𝒐𝒗𝒆𝒓 𝒕𝒊𝒎𝒆 ]
Therefore,
______________________________________________________________________________
𝑷(𝒕) = {𝑷(𝟎) − }𝒆 𝑽 +
⏟ 𝒀 𝒀
→∞
𝒀
( )𝒕
𝑷(𝒕) = {𝑷(𝟎)
⏟ − 𝑷𝒑 } 𝒆 𝑽 + 𝑷𝒑
→∞
𝒀
( )𝒕
{𝑷(𝟎) − 𝑷∗ }𝒆
𝑷(𝒕) = ⏟ 𝑽 + 𝑷∗
→∞
Since addition of (𝑷∗ ) in ∞ shall have no effect on infinity;
𝑷(𝒕) → ∞
𝑷(𝒕) ↛ 𝑷∗
Interpretation: It can be said that provided 𝑽, 𝒀 and 𝑳 are positive, regardless of initial price 𝑷(𝟎)
is greater or smaller than equilibrium price (𝑷∗ ), the time-path of market price shall diverge away
from equilibrium price as time is allowed to pass indefinitely. Hence the time-path is
dynamically unstable.
______________________________________________________________________________
Lesson 21
INCOME DYNAMICS USING DIFFERENTIAL EQUATIONS
TOPIC 086: INCOME DYNAMICS & MONEY MARKET USING DIFFERENTIAL EQUATIONS
On the money market, money supply 𝑴𝒔 is fixed by the government. Furthermore, transaction
demand for money 𝑳 depends on income but is also positively related to its incremental increase
so that:
𝒅𝒀
𝑳 = 𝒇 (𝒀, )
𝒅𝒕
𝒅𝒀
𝑳=𝜶 ⏟
𝒀 +𝜷
𝑁𝑎𝑡𝑖𝑜𝑛𝑎𝑙
⏟
𝒅𝒕
𝐼𝑛𝑐𝑜𝑚𝑒 𝐼𝑛𝑐𝑟𝑒𝑚𝑒𝑛𝑡𝑎𝑙
𝑖𝑛𝑐𝑟𝑒𝑎𝑠𝑒 𝑖𝑛
𝑛𝑎𝑡𝑖𝑜𝑛𝑎𝑙
𝑖𝑛𝑐𝑜𝑚𝑒
𝒅𝒀
𝑳 = 𝜶𝒀 + 𝜷 𝒅𝒕 𝟎 < 𝜶 < 𝟏, 𝜷 > 𝟎
Find the time path of aggregate national income that ensures equilibrium on the money market,
and check whether this time path is dynamically stable.
Solution: Since equilibrium in money market is to be ensured in finding the time-path for
aggregate national income, we equate demand and supply of money:
𝑳 = 𝑴𝒔
𝒅𝒀
𝜶𝒀 + 𝜷 = 𝑴𝒔
𝒅𝒕
𝒅𝒀
𝜶𝒀 + 𝜷 = 𝑴𝒔
𝒅𝒕
Rearranging for developing the first order differential equation:
𝒅𝒀
𝜷 + 𝜶𝒀 = 𝑴𝒔
𝒅𝒕
𝒅𝒀
Dividing by 𝜷 to get the 𝒅𝒕 alone as found in first order differential equation.
𝒅𝒀 𝜶 𝑴𝒔
+ 𝒀=
𝒅𝒕 𝜷 𝜷
𝒅𝒚
Comparing with the standard form of first order differential equation 𝒅𝒕 + 𝒂𝒚 = 𝒃.
𝒅𝒀 𝜶 𝑴𝒔
+ 𝒀=
⏟
𝒅𝒕 𝜷 ⏟ 𝜷
⏟
𝑑𝑦 𝑎 𝑏
𝑑𝑡
𝒅𝒚 𝒅𝒀 𝜶 𝑴𝒔
Here 𝒅𝒕
= 𝒅𝒕
,𝒚 = 𝒀, 𝒂 = 𝜷 , 𝒃 = 𝜷
𝑴𝒔
𝒃= 𝜷
; as money-supply (𝑴𝒔 ) is fixed (some positive number) by the government 𝜷 > 𝟎 (positive
constant), and the ratio will be non-zero positive constant. i.e. 𝒃 ≠ 𝟎 – Non-Homogenous Case.
𝜶
𝒂 = 𝜷; as 𝟎 < 𝜶 < 𝟏 and 𝜷 > 𝟎, both can be considered positive constants, their ratio is the
value of 𝒂 which will also be some positive (non-zero) value. i.e. 𝒂 ≠ 𝟎 – Non-Homogenous
Case with 𝒂 ≠ 𝟎.
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
______________________________________________________________________________

𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
𝑴𝒔 𝑴𝒔
𝜷 𝜶 𝜷
− 𝒕
𝒀(𝒕) = 𝒀(𝟎) − 𝒆 𝜷 +
𝜶 𝜶
𝜷 𝜷
{ }
𝑴𝒔 𝑴𝒔
𝜶
𝜷 −( )𝒕 𝜷
𝒀(𝒕) = {𝒀(𝟎) − 𝜶 } 𝒆 𝜷 + 𝜶
𝜷 𝜷
𝑴𝒔 −(𝜶𝜷)𝒕 𝑴𝒔
𝒀(𝒕) = {𝒀(𝟎) − }𝒆 +
𝜶 𝜶
Definite Time-path for National Income in Dynamic Sense.
Interpretation: Since, 𝑴𝒔 appears in numerator, it has positive relationship with aggregate
national income 𝒀.
it should converge towards equilibrium. We know that equilibrium is particular solution (𝒀𝒑 ).
equation, deviation component is measured by (𝒀𝒄 ). Therefore, for dynamic stability; 𝒀𝒄 → 𝟎 as
𝒕 → ∞.
Using the Definite Time-path for Income level in Dynamic Sense.
𝒀(𝒕) = {𝒀(𝟎) − }𝒆 +
𝜶 𝜶
𝒀(𝒕) = {𝒀(𝟎) − }𝒆 +
⏟ 𝜶 ⏟
𝜶
𝒀𝒄 𝑌𝑝
Need for time path of income {𝒀(𝒕)} to converge to equilibrium income (𝒀∗ ) i.e. 𝒀(𝒕) → 𝒀∗ as
𝒀𝒄 → 𝟎 when 𝒕 → ∞.
𝒀𝒄 → 𝟎
𝑴𝒔 −(𝜶𝜷)𝒕
{𝒀(𝟎) − }𝒆 →𝟎
𝜶
{𝑴𝒔 ,𝜶 >𝟎} {𝜶,𝜷>𝟎}
⏞ 𝜶
⏞ 𝑴𝒔 − ( ) 𝒕
𝜷
{𝒀(𝟎) − } ⏟
𝒆 →𝟎
⏟ 𝜶
⏟ 𝒀(𝟎)≷𝒀𝒑 𝑫𝒆𝒄𝒂𝒚
[𝑫𝒆𝒗𝒊𝒂𝒕𝒊𝒐𝒏 (+𝒗𝒆 𝒐𝒓−𝒗𝒆) 𝒐𝒗𝒆𝒓 𝒕𝒊𝒎𝒆 ]
Therefore,
𝒀(𝒕) = {𝒀(𝟎) − }𝒆 +
⏟ 𝜶 𝜶
→𝟎
𝜶
−( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − 𝒀𝒑 } 𝒆 𝜷 + 𝒀𝒑
⏟
→𝟎
As particular integral is also the equilibrium value of the variable 𝒀𝒑 = 𝒀∗ .
______________________________________________________________________________
𝜶
−( )𝒕
𝒀(𝒕) = {𝒀(𝟎)
⏟ − 𝒀∗ }𝒆 𝜷 + 𝒀∗
→𝟎
𝒀(𝒕) → 𝒀∗
Interpretation: It can be said that provided 𝜶, 𝜷 and 𝑴𝒔 are positive, regardless of initial
income 𝒀(𝟎) is greater or smaller than equilibrium price (𝒀∗ ), the time-path of national income
shall converge towards equilibrium income level as time is allowed to pass indefinitely.
Hence the time-path is dynamically stable.
TOPIC 087: INCOME DYNAMICS & INCOME TAX USING DIFFERENTIAL EQUATIONS
If national income model is modelled by following system of equations:

𝒀=𝑪+𝑰
𝑪 = 𝜶 + 𝜷𝒀 𝜶 > 𝟎, 𝟎 < 𝜷
𝜶 is the autonomous consumption and 𝜷 is the marginal propensity to consume.
𝒅𝒀
𝑰 = 𝜹𝒀 + 𝜸 ( ) 𝜸 > 𝟎, 𝜹 < 𝟏
𝒅𝒕
If an income tax is introduced, the Consumption function becomes:
𝑪 = 𝜶 + 𝜷(𝟏 − 𝝉)𝒀 𝝉<𝟏
𝝉 is the income tax rate. Find the time path and the intertemporal equilibrium level of national
income.
Solution: Expenditure approach to National income is:
𝒀=𝑪+𝑰
Substituting values of 𝑪 and 𝑰.
𝒅𝒀
𝒀 = 𝜶 + 𝜷(𝟏 − 𝝉)𝒀 + 𝜹𝒀 + 𝜸 ( )
𝒅𝒕
𝒅𝒀
𝒀 = 𝜶 + 𝜷(𝟏 − 𝝉)𝒀 + 𝜹𝒀 + 𝜸 ( )
𝒅𝒕
Rearranging to conform to a first order differential equation:
𝒅𝒀
−𝜸 ( ) − 𝜷(𝟏 − 𝝉)𝒀 − 𝜹𝒀 + 𝒀 = 𝜶
𝒅𝒕
𝒅𝒀
−𝜸 ( ) − 𝜷𝒀 + 𝝉𝜷𝒀 − 𝜹𝒀 + 𝒀 = 𝜶
𝒅𝒕
𝒅𝒀
−𝜸 ( ) + 𝒀 − 𝜷𝒀 + 𝝉𝜷𝒀 − 𝜹𝒀 = 𝜶
𝒅𝒕
𝒅𝒀
−𝜸 ( ) + (𝟏 − 𝜷 + 𝝉𝜷 − 𝜹)𝒀 = 𝜶
𝒅𝒕
Dividing by −𝜸 throughout to normalize the differential equation.
𝜸 𝒅𝒀 (𝟏 − 𝜷 + 𝝉𝜷 − 𝜹) 𝜶
− ( )+ 𝒀=
(−𝜸) 𝒅𝒕 (−𝜸) (−𝜸)
𝒅𝒀 𝟏 − 𝜷 + 𝝉𝜷 − 𝜹 𝜶
( )+( )𝒀 = −
𝒅𝒕 −𝜸 𝜸
𝒅𝒀 𝟏 − 𝜷 + 𝝉𝜷 − 𝜹 𝜶
( )+( )𝒀 = −
𝒅𝒚
Comparing with the standard form of first order differential equation; 𝒅𝒕 + 𝒂𝒚 = 𝒃.
𝒅𝒀 𝟏 − 𝜷 + 𝝉𝜷 − 𝜹 𝜶
( )+( ) 𝒀 = (− )
⏟𝒅𝒕 ⏟ −𝜸 ⏟ 𝜸
𝒅𝒚 𝑎 𝑏
𝒅𝒕
______________________________________________________________________________
𝒅𝒚 𝒅𝒀 𝟏−𝜷+𝝉𝜷−𝜹 𝜶
Here 𝒅𝒕
= 𝒅𝒕
,𝒚 = 𝒀, 𝒂 = ( −𝜸
),𝒃 = (− 𝜸)
𝜶
𝒃= (− 𝜸); As 𝜶 > 𝟎 and 𝜸 > 𝟎 (both positive constants), their ratio will be non-zero positive
𝜶
constant. However, the value of 𝒃 = (− 𝜸) has a negative sign. The answer will be negative,
though still non-zero. i.e. 𝒃 ≠ 𝟎 – Non-Homogenous Case.
𝟏−𝜷+𝝉𝜷−𝜹
𝒂=( ); the numerator (𝟏 − 𝜷 + 𝝉𝜷 − 𝜹) can be non-zero if (𝜷 + 𝝉𝜷 − 𝜹) ≠ 𝟏. Also 𝜸 is
−𝜸
positive value. Therefore, value of 𝒂 will be some positive/negative (non-zero) value. i.e. 𝒂 ≠ 𝟎 –
Non-Homogenous Case with 𝒂 ≠ 𝟎.
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
𝜶 𝜶
(− ) 𝟏−𝜷+𝝉𝜷−𝜹 (− )
𝜸 −(
−𝜸
)𝒕 𝜸
𝒀(𝒕) = 𝒀(𝟎) − 𝒆 +
𝟏 − 𝜷 + 𝝉𝜷 − 𝜹 𝟏 − 𝜷 + 𝝉𝜷 − 𝜹
( ) ( )
{ −𝜸 } −𝜸
𝜶 𝜶
( − 𝜸) −(
)𝒕
( − 𝜸)
𝒀(𝒕) = {𝒀(𝟎) − }𝒆 −𝜸 +
𝟏 − 𝜷 + 𝝉𝜷 − 𝜹 𝟏 − 𝜷 + 𝝉𝜷 − 𝜹
( −𝜸 ) ( −𝜸 )
(𝜶) 𝟏−𝜷+𝝉𝜷−𝜹 (𝜶)
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − }𝒆 𝜸 +
(𝟏 − 𝜷 + 𝝉𝜷 − 𝜹) (𝟏 − 𝜷 + 𝝉𝜷 − 𝜹)
𝜶 𝟏−𝜷+𝝉𝜷−𝜹 𝜶
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − ( )} 𝒆 𝜸 +( )
𝟏 − 𝜷 + 𝝉𝜷 − 𝜹 𝟏 − 𝜷 + 𝝉𝜷 − 𝜹

Interpretation: As expected the national income (𝒀) is negatively related to tax rate (𝝉). Its due
𝟏
to presence of 𝝉 in the denominator of the expression. Therefore, 𝒀 ∝ 𝝉 .
𝒕 → ∞.
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − ( )} 𝒆 𝜸 +( )
⏟ 𝟏 − 𝜷 + 𝝉𝜷 − 𝜹 ⏟𝟏 − 𝜷 + 𝝉𝜷 − 𝜹
𝑌𝑐 𝑌𝑝
Need for time path of income {𝒀(𝒕)} to converge to equilibrium income (𝒀∗ ) i.e. 𝒀(𝒕) → 𝒀∗ as 𝒀𝒄 →
𝟎 when 𝒕 → ∞.
𝒀𝒄 → 𝟎
______________________________________________________________________________
𝜶 𝟏−𝜷+𝝉𝜷−𝜹
( )𝒕
{𝒀(𝟎) − ( )} 𝒆 𝜸 →𝟎
𝟏 − 𝜷 + 𝝉𝜷 − 𝜹
{𝜶,𝜷,𝜹,𝝉>𝟎} {𝜷,𝜹,𝜸,𝝉>𝟎}
⏞𝟏−𝜷+𝝉𝜷−𝜹
⏞ 𝜶 𝜸
( ) 𝒕
𝒀(𝟎) − (𝟏−𝜷+𝝉𝜷−𝜹) ⏟ 𝒆 →𝟎
⏟ 𝑬𝒙𝒑𝒐𝒏𝒆𝒏𝒕𝒊𝒂𝒍
{
⏟ 𝒀(𝟎)≷𝒀𝒑 } 𝑮𝒓𝒐𝒘𝒕𝒉
[ 𝑫𝒆𝒗𝒊𝒂𝒕𝒊𝒐𝒏 (+𝒗𝒆 𝒐𝒓−𝒗𝒆) ]
Therefore,
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − ( )} 𝒆 𝜸 +( )
⏟ 𝟏 − 𝜷 + 𝝉𝜷 − 𝜹 𝟏 − 𝜷 + 𝝉𝜷 − 𝜹
→∞
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − ( ) }𝒆 𝜸 + ( )
⏟ 𝟏 − 𝜷 + 𝝉𝜷 − 𝜹 𝟏 − 𝜷 + 𝝉𝜷 − 𝜹
→∞
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − 𝒀𝒑 } 𝒆 𝜸 + 𝒀𝒑
⏟
→∞
( )𝒕
𝒀(𝒕) = {𝒀(𝟎)
⏟ − 𝒀∗ }𝒆 𝜸 + 𝒀∗
→∞
Since addition of (𝒀∗ ) in ∞ shall have no effect on infinity;
𝒀(𝒕) → ∞
𝒀(𝒕) ↛ 𝒀∗
Interpretation: It can be said that provided 𝜷 > 𝟎, 𝜹 < 𝟏, 𝜸 > 𝟎 and 𝝉 < 𝟏, regardless of initial
national income 𝒀(𝟎) is greater or smaller than equilibrium national income (𝒀∗ ), the time-path of
national income shall diverge away from equilibrium level of national income as time is allowed
to pass indefinitely. Hence the time-path is dynamically unstable.
TOPIC 088: DYNAMICS OF CONSUMPTION USING DIFFERENTIAL EQUATIONS
𝒅𝑪
The consumption function of an individual grows with time at a rate given by the equation: 𝒅𝒕 =
−𝟐𝑪 + 𝟏𝟎𝟎. Find the time path of this person’s consumption given that 𝑪(𝟎) = 𝟏𝟎.
𝒅𝑪
Solution: Given the equation for rate of consumption 𝒅𝒕 = −𝟐𝑪 + 𝟏𝟎𝟎; we can rearrange to
compare with the standard form of differential equation.
𝒅𝑪
+ 𝟐𝑪 = 𝟏𝟎𝟎
𝒅𝒕
𝒅𝑪
+⏟ 𝟐 𝑪 = 𝟏𝟎𝟎
⏟
⏟
𝒅𝒕 𝑎 𝑏
𝑑𝑦
𝑑𝑡
𝒅𝒚 𝒅𝑪
Here = , 𝒚 = 𝑪, 𝒂 = 𝟐, 𝒃 = 𝟏𝟎𝟎, 𝒚(𝟎) = 𝑪(𝟎) = 𝟏𝟎
𝒅𝒕 𝒅𝒕
As 𝒃 = 𝟏𝟎𝟎, so 𝒃 ≠ 𝟎: Non-Homogenous Case.
As 𝒂 = 𝟐, so 𝒂 ≠ 𝟎: Non-Homogenous Case with 𝒂 ≠ 𝟎.
______________________________________________________________________________
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
𝟏𝟎𝟎 𝟏𝟎𝟎
𝑪(𝒕) = (𝑪(𝟎) − ) 𝒆− 𝟐 𝒕 +
𝟐 𝟐
𝟏𝟎𝟎 −𝟐𝒕 𝟏𝟎𝟎
𝑪(𝒕) = {𝟏𝟎 − }𝒆 +
𝟐 𝟐
𝑪(𝒕) = {𝟏𝟎 − 𝟓𝟎}𝒆−𝟐𝒕 + 𝟓𝟎
𝑪(𝒕) = {𝟏𝟎 − 𝟓𝟎}𝒆−𝟐𝒕 + 𝟓𝟎
𝑪(𝒕) = −𝟒𝟎𝒆−𝟐𝒕 + 𝟓𝟎
It is the definite Time-Path of Consumption of the individual.
−𝟒𝟎𝒆−𝟐𝒕 + 𝟓𝟎
𝑪(𝒕) = ⏟ ⏟
𝑪𝒄 𝑪𝒑
−𝟐𝒕
Dynamic Stability: The expression; 𝑪(𝒕) = −𝟒𝟎𝒆 + 𝟓𝟎 has an exponential decay term in the
box. i.e.:
______________________________________________________________________________
Time-Path
(Deviation)
𝒕 𝑪𝒄 = −𝟒𝟎𝒆−𝟐𝒕 𝑪(𝒕) = −𝟒𝟎𝒆−𝟐𝒕 + 𝟓𝟎
0 -40 10
1 -5.41341133 44.58658867
2 -0.73262556 49.26737444
3 -0.09915009 49.90084991
4 -0.01341851 49.98658149
5 -0.00181600 49.99818400
6 -0.00024577 49.99975423
7 -0.00003326 49.99996674
8 -0.00000450 49.99999550
9 -0.00000060 49.99999939
10 -0.00000008 49.99999992
… … …
Remarks
Negative deviation decays over Time-path rises and converges to equilibrium
time (𝑪∗ = 𝟓𝟎)
As 𝒕 → ∞, 𝑪𝒄 → 𝟎 ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ (𝑪∗ )
And when 𝑪𝒄 → 𝟎, (𝑪𝒕 )𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑠
Graphically
The straight line shows the equilibrium value of consumption (𝑪∗ ), while curve shows the
deviation from equilibrium (𝑪𝒄 ), which decreases over time and leads the time-path to
convergence.
______________________________________________________________________________
Lesson 22
NATIONAL INCOME DETERMINATION USING DIFFERENTIAL EQUATIONS
TOPIC 089: NATIONAL INCOME DETERMINATION USING DIFFERENTIAL EQUATIONS

𝒀=𝑪+𝑰
𝑪 = 𝜶 + 𝜷𝒀 𝜶 > 𝟎, 𝟎 < 𝜷 < 𝟏
𝒅𝒀
𝑰 = 𝜹𝒀 + 𝜸 ( ) 𝜸 > 𝟎, 𝜹 < 𝟏
𝒅𝒕
Find the time path and the intertemporal equilibrium level of national income.
𝒀=𝑪+𝑰
Substituting values of 𝑪 and 𝑰.
𝒅𝒀
𝒀 = 𝜶 + 𝜷𝒀 + 𝜹𝒀 + 𝜸 ( )
𝒅𝒕
𝒅𝒀
𝒀 = 𝜶 + 𝜷𝒀 + 𝜹𝒀 + 𝜸 ( )
𝒅𝒕
𝒅𝒀
−𝜸 ( ) − 𝜷𝒀 − 𝜹𝒀 + 𝒀 = 𝜶
𝒅𝒕
𝒅𝒀
−𝜸 ( ) − 𝜷𝒀 − 𝜹𝒀 + 𝒀 = 𝜶
𝒅𝒕
𝒅𝒀
−𝜸 ( ) + 𝒀 − 𝜷𝒀 − 𝜹𝒀 = 𝜶
𝒅𝒕
𝒅𝒀
−𝜸 ( ) + (𝟏 − 𝜷 − 𝜹)𝒀 = 𝜶
𝒅𝒕
𝜸 𝒅𝒀 (𝟏 − 𝜷 − 𝜹) 𝜶
− ( )+ 𝒀=
(−𝜸) 𝒅𝒕 (−𝜸) (−𝜸)
𝒅𝒀 𝟏−𝜷−𝜹 𝜶
( )+( )𝒀 = −
( )+( )𝒀 = −
𝒅𝒚
( )+( ) 𝒀 = (− )
⏟𝒅𝒕 ⏟ −𝜸 ⏟ 𝜸
𝑑𝑦 𝑎 𝑏
𝑑𝑡
𝒅𝒚 𝒅𝒀 𝟏−𝜷−𝜹 𝜶
Here 𝒅𝒕
= 𝒅𝒕 , 𝒚 = 𝒀, 𝒂 = ( −𝜸 ) , 𝒃 = (− 𝜸)
𝜶
𝜶
______________________________________________________________________________
𝟏−𝜷−𝜹
𝒂=( −𝜸
); Both 𝜷 and 𝜹 are less than 𝟏. The numerator (𝟏 − 𝜷 − 𝜹) can become 𝟎 only if 𝜷 +
𝟏−𝜷−𝜹
𝜹 = 𝟏. So, we assume that the 𝜷 + 𝜹 ≠ 𝟏 and (𝟏 − 𝜷 − 𝜹 ≠ 𝟎), so ( −𝜸
≠ 𝟎) i.e. 𝒂 ≠ 𝟎 – Non-
Homogenous Case with 𝒂 ≠ 𝟎.
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
𝜶 𝜶
(− 𝜸 ) 𝟏−𝜷−𝜹 (− 𝜸 )
−( )𝒕
−𝜸
𝒀(𝒕) = 𝒀(𝟎) − 𝒆 +
𝟏−𝜷−𝜹 𝟏−𝜷−𝜹
( −𝜸 ) ( −𝜸 )
{ }
𝜶 𝜶
( − 𝜸) −(
𝟏−𝜷−𝜹
)𝒕
( − 𝜸)
𝒀(𝒕) = {𝒀(𝟎) − }𝒆 −𝜸 +
𝟏−𝜷−𝜹 𝟏−𝜷−𝜹
( ) ( )
−𝜸 −𝜸
(𝜶) 𝟏−𝜷−𝜹 (𝜶)
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − }𝒆 𝜸 +
(𝟏 − 𝜷 − 𝜹) (𝟏 − 𝜷 − 𝜹)
𝜶 𝟏−𝜷−𝜹 𝜶
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − ( )} 𝒆 𝜸 +( )
𝟏−𝜷−𝜹 𝟏−𝜷−𝜹
𝒕 → ∞.
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − ( )} 𝒆 𝜸 +( )
⏟ 𝟏−𝜷−𝜹 ⏟𝟏 − 𝜷 − 𝜹
𝑌𝑐 𝑌𝑝
Need for time path of income {𝒀(𝒕)} to converge to equilibrium income (𝒀∗ ) i.e. 𝒀(𝒕) → 𝒀∗ as
𝒀𝒄 → 𝟎 when 𝒕 → ∞.
𝒀𝒄 → 𝟎
𝜶 𝟏−𝜷−𝜹
( )𝒕
{𝒀(𝟎) − ( )} 𝒆 𝜸 →𝟎
𝟏−𝜷−𝜹
{𝜶,𝜷,𝜹 >𝟎} {𝜷,𝜹,𝜸>𝟎}
⏞𝟏−𝜷−𝜹
⏞ 𝜶 ( ) 𝒕
𝒀(𝟎) − ( ) ⏟ 𝒆 𝜸 →𝟎
⏟ 𝟏−𝜷−𝜹
{
Therefore,
______________________________________________________________________________
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − ( )} 𝒆 𝜸 +( )
⏟ 𝟏−𝜷−𝜹 𝟏−𝜷−𝜹
→∞
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − ( ) }𝒆 𝜸 + ( )
⏟ 𝟏−𝜷−𝜹 𝟏−𝜷−𝜹
→∞
𝟏−𝜷−𝜹
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − 𝒀𝒑 } 𝒆 𝜸 + 𝒀𝒑
⏟
→∞
𝟏−𝜷−𝜹
( )𝒕
𝒀(𝒕) = {𝒀(𝟎)
⏟ − 𝒀∗ }𝒆 𝜸 + 𝒀∗
→∞
𝒀(𝒕) → ∞
𝒀(𝒕) ↛ 𝒀∗
Interpretation: It can be said that provided 𝜷 > 𝟎, 𝜹 < 𝟏 and 𝜸 > 𝟎, regardless of initial national
income 𝒀(𝟎) is greater or smaller than equilibrium national income (𝒀∗ ), the time-path of national
income shall diverge away from equilibrium level of national income as time is allowed to pass
indefinitely. Hence the time-path is dynamically unstable.
TOPIC 090: NATIONAL INCOME DETERMINATION: NUMERICAL
Considering a two-sector economy; we have following structural equations:

𝒅𝒀
= 𝟎. 𝟓(𝑪 + 𝑰 − 𝒀)
𝒅𝒕
𝑪 = 𝟎. 𝟖𝒀 + 𝟒𝟎𝟎
𝑰 = 𝟔𝟎𝟎
Find the time-path of national income of the economy; 𝒀(𝟎) = 𝟕𝟎𝟎𝟎.
Solution: Given the equation for rate of national income
𝒅𝒀
𝒅𝒕
= 𝟎. 𝟓(𝑪 + 𝑰 − 𝒀); we can substitute equations of 𝑪 and 𝑰.
𝒅𝒀
= 𝟎. 𝟓(𝑪 + 𝑰 − 𝒀)
𝒅𝒕
𝒅𝒀
= 𝟎. 𝟓(𝟎. 𝟖𝒀 + 𝟒𝟎𝟎 + 𝟔𝟎𝟎 − 𝒀)
𝒅𝒕
𝒅𝒀
= 𝟎. 𝟓(𝟎. 𝟖𝒀 − 𝒀 + 𝟒𝟎𝟎 + 𝟔𝟎𝟎)
𝒅𝒕
𝒅𝒀
= 𝟎. 𝟓(−𝟎. 𝟐𝒀 + 𝟏𝟎𝟎𝟎)
𝒅𝒕
𝒅𝒀
= −𝟎. 𝟏𝒀 + 𝟓𝟎𝟎
𝒅𝒕
𝒅𝒀
+ 𝟎. 𝟏𝒀 = 𝟓𝟎𝟎
𝒅𝒕
First Order Differential Equation for National Income.
𝒅𝒀
+ 𝟎.
⏟ 𝟏 𝒀 = 𝟓𝟎𝟎
⏟
⏟
𝒅𝒕 𝑎 𝑏
𝑑𝑦
𝑑𝑡
______________________________________________________________________________
𝒅𝒚 𝒅𝒀
Here 𝒅𝒕
= 𝒅𝒕 , 𝒚 = 𝒀, 𝒂 = 𝟎. 𝟏, 𝒃 = 𝟓𝟎𝟎, 𝒚(𝟎) = 𝒀(𝟎) = 𝟕𝟎𝟎𝟎
As 𝒃 = 𝟓𝟎𝟎, so 𝒃 ≠ 𝟎: Non-Homogenous Case.
As 𝒂 = 𝟎. 𝟏, so 𝒂 ≠ 𝟎: Non-Homogenous Case with 𝒂 ≠ 𝟎.
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
𝟓𝟎𝟎 𝟓𝟎𝟎
𝒀(𝒕) = (𝒀(𝟎) − ) 𝒆− 𝟎.𝟏 𝒕 +
𝟎. 𝟏 𝟎. 𝟏
𝟓𝟎𝟎 −𝟎.𝟏𝒕 𝟓𝟎𝟎
𝒀(𝒕) = (𝟕𝟎𝟎𝟎 − )𝒆 +
𝟎. 𝟏 𝟎. 𝟏
𝒀(𝒕) = (𝟕𝟎𝟎𝟎 − 𝟓𝟎𝟎𝟎)𝒆−𝟎.𝟏𝒕 + 𝟓𝟎𝟎𝟎
𝒀(𝒕) = (𝟐𝟎𝟎𝟎)𝒆−𝟎.𝟏𝒕 + 𝟓𝟎𝟎𝟎
𝒀(𝒕) = 𝟐𝟎𝟎𝟎 𝒆−𝟎.𝟏𝒕 + 𝟓𝟎𝟎𝟎
It is the definite Time-Path of National Income of the economy.
𝟐𝟎𝟎𝟎𝒆−𝟎.𝟏𝒕 + 𝟓𝟎𝟎𝟎
𝒀(𝒕) = ⏟ ⏟
𝒀𝒄 𝒀𝒑
−𝟎.𝟏𝒕
Dynamic Stability: The expression; 𝒀(𝒕) = 𝟐𝟎𝟎𝟎𝒆 + 𝟓𝟎𝟎𝟎 has an exponential decay term
in the box. i.e.:
______________________________________________________________________________
Complementary Function (Deviation) Time-Path

𝒕 𝒀𝒄 = 𝟐𝟎𝟎𝟎𝒆−𝟎.𝟏𝒕 𝒀(𝒕) = 𝟐𝟎𝟎𝟎𝒆−𝟎.𝟏𝒕 + 𝟓𝟎𝟎𝟎
0 2000 7000
1 1809.675 6809.674836
20 270.6706 5270.670566
40 36.63128 5036.631278
60 4.957504 5004.957504
80 0.670925 5000.670925
100 0.090800 5000.090800
120 0.012288 5000.012288
140 0.001663 5000.001663
160 0.000225 5000.000225
200 0.000004 5000.000004
… … …
Remarks
Positive deviation decays over time Time-path slides and converges to
equilibrium (𝒀∗ = 𝟓𝟎𝟎𝟎)
As 𝒕 → ∞, 𝒀𝒄 → 𝟎 ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ (𝒀∗ )
And when 𝒀𝒄 → 𝟎, (𝒀𝒕 )𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑠
Graphically
The straight line shows the equilibrium value of national income (𝒀∗ ), while curve shows
the deviation from equilibrium (𝒀𝒄 ), which decreases over time and leads the time-path
to convergence.
TOPIC 091: INCOME DYNAMICS & INDUCED GOVERNMENT SPENDING USING

DIFFERENTIAL EQUATIONS

𝒀=𝑪+𝑰+𝑮
𝑪 = 𝜶 + 𝜷𝒀 𝜶 > 𝟎, 𝟎 < 𝜷
Following is the investment function:
𝒅𝒀
𝑰 = 𝜹𝒀 + 𝜸 ( 𝒅𝒕 ) 𝜸 > 𝟎, 𝜹 < 𝟏
Government Spending is included as the 3rd sector:
𝑮 = 𝒈𝒀 𝒈<𝟏
______________________________________________________________________________
𝒈 is the proportion of national income spent as government expenditure.

Find the time path and the intertemporal equilibrium level of national income.
𝒀=𝑪+𝑰+𝑮
Substituting values of 𝑪, 𝑰 and 𝑮.
𝒅𝒀
𝒀 = 𝜶 + 𝜷𝒀 + 𝜹𝒀 + 𝜸 ( ) + 𝒈𝒀
𝒅𝒕
𝒅𝒀
𝒀 = 𝜶 + 𝜷𝒀 + 𝜹𝒀 + 𝜸 ( ) + 𝒈𝒀
𝒅𝒕
𝒅𝒀
−𝜸 ( ) + 𝒀 − 𝜷𝒀 − 𝜹𝒀 − 𝒈𝒀 = 𝜶
𝒅𝒕
𝒅𝒀
−𝜸 ( ) + (𝟏 − 𝜷 − 𝜹 − 𝒈)𝒀 = 𝜶
𝒅𝒕
𝜸 𝒅𝒀 (𝟏 − 𝜷 − 𝜹 − 𝒈) 𝜶
− ( )+ 𝒀=
(−𝜸) 𝒅𝒕 (−𝜸) (−𝜸)
𝒅𝒀 𝟏−𝜷−𝜹−𝒈 𝜶
( )+( )𝒀 = −
( )+( )𝒀 = −
𝒅𝒚
( )+( ) 𝒀 = (− )
⏟𝒅𝒕 ⏟ −𝜸 ⏟ 𝜸
𝒅𝒚 𝑎 𝑏
𝒅𝒕
𝒅𝒚 𝒅𝒀 𝟏−𝜷−𝜹−𝒈 𝜶
Here = ,𝒚 = 𝒀, 𝒂 = ( ),𝒃 = (− )
𝒅𝒕 𝒅𝒕 −𝜸 𝜸
𝜶
𝜶
𝟏−𝜷−𝜹−𝒈
𝒂 = ( −𝜸 ); the numerator (𝟏 − 𝜷 − 𝜹 − 𝒈) can be non-zero if (𝜷 + 𝜹 + 𝒈) ≠ 𝟏. Also 𝜸 is
positive value. Therefore, value of 𝒂 will be some positive/negative (non-zero) value. i.e. 𝒂 ≠ 𝟎 –
Non-Homogenous Case with 𝒂 ≠ 𝟎.
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
𝜶 𝜶
(− 𝜸 ) 𝟏−𝜷−𝜹−𝒈 (− 𝜸 )
−( )𝒕
−𝜸
𝒀(𝒕) = 𝒀(𝟎) − 𝒆 +
𝟏−𝜷−𝜹−𝒈 𝟏−𝜷−𝜹−𝒈
( −𝜸 ) ( −𝜸 )
{ }
______________________________________________________________________________
𝜶 𝜶
( − 𝜸) −(
)𝒕
( − 𝜸)
𝒀(𝒕) = {𝒀(𝟎) − }𝒆 −𝜸 +
( −𝜸 ) ( −𝜸 )
(𝜶) 𝟏−𝜷−𝜹−𝒈 (𝜶)
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − }𝒆 𝜸 +
(𝟏 − 𝜷 − 𝜹 − 𝒈) (𝟏 − 𝜷 − 𝜹 − 𝒈)
𝜶 𝟏−𝜷−𝜹−𝒈 𝜶
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − ( )} 𝒆 𝜸 +( )
Definite Time-path for National Income in Dynamic Sense. And 𝒈 has positive effect on the
national income, via its negative sign in denominator.
𝒕 → ∞.
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − ( )} 𝒆 𝜸 +( )
⏟ 𝟏 − 𝜷 − 𝜹 − 𝒈 ⏟ 𝟏 − 𝜷 − 𝜹−𝒈
𝑌𝑐 𝑌𝑝
Need for time path of income {𝒀(𝒕)} to converge to equilibrium income (𝒀∗ ) i.e. 𝒀(𝒕) → 𝒀∗ as 𝒀𝒄 →
𝟎 when 𝒕 → ∞.
𝒀𝒄 → 𝟎
𝜶 𝟏−𝜷−𝜹−𝒈
( )𝒕
{𝒀(𝟎) − ( )} 𝒆 𝜸 →𝟎
{𝜶,𝜷,𝜹,𝒈>𝟎} {𝜷,𝜹,𝜸,𝒈>𝟎}
⏞𝟏−𝜷−𝜹−𝒈
⏞ 𝜶 (
𝜸
) 𝒕
𝒀(𝟎) − (𝟏−𝜷−𝜹−𝒈) ⏟ 𝒆 →𝟎
{
[ 𝑫𝒆𝒗𝒊𝒂𝒕𝒊𝒐𝒏 (+𝒗𝒆 𝒐𝒓−𝒗𝒆) ]
Therefore,
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − ( )} 𝒆 𝜸 +( )
⏟ 𝟏−𝜷−𝜹−𝒈 𝟏−𝜷−𝜹−𝒈
→∞
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − ( ) }𝒆 𝜸 + ( )
⏟ 𝟏−𝜷−𝜹−𝒈 𝟏−𝜷−𝜹−𝒈
→∞
( )𝒕
𝒀(𝒕) = {𝒀(𝟎) − 𝒀𝒑 } 𝒆 𝜸 + 𝒀𝒑
⏟
→∞
( )𝒕
𝒀(𝒕) = {𝒀(𝟎)
⏟ − 𝒀∗ }𝒆 𝜸 + 𝒀∗
→∞
𝒀(𝒕) → ∞
______________________________________________________________________________
𝒀(𝒕) ↛ 𝒀∗
Interpretation: It can be said that provided 𝜷 > 𝟎, 𝜹 < 𝟏, 𝜸 > 𝟎 and 𝒈 < 𝟏, regardless of initial
national income 𝒀(𝟎) is greater or smaller than equilibrium national income (𝒀∗ ), the time-path of
national income shall diverge away from equilibrium level of national income as time is allowed
to pass indefinitely. Hence the time-path is dynamically unstable.
TOPIC 092: DYNAMICS OF INVESTMENT USING DIFFERENTIAL EQUATIONS
A principal of $𝟔𝟎 is invested. The value, 𝑰(𝒕), of the investment, 𝒕 days later, satisfies the
𝒅𝑰
differential equation 𝒅𝒕 = 𝟎. 𝟎𝟎𝟐𝑰 + 𝟓. Find the value of the investment if the initial level of
investment 𝒚(𝟎) is 𝟐𝟎𝟎𝟎.
𝒅𝑰
Solution: Given the equation for rate of investment 𝒅𝒕 = 𝟎. 𝟎𝟎𝟐𝑰 + 𝟓; we can rearrange to
𝒅𝑰
= 𝟎. 𝟎𝟎𝟐𝑰 + 𝟓
𝒅𝒕
𝒅𝑰
− 𝟎. 𝟎𝟎𝟐𝑰 = 𝟓
𝒅𝒕
𝒅𝑰
+ (𝟎. 𝟎𝟎𝟐) 𝑰 = ⏟
𝟓
⏟ ⏟
𝒅𝒕 𝑎 𝑏
𝑑𝑦
𝑑𝑡
𝒅𝒚 𝒅𝑰
Here = , 𝒚 = 𝑰, 𝒂 = 𝟎. 𝟎𝟎𝟐, 𝒃 = 𝟓, 𝒚(𝟎) = 𝑰(𝟎) = 𝟐𝟎𝟎𝟎.
𝒅𝒕 𝒅𝒕
As 𝒃 = 𝟓, so 𝒃 ≠ 𝟎: Non-Homogenous Case.
As 𝒂 = 𝟎. 𝟎𝟎𝟐, so 𝒂 ≠ 𝟎: Non-Homogenous Case with 𝒂 ≠ 𝟎.
Therefore, using the Definite Solution of time-path of 𝑵𝑯𝑪𝒂≠𝟎:
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
𝟓 𝟓
𝑰(𝒕) = (𝑰(𝟎) − ) 𝒆− 𝟎.𝟎𝟎𝟐 𝒕 +
𝟎. 𝟎𝟎𝟐 𝟎. 𝟎𝟎𝟐
𝟓 𝟓
𝑰(𝒕) = (𝟐𝟎𝟎𝟎 − ) 𝒆−𝟎.𝟎𝟎𝟐𝒕 +
𝟎. 𝟎𝟎𝟐 𝟎. 𝟎𝟎𝟐
𝑰(𝒕) = (𝟐𝟎𝟎𝟎 − 𝟐𝟓𝟎𝟎)𝒆−𝟎.𝟎𝟎𝟐𝒕 + 𝟐𝟓𝟎𝟎
𝑰(𝒕) = −𝟓𝟎𝟎𝒆−𝟎.𝟎𝟎𝟐𝒕 + 𝟐𝟓𝟎𝟎
It is the definite Time-Path of Investment.
−𝟓𝟎𝟎𝒆−𝟎.𝟎𝟎𝟐𝒕 + 𝟐𝟓𝟎𝟎
𝑰(𝒕) = ⏟ ⏟
𝑰𝒄 𝑰𝒑
−𝟎.𝟎𝟎𝟐𝒕
Dynamic Stability: The expression; 𝑰(𝒕) = −𝟓𝟎𝟎𝒆 + 𝟐𝟓𝟎𝟎 has an exponential decay term
in the box. i.e.:
______________________________________________________________________________
Complementary Function (Deviation) Time-Path

𝒕 𝑰𝒄 = −𝟓𝟎𝟎𝒆−𝟎.𝟎𝟎𝟐𝒕 𝑰(𝒕) = −𝟓𝟎𝟎𝒆−𝟎.𝟎𝟎𝟐𝒕 + 𝟐𝟓𝟎𝟎
0 -500.0000 2000.0000
1 -499.0009 2000.9990
2 -498.0040 2001.9960
3 -497.0089 2002.9910
4 -496.0160 2003.9840
5 -495.0249 2004.9751
6 -494.0359 2005.9641
7 -493.0488 2006.9512
8 -492.0637 2007.9363
9 -491.0805 2008.9194
10 -490.0993 2009.9006
… … …
Remarks
Negative deviation decays over time Time-path rises and converges to equilibrium
(𝑰∗ = 𝟐𝟓𝟎𝟎)
As 𝒕 → ∞, 𝑰𝒄 → 𝟎 ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ (𝑰∗ )
And when 𝑰𝒄 → 𝟎, (𝑰𝒕 )𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑠
Graphically
The straight line shows the equilibrium value of consumption (𝑰∗ ), while curve shows the
deviation from equilibrium (𝑰𝒄 ), and this deviation decreases over time and leads the time-
path to convergence.
TOPIC 093: DYNAMICS OF INTEREST RATE IN CAPITAL MARKET USING DIFFERENTIAL

EQUATIONS
The demand and supply for borrowed capital in the capital market are given by the investment
and savings functions, respectively:
𝒅𝒊 𝒅𝒊
𝑰 = 𝜶 − 𝜷𝒊 + 𝝆 𝒅𝒕 𝑺 = −𝜸 + 𝜹𝒊 + 𝝈 𝒅𝒕 (𝜶, 𝜷, 𝜸, 𝜹)
Determine the intertemporal equilibrium level and time path of interest rate 𝒊(𝒕). What values of
the parameters 𝝆 and 𝝈 ensure dynamic stability?
______________________________________________________________________________
Solution: Equilibrium in the market requires equality of Investment and Savings: (𝑺 = 𝑰).
𝒅𝒊 𝒅𝒊
𝜶 − 𝜷𝒊 + 𝝆 = −𝜸 + 𝜹𝒊 + 𝝈
𝒅𝒕 𝒅𝒕
Rearranging to compare with the standard form of differential equation.
𝒅𝒊 𝒅𝒊
−𝜷𝒊 − 𝜹𝒊 + 𝝆 − 𝝈 = −𝜶 − 𝜸
𝒅𝒕 𝒅𝒕
𝒅𝒊 𝒅𝒊
𝝆 − 𝝈 − 𝜷𝒊 − 𝜹𝒊 = −𝜶 − 𝜸
𝒅𝒕 𝒅𝒕
𝒅𝒊
(𝝆 − 𝝈) − (𝜷 + 𝜹)𝒊 = −(𝜶 + 𝜸)
𝒅𝒕
Dividing by (𝝆 − 𝝈) for normalizing the differential equation.
(𝝆 − 𝝈) 𝒅𝒊 (𝜷 + 𝜹) (𝜶 + 𝜸)
− 𝒊=−
(𝝆 − 𝝈) 𝒅𝒕 (𝝆 − 𝝈) (𝝆 − 𝝈)
𝒅𝒊 (𝜷 + 𝜹) (𝜶 + 𝜸)
− 𝒊=−
𝒅𝒕 (𝝆 − 𝝈) (𝝆 − 𝝈)
𝒅𝒚
Comparing with the standard form of the differential equation, 𝒅𝒕 + 𝒂𝒚 = 𝒃.
𝒅𝒊 (𝜷 + 𝜹) (𝜶 + 𝜸)
+ {− } ⏟𝒊 = −
⏟ ⏟ (𝝆 − 𝝈)
𝒅𝒕 ⏟ (𝝆 − 𝝈)
𝑦
𝑑𝑦 𝑏
𝑎
𝑑𝑡
𝒅𝒚 𝒅𝒊 (𝜷+𝜹) (𝜶+𝜸)
Here = , 𝒚 = 𝒊, 𝒂 = {− (𝝆−𝝈)} , 𝒃 = {− (𝝆−𝝈)}
𝒅𝒕 𝒅𝒕
(𝜶+𝜸)
𝒃= {− (𝝆−𝝈)}; As 𝜶 > 𝟎 and 𝜸 > 𝟎 (both positive constants), their ratio will be non-zero positive
(𝜶+𝜸)
constant. However, the value of 𝒃 = {− (𝝆−𝝈)} who’s denominator should be non-zero to avoid
undefined answer (∞). The answer will be negative, though still non-zero. i.e. 𝒃 ≠ 𝟎 – Non-
Homogenous Case.
(𝜷+𝜹)
𝒂 = {− (𝝆−𝝈)}; Both 𝜷 and 𝜹 are less than 𝟏. The numerator (𝜷 + 𝜹) shall be a non-zero positive
(𝜷+𝜹)
constant. However, 𝒂 = {− (𝝆−𝝈)} so value of numerator (𝝆 − 𝝈) should be non-zero to avoid
(𝜷+𝜹)
undefined answer (∞). So [{− (𝝆−𝝈)} ≠ 𝟎] i.e. 𝒂 ≠ 𝟎 – Non-Homogenous Case with 𝒂 ≠ 𝟎.
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
(𝜶 + 𝜸) (𝜶 + 𝜸)
{− } (𝜷+𝜹) {− }
(𝝆 − 𝝈) − {−
(𝝆−𝝈)
}𝒕 (𝝆 − 𝝈)
𝒊(𝒕) = 𝒊(𝟎) − 𝒆 +
(𝜷 + 𝜹) (𝜷 + 𝜹)
{− } {− }
{ (𝝆 − 𝝈) } (𝝆 − 𝝈)
(𝜶 + 𝜸) (𝜶 + 𝜸)
{− } (𝜷+𝜹) {− }
(𝝆 − 𝝈) −{− }𝒕
(𝝆−𝝈) +
(𝝆 − 𝝈)
𝒊(𝒕) = [𝒊(𝟎) − ]𝒆
(𝜷 + 𝜹) (𝜷 + 𝜹)
{− } {− }
(𝝆 − 𝝈) (𝝆 − 𝝈)
______________________________________________________________________________
(𝜶 + 𝜸) (𝜶 + 𝜸)
{ } 𝜷+𝜹 { }
(𝝆 − 𝝈) ( )𝒕 (𝝆 − 𝝈)
𝒊(𝒕) = [𝒊(𝟎) − ] 𝒆 𝝆−𝝈 +
(𝜷 + 𝜹) (𝜷 + 𝜹)
{ } { }
(𝝆 − 𝝈) (𝝆 − 𝝈)
𝜶+𝜸 𝜷+𝜹 𝜶+𝜸
( )𝒕
𝒊(𝒕) = {𝒊(𝟎) − ( )} 𝒆 𝝆−𝝈 + ( )
𝜷+𝜹 𝜷+𝜹
( )𝒕
𝒊(𝒕) = {𝒊(𝟎) − ( )} 𝒆 𝝆−𝝈 + ( )
𝜷+𝜹 𝜷+𝜹
Definite Time-path for Interest rate in Dynamic Sense.
it should converge towards equilibrium. We know that equilibrium is particular solution (𝒊𝒑 ).
equation, deviation component is measured by (𝒊𝒄 ). Therefore, for dynamic stability; 𝒊𝒄 → 𝟎 as
𝒕 → ∞.
Using the Definite Time-path for interest rate in Dynamic Sense.
( )𝒕
𝒊(𝒕) = {𝒊(𝟎) − ( )} 𝒆 𝝆−𝝈 + ( )
⏟ 𝜷+𝜹 ⏟𝜷 + 𝜹
𝒊𝒄 𝒊𝒑
Need for time path of income {𝒊(𝒕)} to converge to equilibrium level of interest rate (𝒊∗ ) i.e.
𝒊(𝒕) → 𝒊∗ as 𝒊𝒄 → 𝟎 when 𝒕 → ∞.
𝒊𝒄 → 𝟎
𝜷+𝜹
𝜶+𝜸 ( )𝒕
𝝆−𝝈
{𝒊(𝟎) − ( )} 𝒆 →𝟎
𝜷+𝜹
𝜷+𝜹
( 𝝆−𝝈 ) is the rate of growth/decay of the exponential expression. Therefore, its negativity shall
𝜷+𝜹
ensure its decay. i.e. ( 𝝆−𝝈 ) < 𝟎. It’s already known that 𝜷 and 𝜹 are positive and their sum
(Numerator; 𝜷 + 𝜹) will also be positive. So, denominator (𝝆 − 𝝈) must be negative. i.e.
(𝝆 − 𝝈) < 𝟎. In comparative terms, (𝝆 < 𝝈). Therefore, the condition for dynamic stability of
the time-path of interest rate 𝒊(𝒕) is (𝝆 < 𝝈):
{𝜶,𝜷,𝜹,𝜸 >𝟎} {𝜷,𝜹>𝟎 &(𝝆< 𝝈)}
⏞
⏞𝜶+𝜸 (
𝜷+𝜹
𝝆−𝝈
) 𝒕
𝒊(𝟎) − (𝜷+𝜹) 𝒆
⏟ →𝟎
{
⏟ 𝒀(𝟎)≷𝒀𝒑 } 𝑫𝒆𝒄𝒂𝒚
Therefore,
( )𝒕
𝒊(𝒕) = {𝒊(𝟎) − ( )} 𝒆 𝝆−𝝈 + ( )
⏟ 𝜷+𝜹 𝜷+𝜹
→0
( )𝒕
𝒊(𝒕) = {𝒊(𝟎) − ( ) } 𝒆 𝝆−𝝈 + ( )
⏟ 𝜷+𝜹 𝜷+𝜹
→0
𝜷+𝜹
( )𝒕
𝒊(𝒕) = {𝒊(𝟎) − (𝒊𝒑 ) } 𝒆 𝝆−𝝈 + (𝒊𝒑 )
⏟
→0
______________________________________________________________________________
𝜷+𝜹
( )𝒕
𝒊(𝒕) = {𝒊(𝟎) − 𝒊𝒑 } 𝒆 𝝆−𝝈 + 𝒊𝒑
⏟
→0
As particular integral is also the equilibrium value of the variable 𝒊𝒑 = 𝒊∗ .
𝜷+𝜹
( )𝒕
{𝒊(𝟎) − 𝒊∗ }𝒆
𝒊(𝒕) = ⏟ 𝝆−𝝈 + 𝒊∗
→0
𝒊(𝒕) → ∞
Interpretation: It can be said that provided 𝜷 > 𝟎, 𝜹 > 𝟎 with 𝝆 < 𝝈, regardless of initial interest
rate 𝒊(𝟎) is greater or smaller than equilibrium interest rate (𝒊∗ ), the time-path of interest rate shall
converge towards equilibrium level of interest rate as time is allowed to pass indefinitely. Hence
the time-path is dynamically stable.
______________________________________________________________________________
Lesson 23
VARIOUS ECONOMIC DYNAMICS USING DIFFERENTIAL EQUATIONS
TOPIC 094: DYNAMICS OF VALUE OF OIL WELL USING DIFFERENTIAL EQUATIONS
The value of an oil well, which is explored, initially is 𝑽(𝟎) = $𝟐𝟕𝟎𝟎, and is decaying at the rate
𝒅𝑽
𝒅𝒕
+ 𝟎. 𝟎𝟒 𝑽(𝒕) = 𝟐𝟖.
a) What is the general time path that describes the value of this well?
b) What is the future value of the well al time 𝒕 = 𝟏𝟎?
𝒅𝑽
Solution: Given the equation for rate of value of oil-well 𝒅𝒕 + 𝟎. 𝟎𝟒 𝑽(𝒕) = 𝟐𝟖; we can rewrite to
compare with the standard form of differential equation:
𝒅𝑽
+ 𝟎. 𝟎𝟒𝑽 = 𝟐𝟖
𝒅𝒕
𝒅𝑽
+ (𝟎. 𝟎𝟒) 𝑽
⏟ = 𝟐𝟖
⏟
⏟ ⏟
𝒅𝒕 𝑎 𝒚 𝑏
𝑑𝑦
𝑑𝑡
𝒅𝒚 𝒅𝑽
Here 𝒅𝒕
= 𝒅𝒕 , 𝒚 = 𝑽, 𝒂 = 𝟎. 𝟎𝟒, 𝒃 = 𝟐𝟖, 𝒚(𝟎) = 𝑽(𝟎) = 𝟐𝟕𝟎𝟎.
As 𝒃 = 𝟐𝟖, so 𝒃 ≠ 𝟎: Non-Homogenous Case.
As 𝒂 = 𝟎. 𝟎𝟒, so 𝒂 ≠ 𝟎: Non-Homogenous Case with 𝒂 ≠ 𝟎.
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
𝟐𝟖 𝟐𝟖
𝑽(𝒕) = {𝑽(𝟎) − } 𝒆− 𝟎.𝟎𝟒 𝒕 +
𝟎. 𝟎𝟒 𝟎. 𝟎𝟒
𝟐𝟖 𝟐𝟖
𝑽(𝒕) = (𝟐𝟕𝟎𝟎 − ) 𝒆−𝟎.𝟎𝟒𝒕 +
𝟎. 𝟎𝟒 𝟎. 𝟎𝟒
𝑽(𝒕) = (𝟐𝟕𝟎𝟎 − 𝟕𝟎𝟎)𝒆−𝟎.𝟎𝟒𝒕 + 𝟕𝟎𝟎
𝑽(𝒕) = (𝟐𝟎𝟎𝟎)𝒆−𝟎.𝟎𝟒𝒕 + 𝟕𝟎𝟎
It is the definite Time-Path of Value of Oil-Well.
𝟐𝟎𝟎𝟎 𝒆−𝟎.𝟎𝟒𝒕 + 𝟕𝟎𝟎
𝑽(𝒕) = ⏟ ⏟
𝑽𝒄 𝑽𝒑
−𝟎.𝟎𝟒𝒕
Dynamic Stability: The expression; 𝑽(𝒕) = 𝟐𝟎𝟎𝟎 𝒆 + 𝟕𝟎𝟎 has an exponential decay term
in the box. i.e.:
______________________________________________________________________________
Time-Path
(Deviation)
𝒕 𝑽𝒄 = 𝟐𝟎𝟎𝟎 𝒆−𝟎.𝟎𝟒𝒕 𝑽(𝒕) = 𝟐𝟎𝟎𝟎 𝒆−𝟎.𝟎𝟒𝒕 + 𝟕𝟎𝟎
0 2000 2700
1 1921.58 2621.58
40 403.793 1103.79
80 81.5244 781.524
120 16.4595 716.459
160 3.32311 703.323
200 0.67093 700.671
240 0.13546 700.135
280 0.02735 700.027
320 0.00552 700.006
340 0.00248 700.002
… … …
Remarks
Positive deviation decays Time-path slides and converges to equilibrium
over time (𝑽∗ = 𝟕𝟎𝟎)
As 𝒕 → ∞, 𝑽𝒄 → 𝟎 ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ (𝑽∗ )
And when 𝑽𝒄 → 𝟎, (𝑽𝒕 )𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑠
Graphically
The straight line shows the equilibrium value of Value of Oil-Well (𝑽∗ ), while curve
shows the deviation from equilibrium (𝑽𝒄 ), and this deviation decreases over time and
leads the time-path to convergence.
TOPIC 095: DYNAMICS OF VALUE OF GOLD MINE USING DIFFERENTIAL EQUATIONS
𝒅𝑮
The value of a gold mine is initially 𝑮(𝟎) = $𝟏, 𝟒𝟎𝟎 and is decaying at the rate 𝒅𝒕
+ 𝟎. 𝟎𝟒 𝑮(𝒕) =
𝟏𝟔.
a) What is the general time path that describes the value of the gold mine?
b) What is the future value of the gold mine at time 𝒕 = 𝟏𝟎?
______________________________________________________________________________
𝒅𝑮
Solution: Given the rate of change of value of Gold mine 𝒅𝒕 + 𝟎. 𝟎𝟒 𝑮(𝒕) = 𝟏𝟔, we can rewrite
to compare with the standard form of differential equation:
𝒅𝑮
+ 𝟎. 𝟎𝟒 𝑮 = 𝟏𝟔
𝒅𝒕
𝒅𝑮
+ (𝟎. 𝟎𝟒) 𝑮
⏟ = 𝟏𝟔
⏟
⏟ ⏟
𝑑𝑦
𝑑𝑡
𝒅𝒚 𝒅𝑮
Here = , 𝒚 = 𝑮, 𝒂 = 𝟎. 𝟎𝟒, 𝒃 = 𝟏𝟔, 𝒚(𝟎) = 𝑮(𝟎) = 𝟏𝟒𝟎𝟎.
𝒅𝒕 𝒅𝒕
As 𝒃 = 𝟏𝟔, so 𝒃 ≠ 𝟎: Non-Homogenous Case.
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
𝟏𝟔 𝟏𝟔
𝑮(𝒕) = {𝑮(𝟎) − } 𝒆− 𝟎.𝟎𝟒 𝒕 +
𝟎. 𝟎𝟒 𝟎. 𝟎𝟒
𝟏𝟔 𝟏𝟔
𝑮(𝒕) = (𝟏𝟒𝟎𝟎 − ) 𝒆−𝟎.𝟎𝟒𝒕 +
𝟎. 𝟎𝟒 𝟎. 𝟎𝟒
𝑮(𝒕) = (𝟏𝟒𝟎𝟎 − 𝟒𝟎𝟎)𝒆−𝟎.𝟎𝟒𝒕 + 𝟒𝟎𝟎
𝑮(𝒕) = (𝟏𝟎𝟎𝟎)𝒆−𝟎.𝟎𝟒𝒕 + 𝟒𝟎𝟎
It is the definite Time-Path of Value of Gold-Mine.
𝑮(𝒕) = ⏟𝟏𝟎𝟎𝟎 𝒆−𝟎.𝟎𝟒𝒕 + 𝟒𝟎𝟎
⏟
𝑮𝒄 𝑮𝒑
−𝟎.𝟎𝟒𝒕
Dynamic Stability: The expression; 𝑮(𝒕) = 𝟏𝟎𝟎𝟎 𝒆 + 𝟒𝟎𝟎 has an exponential decay term
in the box. i.e.:
______________________________________________________________________________
Time-Path
(Deviation)
𝒕 𝑮𝒄 = 𝟏𝟎𝟎𝟎 𝒆−𝟎.𝟎𝟒𝒕 𝑮(𝒕) = 𝟏𝟎𝟎𝟎 𝒆−𝟎.𝟎𝟒𝒕 + 𝟒𝟎𝟎
0 1000 1400
1 960.79 1360.789
40 201.90 601.8965
80 40.762 440.7622
120 8.2297 408.2297
160 1.6616 401.6615
200 0.3355 400.3354
240 0.0677 400.0677
280 0.0137 400.0136
320 0.0028 400.0027
360 0.0006 400.0006
… … …
Remarks
over time (𝑮∗ = 𝟒𝟎𝟎)
As 𝒕 → ∞, 𝑮𝒄 → 𝟎 ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ (𝑮∗ )
And when 𝑮𝒄 → 𝟎, (𝑮𝒕 )𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑠
Graphically
The straight line shows the equilibrium value of Value of Oil-Well (𝑮∗ ), while curve
shows the deviation from equilibrium (𝑮𝒄 ), and this deviation decreases over time and
TOPIC 096: HOTELLING'S RULE USING DIFFERENTIAL EQUATIONS
Hotelling’s rule is named after statistician and economic theorist Harold Hotelling.
It studies the net revenue of extraction from a nonrenewable natural resource as a function of
time 𝝅(𝒕).
Hotelling found the optimum point to be where net revenue grows exactly at the rate of interest 𝒓.
Thus, the simple rule is:
𝝅′ (𝒕)
= ⏟ 𝒓
𝝅(𝒕)
⏟ 𝑟𝑎𝑡𝑒 𝑜𝑓
𝑁𝑒𝑡 𝐼𝑛𝑡𝑒𝑟𝑒𝑠𝑡
𝑟𝑒𝑣𝑒𝑛𝑢𝑒
______________________________________________________________________________
Rewriting the Hotelling’s Rule:

𝝅′ (𝒕)
=𝒓
𝝅(𝒕)
𝝅′ (𝒕) = 𝒓 × 𝝅(𝒕)
𝒅{𝝅(𝒕)}
= 𝒓 × 𝝅(𝒕)
𝒅𝒕
In simple terms:
𝒅𝝅
= 𝒓𝝅
𝒅𝒕
Rewriting to compare with standard form:
𝒅𝝅
− 𝒓𝝅 = 𝟎
𝒅𝒕
𝒅𝝅
−⏟𝒓𝝅
⏟=⏟ 𝟎
⏟
𝒅𝒕 𝑎 𝑦 𝑏
𝒅𝒚
𝒅𝒕
𝒅𝒚 𝒅𝝅
Here 𝒅𝒕 = 𝒅𝒕 , 𝒚 = 𝝅, 𝒂 = 𝒓, 𝒃 = 𝟎
𝒃 = 𝟎 – Homogenous Case.
𝒂 = 𝒓; interest rate is usually a positive number 𝒓 > 𝟎 i.e. 𝒂 ≠ 𝟎 – Homogenous Case with 𝒂 ≠
𝟎.
Therefore, using the definite Solution of time-path of 𝑯𝑪:
𝒚(𝒕) = {𝒚(𝟎)}𝒆−𝒂𝒕
𝒚(𝒕) = {𝒚(𝟎)} 𝒆− 𝒂 𝒕
Substituting values of 𝒚 and 𝒂, respectively:
𝝅(𝒕) = {𝝅(𝟎)} 𝒆− 𝒓 𝒕
𝝅(𝒕) = {𝝅(𝟎)}𝒆−𝒓𝒕
Definite Time-path for net revenue in Dynamic Sense.
the deviation expression should fade. In homogenous case (𝑯𝑪), the particular solution is zero;
𝝅𝒑 = 𝟎. Therefore, the convergence should take place towards zero (𝟎). In this case, we have a
negative exponent 𝝅(𝒕) = {𝝅(𝟎)}𝒆 −𝒓 𝒕 which shall approach to zero (𝟎), when 𝒕 → ∞. In a
differential equation, deviation component is measured by (𝝅𝒄 ). Therefore, for dynamic stability;
𝝅𝒄 → 𝟎 as 𝒕 → ∞.
(𝒓>𝟎)
⏞
−(𝒓) 𝒕
{𝝅(𝟎)}
⏟ 𝒆
⏟ →𝟎
𝑫𝒆𝒗𝒊𝒂𝒕𝒊𝒐𝒏 𝑬𝒙𝒑𝒐𝒏𝒆𝒏𝒕𝒊𝒂𝒍
𝑫𝒆𝒄𝒂𝒚
[ 𝒐𝒗𝒆𝒓 𝒕𝒊𝒎𝒆 ]
Therefore, the time-path of net revenue shall converge towards equilibrium as time is allowed
to pass indefinitely. Hence the time-path is dynamically stable.
______________________________________________________________________________
TOPIC 097: FUTURE VALUE OF INVESTMENT USING DIFFERENTIAL EQUATIONS
An initial amount 𝑫(𝟎) = $𝟏, 𝟑𝟎𝟎 deposited in a bank is growing according to the equation:
𝒅𝑫
− 𝟎. 𝟎𝟓 𝑫(𝒕) = 𝟏𝟓
𝒅𝒕
a) What is the general time path that describes the growth of this account?
b) What is the future value of this account at time 𝒕 = 𝟔?
𝒅𝑫
Solution: Given the rate of change of value of deposit 𝒅𝒕 − 𝟎. 𝟎𝟓 𝑫 (𝒕) = 𝟏𝟓, we can rewrite to
compare with the standard form of differential equation:
𝒅𝑫
+ 𝟎. 𝟎𝟓 𝑫 = 𝟏𝟓
𝒅𝒕
𝒅𝑫
+ (𝟎. 𝟎𝟓) 𝑫
⏟ = 𝟏𝟓
⏟
⏟ ⏟
𝑑𝑦
𝑑𝑡
𝒅𝒚 𝒅𝑫
Here 𝒅𝒕
= 𝒅𝒕 , 𝒚 = 𝑫, 𝒂 = 𝟎. 𝟎𝟓, 𝒃 = 𝟏𝟓, 𝒚(𝟎) = 𝑫(𝟎) = 𝟏𝟑𝟎𝟎.
As 𝒃 = 𝟏𝟓, so 𝒃 ≠ 𝟎: Non-Homogenous Case.
As 𝒂 = 𝟎. 𝟎𝟓, so 𝒂 ≠ 𝟎: Non-Homogenous Case with 𝒂 ≠ 𝟎.
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
𝟏𝟓 𝟏𝟓
𝑫(𝒕) = {𝑫(𝟎) − } 𝒆− 𝟎.𝟎𝟓 𝒕 +
𝟎. 𝟎𝟓 𝟎. 𝟎𝟓
𝑫(𝒕) = (𝟏𝟑𝟎𝟎 − 𝟑𝟎𝟎)𝒆−𝟎.𝟎𝟓𝒕 + 𝟑𝟎𝟎
𝑫(𝒕) = (𝟏𝟑𝟎𝟎 − 𝟑𝟎𝟎)𝒆−𝟎.𝟎𝟓𝒕 + 𝟑𝟎𝟎
𝑫(𝒕) = (𝟏𝟎𝟎𝟎)𝒆−𝟎.𝟎𝟓𝒕 + 𝟑𝟎𝟎
It is the definite Time-Path of Value of Deposit.
______________________________________________________________________________
𝟏𝟎𝟎𝟎 𝒆−𝟎.𝟎𝟓𝒕 + 𝟑𝟎𝟎

𝑫(𝒕) = ⏟ ⏟
𝑫𝒄 𝑫𝒑
Dynamic Stability: The expression; 𝑫(𝒕) = 𝟏𝟎𝟎𝟎 𝒆−𝟎.𝟎𝟓𝒕 + 𝟑𝟎𝟎 has an exponential decay term
in the box. i.e.:
Time-Path
(Deviation)
𝒕 𝑫𝒄 = 𝟏𝟎𝟎𝟎 𝒆−𝟎.𝟎𝟓𝒕 𝑫(𝒕) = 𝟏𝟎𝟎𝟎 𝒆−𝟎.𝟎𝟓𝒕 + 𝟑𝟎𝟎
0 1000 1300
1 951.2 1251.22
30 223.1 523.130
60 49.78 349.787
90 11.10 311.108
120 2.478 302.478
150 0.553 300.553
180 0.123 300.123
210 0.027 300.027
240 0.006 300.006
270 0.001 300.001
… … …
Remarks
over time (𝑫∗ = 𝟒𝟎𝟎)
As 𝒕 → ∞, 𝑫𝒄 → 𝟎 ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ (𝑫∗ )
And when 𝑫𝒄 → 𝟎, (𝑫𝒕 )𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑠
Graphically
The straight line shows the equilibrium value of Value of Deposit (𝑫∗ ), while curve
shows the deviation from equilibrium (𝑫𝒄 ), and this deviation decreases over time and
TOPIC 098: DYNAMICS OF MONTHLY SALES USING DIFFERENTIAL EQUATIONS
A model predicts that monthly sales, 𝑺(𝒕), of a new product after 𝒕 months satisfy the differential
equation:
𝒅𝑺
= 𝒌(𝑴 − 𝑺) with initial condition 𝑺(𝟎) = 𝑪.
𝒅𝒕
Where 𝒌, 𝑴 and 𝑪 are positive constants.
______________________________________________________________________________
𝒅𝑺
Solution: Given the rate of change of monthly sales; 𝒅𝒕 = 𝒌(𝑴 − 𝑺), we can rewrite it for the sake
of comparison with first order differential equation.
𝒅𝑺
= 𝒌(𝑴 − 𝑺)
𝒅𝒕
𝒅𝑺
= 𝒌𝑴 − 𝒌𝑺
𝒅𝒕
𝒅𝑺
+𝒌⏟⏟𝑺 = 𝒌𝑴
⏟
⏟
𝑑𝑦
𝑑𝑡
𝒅𝒚 𝒅𝑺
Here 𝒅𝒕 = 𝒅𝒕 , 𝒚 = 𝑺, 𝒂 = 𝒌, 𝒃 = 𝒌𝑴, 𝒚(𝟎) = 𝑺(𝟎) = 𝑪
𝒃 = 𝒌𝑴; As 𝒌 > 𝟎 and 𝑴 > 𝟎 (both positive constants), their product (𝒌 × 𝑴) will be non-zero
positive constant. i.e. 𝒃 ≠ 𝟎 – Non-Homogenous Case.
𝒂 = 𝒌; 𝒌 > 𝟎 (non-zero positive constants). i.e. 𝒂 ≠ 𝟎 – Non-Homogenous Case with 𝒂 ≠ 𝟎.
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
𝒌𝑴 − 𝒌 𝒕 𝒌𝑴
𝑺(𝒕) = {𝑺(𝟎) − }𝒆 +
𝒌 𝒌
𝒌𝑴 −𝒌𝒕 𝒌𝑴
𝑺(𝒕) = { 𝑺(𝟎) − }𝒆 +
𝒌 𝒌
𝑺(𝒕) = ( 𝑪 − 𝑴)𝒆−𝒌𝒕 + 𝑴
𝑺(𝒕) = (𝑪 − 𝑴)𝒆−𝒌𝒕 + 𝑴
Definite Time-path for Monthly Sales in Dynamic Sense.
it should converge towards equilibrium. We know that equilibrium is particular solution (𝑺𝒑 ).
equation, deviation component is measured by (𝑺𝒄 ). Therefore, for dynamic stability; 𝑺𝒄 → 𝟎 as
𝒕 → ∞.
(𝑪 − 𝑴)𝒆−𝒌𝒕 + 𝑴
𝑺(𝒕) = ⏟ ⏟
𝑆𝑐 𝑆𝑝
Need for time path of Monthly Sales {𝑺(𝒕)} to converge to equilibrium income (𝑺∗ ) i.e. 𝑺(𝒕) →
𝑺∗ as 𝑺𝒄 → 𝟎 when 𝒕 → ∞.
𝑺𝒄 → 𝟎
(𝑪 − 𝑴)𝒆−𝒌𝒕 → 𝟎
{𝑴 >𝟎} {𝒌>𝟎}
⏞ ⏞
⏟− 𝑴 } 𝒆⏟− 𝒌 𝒕
{𝑪 →𝟎
⏟{𝑺(𝟎)=𝑪}≷𝑺𝒑 𝑫𝒆𝒄𝒂𝒚
𝑫𝒆𝒗𝒊𝒂𝒕𝒊𝒐𝒏 𝒐𝒗𝒆𝒓 𝒕𝒊𝒎𝒆
[ (+𝒗𝒆 𝒐𝒓−𝒗𝒆) ]
______________________________________________________________________________
Therefore,
⏟ − 𝑴)𝒆−𝒌𝒕 + 𝑴
𝑺(𝒕) = (𝑪
→0
(𝑪 − 𝑴 )𝒆−𝒌𝒕 + 𝑴
𝑺(𝒕) = ⏟
→0
−𝒌𝒕
𝑺(𝒕) = (𝑪
⏟ − 𝑺𝒑 ) 𝒆 + 𝑺𝒑
→0
As particular integral is also the equilibrium value of the variable 𝑺𝒑 = 𝑺∗ .
(𝑪 − 𝑺∗ )𝒆−𝒌𝒕 + 𝑺∗
𝑺(𝒕) = ⏟
→0
𝑺(𝒕) → 𝑺∗
Interpretation: It can be said that provided 𝑪 > 𝟎, 𝒌 > 𝟎, regardless of initial national income 𝑺(𝟎)
is greater or smaller than equilibrium monthly sales (𝑺∗ ), the time-path of monthly sales shall
converge towards equilibrium level of monthly sales as time is allowed to pass indefinitely.
Hence the time-path is dynamically stable.
TOPIC 099: DYNAMICS OF POPULATION USING DIFFERENTIAL EQUATIONS
The population, 𝑵, of a small country is currently 𝟓 million. It is expected that this population will
grow according to the differential equation.
𝒅𝑵
= 𝟎. 𝟎𝟏𝟓𝑵
𝒅𝒕
where 𝒕 is measured in years.
Solution: Current population of the country is 𝟓 million. It implies; 𝑵(𝟎) = 𝟓, 𝟎𝟎𝟎, 𝟎𝟎𝟎.
𝒅𝑵
Given the rate of change of population; = 𝟎. 𝟎𝟏𝟓𝑵, rewrite to compare with standard form:
𝒅𝒕
𝒅𝑵
= 𝟎. 𝟎𝟏𝟓𝑵
𝒅𝒕
𝒅𝑵
− 𝟎. 𝟎𝟏𝟓𝑵 = 𝟎
𝒅𝒕
𝒅𝑵
+ (−𝟎. 𝟎𝟏𝟓)𝑵 = 𝟎
𝒅𝒕
𝒅𝑵
+ (−𝟎. 𝟎𝟏𝟓) 𝑵 ⏟=⏟ 𝟎
⏟ ⏟
𝒅𝒚
𝒅𝒕
𝒅𝒚 𝒅𝑵
Here = 𝒅𝒕 , 𝒚 = 𝑵, 𝒂 = −𝟎. 𝟎𝟏𝟓, 𝒃 = 𝟎, 𝒚(𝟎) = 𝑵(𝟎) = 𝟓𝟎𝟎𝟎𝟎𝟎𝟎
𝒅𝒕
𝒃 = 𝟎; – Homogenous Case.
𝒂 = −𝟎. 𝟎𝟏𝟓; i.e. 𝒂 ≠ 𝟎 – Homogenous Case with 𝒂 ≠ 𝟎.
𝒚(𝒕) = {𝒚(𝟎)}𝒆−𝒂𝒕
𝒚(𝒕) = {𝒚(𝟎)} 𝒆− 𝒂 𝒕
( − 𝟎.𝟎𝟏𝟓) 𝒕
𝑵(𝒕) = {𝑵(𝟎)} 𝒆 −
𝑵(𝒕) = (𝟓𝟎𝟎𝟎𝟎𝟎𝟎)𝒆𝟎.𝟎𝟏𝟓𝒕
Definite Time-path for net revenue in Dynamic Sense.
______________________________________________________________________________
𝑵𝒑 = 𝟎. Therefore, the convergence should take place towards zero (𝟎). In this case, we have a
positive exponent 𝑵(𝒕) = (𝟓𝟎𝟎𝟎𝟎𝟎𝟎)𝒆 𝟎.𝟎𝟏𝟓 𝒕 which shall approach to infinity (∞), when 𝒕 → ∞. In
a differential equation, deviation component is measured by (𝑵𝒄 ). Here, there is dynamic
instability; 𝑵𝒄 → ∞ as 𝒕 → ∞.
(𝒓>𝟎)
⏞
(𝟎.𝟎𝟏𝟓) 𝒕
{𝟓𝟎𝟎𝟎𝟎𝟎𝟎}
⏟ 𝒆
⏟ →𝟎
𝑮𝒓𝒐𝒘𝒕𝒉
Therefore, the time-path of population shall diverge away from equilibrium as time is allowed
to pass indefinitely. Hence the time-path is dynamically instable.
______________________________________________________________________________
Policy Problem: From policy perspective, its important to know about the time the population
will take to get double. i.e. {𝑵(𝑻) = 𝟏𝟎, 𝟎𝟎𝟎, 𝟎𝟎𝟎}𝑻=?
Using the solution 𝑵(𝒕) = (𝟓𝟎𝟎𝟎𝟎𝟎𝟎)𝒆𝟎.𝟎𝟏𝟓𝒕 for substitution, we get:
𝑵(𝑻) = (𝟓𝟎𝟎𝟎𝟎𝟎𝟎)𝒆𝟎.𝟎𝟏𝟓 (𝑻)
𝟏𝟎𝟎𝟎𝟎𝟎𝟎𝟎 = (𝟓𝟎𝟎𝟎𝟎𝟎𝟎)𝒆𝟎.𝟎𝟏𝟓 𝑻
𝟏𝟎𝟎𝟎𝟎𝟎𝟎𝟎
= 𝒆𝟎.𝟎𝟏𝟓 𝑻
𝟓𝟎𝟎𝟎𝟎𝟎𝟎
𝒆𝟎.𝟎𝟏𝟓 𝑻 = 𝟐
Raising both sides to exponent:
𝒍𝒏(𝒆𝟎.𝟎𝟏𝟓 𝑻 ) = 𝒍𝒏(𝟐)
𝟎. 𝟎𝟏𝟓 𝑻 = 𝟎. 𝟔𝟗𝟑𝟏𝟒𝟕𝟏𝟖𝟎𝟓𝟓𝟗𝟗
𝟎. 𝟔𝟗𝟑𝟏𝟒𝟕𝟏𝟖𝟎𝟓𝟓𝟗𝟗
𝑻 =
𝟎. 𝟎𝟏𝟓
𝑻 = 𝟒𝟔. 𝟐
Interpretation: It will take 𝟒𝟔. 𝟐 𝒚𝒆𝒂𝒓𝒔 approximately for the population of the country to get
double i.e. from 𝟓, 𝟎𝟎𝟎, 𝟎𝟎𝟎 to 𝟏𝟎, 𝟎𝟎𝟎, 𝟎𝟎𝟎.
______________________________________________________________________________
Lesson 24
VARIOUS ECONOMIC DYNAMICS USING DIFFERENTIAL EQUATIONS (CONTINUED)
TOPIC 100: DYNAMICS OF GOVERNMENT SPENDING USING DIFFERENTIAL EQUATIONS
Government expenditure, 𝑮, satisfies the differential equation:

𝒅𝑮
= −𝟎. 𝟎𝟓𝑮
𝒅𝒕
Where, 𝑮 is measured in billions of dollars and time 𝒕 is in years. The initial condition is 𝑮(𝟎) =
𝟓𝟎𝟎.
(a) Solve this equation to express 𝑮 in terms of 𝒕.
(b) Find the value of 𝑮 when 𝒕 = 𝟔. Round your answers to three significant figures.
𝒅𝑮
Solution: Given the rate of change of Government Expenditure; = −𝟎. 𝟎𝟓𝑮, rewrite to compare
𝒅𝒕
with standard form:
𝒅𝑮
= −𝟎. 𝟎𝟓𝑮
𝒅𝒕
𝒅𝑮
+ 𝟎. 𝟎𝟓𝑮 = 𝟎
𝒅𝒕
𝒅𝑮
+ 𝟎. 𝟎𝟓𝑮 = 𝟎
𝒅𝒕
𝒅𝑮
+ (𝟎. 𝟎𝟓) 𝑮
⏟=⏟𝟎
⏟ ⏟
𝒅𝒚
𝒅𝒕
𝒅𝒚 𝒅𝑮
Here = , 𝒚 = 𝑮, 𝒂 = 𝟎. 𝟎𝟓, 𝒃 = 𝟎, 𝒚(𝟎) = 𝑮(𝟎) = 𝟓𝟎𝟎
𝒅𝒕 𝒅𝒕
𝒂 = 𝟎. 𝟎𝟓; i.e. 𝒂 ≠ 𝟎 – Homogenous Case with 𝒂 ≠ 𝟎.
𝒚(𝒕) = {𝒚(𝟎)}𝒆−𝒂𝒕
𝒚(𝒕) = {𝒚(𝟎)} 𝒆− 𝒂 𝒕
𝑮(𝒕) = {𝑮(𝟎)} 𝒆− (𝟎.𝟎𝟓) 𝒕
𝑮(𝒕) = (𝟓𝟎𝟎)𝒆−𝟎.𝟎𝟓𝒕
𝑮(𝒕) = (𝟓𝟎𝟎)𝒆−𝟎.𝟎𝟓𝒕
Definite Time-path for Government Expenditure in Dynamic Sense.
𝑮𝒑 = 𝟎. Therefore, the convergence should take place towards zero (𝟎). In this case, we have a
negative exponent 𝑮(𝒕) = (𝟓𝟎𝟎)𝒆− 𝟎.𝟎𝟓 𝒕 which shall approach to zero (𝟎), when 𝒕 → ∞. In a
differential equation, deviation component is measured by (𝑮𝒄 ). Therefore, for dynamic stability;
𝑮𝒄 → 𝟎 as 𝒕 → ∞.
______________________________________________________________________________
{𝟓𝟎𝟎}
⏟ ⏟−(𝟎.𝟎𝟓) 𝒕
𝒆 →𝟎
Therefore, the time-path of Government Expenditure shall converge towards equilibrium as
time is allowed to pass indefinitely. Hence the time-path is dynamically stable.
Policy Problem: From policy perspective, it’s important to know about the level of government
expenditure after some time.
For example; after 6 years: {𝑮(𝟔) =? }
Using the solution 𝑮(𝒕) = (𝟓𝟎𝟎)𝒆−𝟎.𝟎𝟓𝒕 for substitution, we get:
𝑮(𝟔) = (𝟓𝟎𝟎)𝒆−𝟎.𝟎𝟓 (𝟔)
𝑮(𝟔) = (𝟓𝟎𝟎)𝒆−𝟎.𝟑
𝑮(𝟔) = $ 𝟑𝟕𝟎. 𝟒𝟎𝟗 𝑏𝑖𝑙𝑙𝑖𝑜𝑛
TOPIC 101: DYNAMICS OF FUEL CONSUMPTION USING DIFFERENTIAL EQUATIONS
Let 𝑭(𝒙) denote the number of litre(s) of fuel left in an aircraft’s fuel tank if it has flown 𝒙 km.
Suppose that 𝑭(𝒙) satisfies the following differential equation: 𝑭′ (𝒙) = −𝒂𝑭(𝒙) − 𝒃. Here, the fuel
consumption per km is a constant 𝒃 > 𝟎. The term −𝒂𝑭(𝒙), with 𝒂 > 𝟎, is due to the weight of the
fuel.
Numerically speaking, the differential equation is: 𝑭′ (𝒙) = −𝟎. 𝟎𝟎𝟏𝑭(𝒙) − 𝟖. The initial level of fuel
in the plane is 𝑭(𝟎) = 𝟏𝟐𝟎𝟎𝟎:
Find the solution of the equation with 𝑭(𝟎) = 𝑭𝟎 .
Solution: Rate of change of remaining fuel; 𝑭′ (𝒙) = −𝟎. 𝟎𝟎𝟏𝑭(𝒙) − 𝟖.
Rewriting equation to compare with the standard form:
𝑭′ (𝒙) = −𝟎. 𝟎𝟎𝟏𝑭(𝒙) − 𝟖
𝑭′ (𝒙) + 𝟎. 𝟎𝟎𝟏𝑭(𝒙) = −𝟖
𝒅𝑭
+ (𝟎. 𝟎𝟎𝟏) 𝑭⏟ = −𝟖
⏟
⏟ ⏟
𝒅𝒙 𝑎 𝑦 𝑏
𝑑𝑦
𝑑𝑥
𝒅𝒚 𝒅𝑭
Here 𝒅𝒕
= 𝒅𝒕
, 𝒚 = 𝑭, 𝒂 = 𝟎. 𝟎𝟎𝟏, 𝒃 = −𝟖, 𝒚(𝟎) = 𝑭(𝟎) = 𝟏𝟐𝟎𝟎𝟎.
______________________________________________________________________________

As 𝒃 = −𝟖, so 𝒃 ≠ 𝟎: Non-Homogenous Case.
As 𝒂 = 𝟎. 𝟎𝟎𝟏, so 𝒂 ≠ 𝟎: Non-Homogenous Case with 𝒂 ≠ 𝟎.
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
(−𝟖) (−𝟖)
𝑭(𝒕) = {𝑭(𝟎) − } 𝒆− 𝟎.𝟎𝟎𝟏 𝒕 +
𝟎. 𝟎𝟎𝟏 𝟎. 𝟎𝟎𝟏
( − 𝟖) −(𝟎.𝟎𝟎𝟏)𝒕 ( − 𝟖)
𝑭(𝒕) = {𝟏𝟐𝟎𝟎𝟎 − }𝒆 +
𝟎. 𝟎𝟎𝟏 𝟎. 𝟎𝟎𝟏
𝟖 𝟖
𝑭(𝒕) = {𝟏𝟐𝟎𝟎𝟎 + } 𝒆−(𝟎.𝟎𝟎𝟏)𝒕 −
𝟎. 𝟎𝟎𝟏 𝟎. 𝟎𝟎𝟏
𝑭(𝒕) = {𝟏𝟐𝟎𝟎𝟎 + 𝟖𝟎𝟎𝟎}𝒆−(𝟎.𝟎𝟎𝟏)𝒕 − 𝟖𝟎𝟎𝟎
𝑭(𝒕) = (𝟐𝟎𝟎𝟎𝟎)𝒆−(𝟎.𝟎𝟎𝟏)𝒕 − 𝟖𝟎𝟎𝟎
It is the definite Time-Path of remaining Fuel.
𝟐𝟎𝟎𝟎𝟎 𝒆−𝟎.𝟎𝟎𝟏𝒕 − 𝟖𝟎𝟎𝟎
𝑭(𝒕) = ⏟ ⏟
𝑭𝒄 𝑭𝒑
−𝟎.𝟎𝟎𝟏𝒕
Dynamic Stability: The expression; 𝑭(𝒕) = 𝟐𝟎𝟎𝟎𝟎 𝒆 + 𝟖𝟎𝟎𝟎 has an exponential decay
term in the box. i.e.:
Time-Path
(Deviation)
𝒕 𝑭𝒄 = 𝟐𝟎𝟎𝟎𝟎 𝒆−𝟎.𝟎𝟎𝟏𝒕 𝑭(𝒕) = 𝟐𝟎𝟎𝟎𝟎 𝒆−𝟎.𝟎𝟎𝟏𝒕 + 𝟖𝟎𝟎𝟎
0 20000 28000
1 19980.0 27980.0
2 19960.0 27960.0
3 19940.0 27940.0
4 19920.1 27920.1
5 19900.2 27900.2
6 19880.3 27880.3
7 19860.4 27860.4
8 19840.6 27840.6
9 19820.8 27820.8
10 19800.9 27800.9
… … …
Remarks
Positive deviation decays over Time-path slides and converges to equilibrium
time (𝑭∗ = 𝟖𝟎𝟎𝟎)
As 𝒕 → ∞, 𝑭𝒄 → 𝟎 ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ (𝑭∗ )
And when 𝑭𝒄 → 𝟎, (𝑭𝒕 )𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑠
Graphically
______________________________________________________________________________
The straight line shows the equilibrium value of remaining Fuel (𝑭∗ ), while curve shows
the deviation from equilibrium (𝑭𝒄 ), and this deviation decreases over time and leads the
time-path to convergence.
TOPIC 102: DYNAMICS OF UNEMPLOYMENT USING DIFFERENTIAL EQUATIONS
If unemployment data is measured in thousands, unemployment in a small economy is measured

by following differential equation:
𝒅𝑼
= −𝟑𝑼 + 𝟏𝟖𝟎
𝒅𝒕
Here 𝑼 is the unemployment level (in thousands) and 𝒕 is time in years. If unemployment is 𝟏𝟓𝟎
(in thousands) at the beginning of analysis, find the time-path of unemployment in the economy.
Solution: Given the rate of unemployment in the economy;
𝒅𝑼
= −𝟑𝑼 + 𝟏𝟖𝟎
𝒅𝒕
we can rewrite compare it with the standard form of differential equation.
𝒅𝑼
+ 𝟑𝑼 = 𝟏𝟖𝟎
𝒅𝒕
𝒅𝑼
+⏟ 𝟑𝑼⏟ = 𝟏𝟖𝟎
⏟
⏟
𝑑𝑦
𝑑𝑡
𝒅𝒚 𝒅𝑼
Here = 𝒅𝒕 , 𝒚 = 𝑼, 𝒂 = 𝟑, 𝒃 = 𝟏𝟖𝟎, 𝒚(𝟎) = 𝑼(𝟎) = 𝟏𝟓𝟎.
𝒅𝒕
As 𝒃 = 𝟏𝟖𝟎, so 𝒃 ≠ 𝟎: Non-Homogenous Case.
As 𝒂 = 𝟑, so 𝒂 ≠ 𝟎: Non-Homogenous Case with 𝒂 ≠ 𝟎.
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
𝟏𝟖𝟎 − 𝟑 𝒕 𝟏𝟖𝟎
𝑼(𝒕) = {𝑼(𝟎) − }𝒆 +
𝟑 𝟑
𝟏𝟖𝟎 − 𝟑 𝒕 𝟏𝟖𝟎
𝑼(𝒕) = {𝟏𝟓𝟎 − }𝒆 +
𝟑 𝟑
−𝟑𝒕
𝑼(𝒕) = (𝟏𝟓𝟎 − 𝟔𝟎)𝒆 + 𝟔𝟎
______________________________________________________________________________
𝑼(𝒕) = (𝟗𝟎)𝒆−𝟑𝒕 + 𝟔𝟎
𝑼(𝒕) = (𝟗𝟎)𝒆−𝟑𝒕 + 𝟔𝟎
It is the definite Time-Path of Unemployment.
𝑼(𝒕) = ⏟𝟗𝟎 𝒆−𝟑𝒕 + 𝟔𝟎
⏟
𝑼𝒄 𝑼𝒑
−𝟑𝒕
Dynamic Stability: The expression; 𝑼(𝒕) = 𝟗𝟎 𝒆 + 𝟔𝟎 has an exponential decay term in the
box. i.e.:
Time-Path
(Deviation)
𝒕 𝑼𝒄 = 𝟗𝟎 𝒆−𝟑𝒕 𝑼(𝒕) = 𝟗𝟎 𝒆−𝟑𝒕 + 𝟔𝟎
0 90 150
1 4.480836153108 64.480836153108
2 0.223087695900 60.223087695900
3 0.011106882368 60.011106882368
4 0.000552979112 60.000552979112
5 0.000027531209 60.000027531209
6 0.000001370698 60.000001370698
7 0.000000068243 60.000000068243
8 0.000000003398 60.000000003398
9 0.000000000169 60.000000000169
10 0.000000000008 60.000000000008
… … …
Remarks
Positive deviation decays over time Time-path slides and converges to equilibrium
(𝑼∗ = 𝟔𝟎)
As 𝒕 → ∞, 𝑼𝒄 → 𝟎 ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ (𝑼∗ )
And when 𝑼𝒄 → 𝟎, (𝑼𝒕 )𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑠
Graphically
______________________________________________________________________________
The straight line shows the equilibrium value of Unemployment (𝑼∗ ), while curve shows
the deviation from equilibrium (𝑼𝒄 ), and this deviation decreases over time and leads the
TOPIC 103: LIMITED GROWTH USING DIFFERENTIAL EQUATIONS
Situations where growth of a variable is fast initially but levels-off (stabilize) just short of a
maximum limit. Examples include mobile adoption, wireless internet etc.
𝒅𝒚
Using differential equations of the type 𝒅𝒕 = 𝒓(𝑨 − 𝒚) represent a similar situation where 𝑨 is the
maximum limit. Graphically it is:
The rate at which the volume of sales (𝑸) for a new type of printer increases after an advertising
𝒅𝑸
campaign is given by the equation 𝒅𝒕 = 𝟎. 𝟎𝟒(𝟕𝟎𝟎– 𝑸), given that 𝑸 = 𝟎 at 𝒕 = 𝟎. 𝑸 is the number
of printers sold, 𝒕 is time in years.
(a) Solve the differential equation to obtain an expression for 𝑸 in terms of 𝒕.
(b) Sketch the solution curve, hence describe verbally how the volume of sales should increase
in time.
𝒅𝑸
Solution: Given the rate of change of Volume of Sales; 𝒅𝒕 = 𝟎. 𝟎𝟒(𝟕𝟎𝟎– 𝑸), we can rewrite it to
compare with standard form of differential equation.
𝒅𝑸
= 𝟎. 𝟎𝟒(𝟕𝟎𝟎– 𝑸)
𝒅𝒕
𝒅𝑸
= 𝟐𝟖 − 𝟎. 𝟎𝟒𝑸
𝒅𝒕
𝒅𝑸
+⏟𝟎. 𝟎𝟒 𝑸
⏟ = 𝟐𝟖
⏟
⏟
𝑑𝑦
𝑑𝑡
𝒅𝒚 𝒅𝑸
Here 𝒅𝒕
= 𝒅𝒕 , 𝒚 = 𝑸, 𝒂 = 𝟎. 𝟎𝟒, 𝒃 = 𝟐𝟖, 𝒚(𝟎) = 𝑸(𝟎) = 𝟎.
As 𝒃 = 𝟐𝟖, so 𝒃 ≠ 𝟎: Non-Homogenous Case.
______________________________________________________________________________

𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
𝟐𝟖 𝟐𝟖
𝑸(𝒕) = {𝑸(𝟎) − } 𝒆− 𝟎.𝟎𝟒 𝒕 +
𝟎. 𝟎𝟒 𝟎. 𝟎𝟒
𝟐𝟖 𝟐𝟖
𝑸(𝒕) = {𝟎 − } 𝒆−𝟎.𝟎𝟒𝒕 +
𝟎. 𝟎𝟒 𝟎. 𝟎𝟒
𝑸(𝒕) = (𝟎 − 𝟕𝟎𝟎)𝒆−𝟎.𝟎𝟒𝒕 + 𝟕𝟎𝟎
𝑸(𝒕) = (−𝟕𝟎𝟎)𝒆−𝟎.𝟎𝟒𝒕 + 𝟕𝟎𝟎
It is the definite Time-Path of Quantity Supplied.
−𝟕𝟎𝟎 𝒆−𝟎.𝟎𝟒𝒕 + 𝟕𝟎𝟎
𝑸(𝒕) = ⏟ ⏟
𝑸𝒄 𝑸𝒑
Dynamic Stability: The expression; 𝑸(𝒕) = −𝟕𝟎𝟎 𝒆−𝟎.𝟎𝟒𝒕 + 𝟕𝟎𝟎 has an exponential decay term
in the box. i.e.:
Time-Path
(Deviation)
𝒕 𝑸𝒄 = −𝟕𝟎𝟎 𝒆−𝟎.𝟎𝟒𝒕 𝑸(𝒕) = −𝟕𝟎𝟎 𝒆−𝟎.𝟎𝟒𝒕 + 𝟕𝟎𝟎
0 -700 0
1 -672.552 27.447392
50 -94.7346 605.26530
100 -12.8209 687.17905
150 -1.73512 698.26487
200 -0.23482 699.76517
250 -0.03177 699.96822
300 -0.00430 699.99569
350 -0.00058 699.99941
400 -0.00007 699.99992
450 -0.00001 699.99998
… … …
Remarks
Negative deviation decays Time-path rises and converges to equilibrium
over time (𝑸∗ = 𝟕𝟎𝟎)
As 𝒕 → ∞, 𝑸𝒄 → 𝟎 ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ (𝑸∗ )
And when 𝑸𝒄 → 𝟎, (𝑸𝒕 )𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑠
Graphically
______________________________________________________________________________
The straight line shows the equilibrium value of Sales (𝑸∗ ), while curve shows the
deviation from equilibrium (𝑸𝒄 ), and this deviation decreases over time and leads the
TOPIC 104: DYNAMICS OF ORE MINING USING DIFFERENTIAL EQUATIONS
𝒅𝑴
The rate at which an ore is mined is given by the differential equation 𝒅𝒕 = 𝟎. 𝟎𝟐𝑴. Solve the
differential equation, if the total reserves of ore in 𝟏𝟗𝟖𝟗 were 𝟒𝟎𝟎 million tons.
𝒅𝑴
Solution: Given the rate of change of Ore Mining ( = 𝟎. 𝟎𝟐𝑴), we can rearrange it to compare
𝒅𝒕
with the standard form of differential equations.
𝒅𝑴
= 𝟎. 𝟎𝟐𝑴
𝒅𝒕
𝒅𝑴
− 𝟎. 𝟎𝟐𝑴 = 𝟎
𝒅𝒕
𝒅𝑴
−⏟𝟎. 𝟎𝟐 𝑴
⏟ =⏟𝟎
⏟𝒅𝒕 𝑎 𝑦 𝑏
𝑑𝑦
𝑑𝑡
𝒅𝒚 𝒅𝑴
Here = , 𝒚 = 𝑴, 𝒂 = −𝟎. 𝟎𝟐, 𝒃 = 𝟎, 𝒚(𝟎) = 𝑴(𝟎) = 𝟒𝟎𝟎
𝒅𝒕 𝒅𝒕
𝒂 = −𝟎. 𝟎𝟐; i.e. 𝒂 ≠ 𝟎 – Homogenous Case with 𝒂 ≠ 𝟎.
𝒚(𝒕) = {𝒚(𝟎)}𝒆−𝒂𝒕
𝒚(𝒕) = {𝒚(𝟎)} 𝒆− 𝒂 𝒕
( − 𝟎.𝟎𝟐) 𝒕
𝑴(𝒕) = {𝑴(𝟎)} 𝒆 −
𝑴(𝒕) = (𝟒𝟎𝟎)𝒆𝟎.𝟎𝟐𝒕
Definite Time-path for Ore Mining in Dynamic Sense.
𝑴𝒑 = 𝟎. Therefore, the convergence should take place towards zero (𝟎). In this case, we have a
positive exponent 𝑴(𝒕) = (𝟒𝟎𝟎)𝒆 𝟎.𝟎𝟐 𝒕 which shall approach to infinity (∞), when 𝒕 → ∞. In a
______________________________________________________________________________
differential equation, deviation component is measured by (𝑴𝒄 ). Here, there is dynamic instability;
𝑴𝒄 → ∞ as 𝒕 → ∞.
{𝟒𝟎𝟎}
⏟ ⏟(𝟎.𝟎𝟐)
𝒆 𝒕
→𝟎
𝑮𝒓𝒐𝒘𝒕𝒉
Therefore, the time-path of Ore Mining shall diverge away from equilibrium as time is allowed
to pass indefinitely. Hence the time-path is dynamically instable.
TOPIC 105: DYNAMICS OF POLLUTANT USING DIFFERENTIAL EQUATIONS
𝒅(𝑷)
The amount of sulphur dioxide (in tons) present in the atmosphere is given by the equation 𝒅𝒕 =
−𝟎. 𝟐𝑷 (𝒕 is in hours). If 𝟓𝟎 ton(s) is released into the atmosphere (i.e., 𝑷 = 𝟓𝟎 at 𝒕 = 𝟎), derive
an expression for the amount of sulphur dioxide present at any time 𝒕.
𝒅(𝑷)
Solution: Given the rate of change of Sulphur Dioxide (𝑺𝑶𝟐 ): 𝒅𝒕 = −𝟎. 𝟐𝑷, we can rewrite it to
compare it with the standard form of differential equation.
𝒅(𝑷)
= −𝟎. 𝟐𝑷
𝒅𝒕
𝒅𝑷
+ 𝟎. 𝟐𝑷 = 𝟎
𝒅𝒕
______________________________________________________________________________
𝒅𝑷
+ 𝟎.
⏟ 𝟐𝑷
⏟=⏟
𝟎
⏟
𝑑𝑦
𝑑𝑡
𝒅𝒚 𝒅𝑷
Here = , 𝒚 = 𝑷, 𝒂 = 𝟎. 𝟐, 𝒃 = 𝟎, 𝒚(𝟎) = 𝑷(𝟎) = 𝟓𝟎
𝒅𝒕 𝒅𝒕
𝒂 = 𝟎. 𝟐; i.e. 𝒂 ≠ 𝟎 – Homogenous Case with 𝒂 ≠ 𝟎.
𝒚(𝒕) = {𝒚(𝟎)}𝒆−𝒂𝒕
𝒚(𝒕) = {𝒚(𝟎)} 𝒆− 𝒂 𝒕
𝑷(𝒕) = {𝑷(𝟎)} 𝒆− (𝟎.𝟐) 𝒕
𝑷(𝒕) = (𝟓𝟎)𝒆−𝟎.𝟐𝒕
𝑷(𝒕) = (𝟓𝟎)𝒆−𝟎.𝟐𝒕
Definite Time-path for Government Expenditure in Dynamic Sense.
𝑷𝒑 = 𝟎. Therefore, the convergence should take place towards zero (𝟎). In this case, we have a
negative exponent 𝑷(𝒕) = (𝟓𝟎)𝒆− 𝟎.𝟐 𝒕 which shall approach to zero (𝟎), when 𝒕 → ∞. In a
differential equation, deviation component is measured by (𝑷𝒄 ). Therefore, for dynamic stability;
𝑷𝒄 → 𝟎 as 𝒕 → ∞.
{𝟓𝟎}
⏟ ⏟−(𝟎.𝟐) 𝒕
𝒆 →𝟎
Therefore, the time-path of Pollutant shall converge towards equilibrium as time is allowed to
pass indefinitely. Hence the time-path is dynamically stable.
______________________________________________________________________________
TOPIC 106: DYNAMICS OF FISH STOCK USING DIFFERENTIAL EQUATIONS
𝒅𝑭
The rate at which the fish stocks increase in an inland lake is given by the equation 𝒅𝒕 =
𝟎. 𝟏𝟐(𝟖𝟎 – 𝑭), where 𝑭 is in thousands, 𝒕 is in years. (a) Derive an expression for the total number
of fish present at any time 𝒕 given 𝑭 = 𝟎 at 𝒕 = 𝟎.
(b) Graph the total fish stock as a function of time from 𝒕 = 𝟎 to 𝒕 = 𝟑𝟎. Does the lake appear
to have a maximum carrying capacity?
(c) Calculate the number of years taken for the fish stocks to reach 40000.
𝒅𝑭
Solution: Given the rate of change of fish stock; 𝒅𝒕 = 𝟎. 𝟏𝟐(𝟖𝟎 – 𝑭), we can rearrange it to
𝒅𝑭
= 𝟎. 𝟏𝟐(𝟖𝟎 – 𝑭)
𝒅𝒕
𝒅𝑭
= 𝟗. 𝟔 − 𝟎. 𝟏𝟐𝑭
𝒅𝒕
𝒅𝑭
+⏟𝟎. 𝟏𝟐 𝑭
⏟ = 𝟗.
⏟ 𝟔
⏟
𝑑𝑦
𝑑𝑡
𝒅𝒚 𝒅𝑭
Here = 𝒅𝒕 , 𝒚 = 𝑭, 𝒂 = 𝟎. 𝟏𝟐, 𝒃 = 𝟗. 𝟔, 𝒚(𝟎) = 𝑭(𝟎) = 𝟎.
𝒅𝒕
As 𝒃 = 𝟗. 𝟔, so 𝒃 ≠ 𝟎: Non-Homogenous Case.
As 𝒂 = 𝟎. 𝟏𝟐, so 𝒂 ≠ 𝟎: Non-Homogenous Case with 𝒂 ≠ 𝟎.
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
______________________________________________________________________________

𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆− 𝒂 𝒕 +
𝒂 𝒂
𝟗. 𝟔 𝟗. 𝟔
𝑭(𝒕) = {𝑭(𝟎) − } 𝒆− 𝟎.𝟏𝟐 𝒕 +
𝟎. 𝟏𝟐 𝟎. 𝟏𝟐
𝟗. 𝟔 −𝟎.𝟏𝟐𝒕 𝟗. 𝟔
𝑭(𝒕) = {𝟎 − }𝒆 +
𝟎. 𝟏𝟐 𝟎. 𝟏𝟐
−𝟎.𝟏𝟐𝒕
𝑭(𝒕) = (𝟎 − 𝟖𝟎)𝒆 + 𝟖𝟎
𝑭(𝒕) = (−𝟖𝟎)𝒆−𝟎.𝟏𝟐𝒕 + 𝟖𝟎
It is the definite Time-Path of Fish Stock.
−𝟖𝟎 𝒆−𝟎.𝟏𝟐𝒕 + 𝟖𝟎
𝑭(𝒕) = ⏟ ⏟
𝑭𝒄 𝑭𝒑
−𝟎.𝟏𝟐𝒕
Dynamic Stability: The expression; 𝑭(𝒕) = −𝟖𝟎 𝒆 + 𝟖𝟎 has an exponential decay term in
the box. i.e.:
Time-Path
(Deviation)
𝒕 𝑭𝒄 = −𝟖𝟎 𝒆−𝟎.𝟏𝟐𝒕 𝑭(𝒕) = −𝟖𝟎 𝒆−𝟎.𝟏𝟐𝒕 + 𝟖𝟎
0 -80 0
1 -70.95 9.0464
10 -24.10 55.904
20 -7.257 72.743
30 -2.186 77.814
40 -0.658 79.342
50 -0.198 79.802
60 -0.060 79.940
70 -0.018 79.982
80 -0.005 79.995
90 -0.002 79.998
… … …
______________________________________________________________________________
Remarks
Negative deviation decays over time Time-path rises and converges to equilibrium
(𝑭∗ = 𝟖𝟎)
As 𝒕 → ∞, 𝑭𝒄 → 𝟎 ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ (𝑭∗ )
And when 𝑭𝒄 → 𝟎, (𝑭𝒕 )𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑠
Graphically
The straight line shows the equilibrium value of Fish Stock (𝑭∗ ), while curve shows the
deviation from equilibrium (𝑭𝒄 ), and this deviation decreases over time and leads the
______________________________________________________________________________
Lesson 25
EXACT DIFFERENTIAL EQUATIONS
TOPIC 107: FIRST-ORDER DIFFERENTIAL EQUATIONS: VARIABLE TERM & VARIABLE

COEFFICIENT: HOMOGENEOUS CASE

𝒅𝒚
+ 𝒖(𝒕) ∙ 𝒚 = 𝒘(𝒕)
𝒅𝒕
Here, 𝒖(𝒕) and 𝒘(𝒕) are functions of time and not constants i.e. 𝒕𝒏 when 𝒏 ≠ 𝟎.
Depending on the values of 𝒖(𝒕) and 𝒘(𝒕), we get 2 possible cases:
Case – I: 𝒖(𝒕) ≠ 𝟎, 𝒘(𝒕) = 𝟎 [Homogenous Case]
Case – II: 𝒖(𝒕) ≠ 𝟎, 𝒘(𝒕) ≠ 𝟎 [Non-Homogenous Case]
Case – I: Homogenous Case: 𝒖(𝒕) ≠ 𝟎, 𝒘(𝒕) = 𝟎

𝒅𝒚
Standard form of First-Order Differential Equation 𝒅𝒕 + 𝒖(𝒕)𝒚 = 𝒘(𝒕), becomes:
𝒅𝒚
+ 𝒖(𝒕)
⏟ 𝒚
⏟ =𝟎
𝒅𝒕 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒 𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡
𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑣𝑎𝑟𝑎𝑖𝑏𝑙𝑒
𝒅𝒚
= −𝒖(𝒕)𝒚
𝒅𝒕
𝟏 𝒅𝒚
= −𝒖(𝒕)
𝒚 𝒅𝒕
𝒅𝒚
( )
𝒅𝒕
= −𝒖(𝒕)
𝒚
Integrating w.r.t. 𝒕
𝒅𝒚
( )
𝒅𝒕
∫ 𝒅𝒕 = − ∫ 𝒖(𝒕) 𝒅𝒕
𝒚
(𝒅𝒚)
∫ = − ∫ 𝒖(𝒕) 𝒅𝒕
𝒚
𝟏
∫ ( ) 𝒅𝒚 = − ∫ 𝒖(𝒕) 𝒅𝒕
𝒚
𝒍𝒏|𝒚| + 𝒄 = − ∫ 𝒖(𝒕) 𝒅𝒕
𝒍𝒏|𝒚| = − ∫ 𝒖(𝒕) 𝒅𝒕 − 𝒄
𝒍𝒏|𝒚| = − ∫ 𝒖(𝒕) 𝒅𝒕 − 𝒄
To get rid of natural logarithm, raise exponent on both sides:
𝒆𝒍𝒏|𝒚| = 𝒆{− ∫ 𝒖(𝒕) 𝒅𝒕 − 𝒄}
𝒚 = 𝒆− ∫ 𝒖(𝒕) 𝒅𝒕 𝒆−𝒄
𝒚(𝒕) = 𝒆− ∫ 𝒖(𝒕) 𝒅𝒕 𝑨 [∵ 𝒆−𝒄 = 𝑨] 𝑦𝑒𝑡 𝑎𝑛𝑜𝑡ℎ𝑒𝑟 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
𝒚(𝒕) = 𝑨𝒆− ∫ 𝒖(𝒕) 𝒅𝒕 (General Solution)
Finding definite solution can be easier once ∫ 𝒖(𝒕) 𝒅𝒕 has a specific form of 𝒖(𝒕).
______________________________________________________________________________
𝒅𝒚
Numerical Example: Find the general solution of the equation 𝒅𝒕 + 𝟑𝒕𝟐 𝒚 = 𝟎
Solution: Comparing the general form of first order differential equation (variable coefficient)
with given specific form:
𝒅𝒚
+ 𝒖(𝒕) × 𝒚 = 𝒘(𝒕)
𝒅𝒕
𝒅𝒚
+ 𝟑𝒕𝟐 × 𝒚 = 𝟎
𝒅𝒕
Here, 𝒖(𝒕) = 𝟑𝒕𝟐 , 𝒘(𝒕) = 𝟎.
Using the formula for general solution:
− 𝒖(𝒕) 𝒅𝒕
𝒚(𝒕) = 𝑨𝒆 ∫
Solving the exponent ∫ 𝒖(𝒕) 𝒅𝒕, separately before substituting, will avoid clutter:
= ∫ 𝒖(𝒕) 𝒅𝒕
= ∫ 𝟑𝒕𝟐 𝒅𝒕
= 𝟑 ∫ 𝒕𝟐 𝒅𝒕
𝒕𝟑
= 𝟑( )+ 𝒄
𝟑
∫ 𝒖(𝒕) 𝒅𝒕 = 𝒕𝟑 + 𝒄
Substituting the value of exponent back into formula of General Solution:

− 𝒖(𝒕) 𝒅𝒕
𝒚(𝒕) = 𝑨𝒆 ∫
−(𝒕𝟑 +𝒄)
𝟑
𝒚(𝒕) = 𝑨 × 𝒆−𝒕 −𝒄
𝟑
𝒚(𝒕) = 𝑨 × 𝒆−𝒕 𝒆−𝒄
𝟑
𝒚(𝒕) = 𝑨 × 𝒆−𝒄 𝒆−𝒕
𝟑
𝒚(𝒕) = 𝑨 × 𝒆−𝒄 𝒆−𝒕
𝟑
𝒚(𝒕) = 𝑨 × 𝑨′ × 𝒆−𝒕
𝟑
𝒚(𝒕) = 𝑩 × 𝒆−𝒕 [∵ 𝑨 × 𝑨′ = 𝑩]
𝟑
𝒚(𝒕) = 𝑩𝒆−𝒕
𝟑
Interpretation: 𝒚(𝒕) = 𝑩𝒆−𝒕 is the general solution of first order differential equation with
𝒅𝒚
variable coefficient + 𝟑𝒕𝟐 𝒚 = 𝟎 .
𝒅𝒕
TOPIC 108: FIRST-ORDER DIFFERENTIAL EQUATIONS: VARIABLE TERM & VARIABLE
COEFFICIENT: NON-HOMOGENEOUS CASE

𝒅𝒚
+ 𝒖(𝒕) ∙ 𝒚 = 𝒘(𝒕)
𝒅𝒕
Here, 𝒖(𝒕) and 𝒘(𝒕) are functions of time and not constants i.e. 𝒕𝒏 when 𝒏 ≠ 𝟎.
Depending on the values of 𝒖(𝒕) and 𝒘(𝒕), we get 2 possible cases:
Case – I: 𝒖(𝒕) ≠ 𝟎, 𝒘(𝒕) = 𝟎 [Homogenous Case]
Case – II: 𝒖(𝒕) ≠ 𝟎, 𝒘(𝒕) ≠ 𝟎 [Non-Homogenous Case]
______________________________________________________________________________
Case – II: Non-Homogenous Case: 𝒖(𝒕) ≠ 𝟎, 𝒘(𝒕) ≠ 𝟎

𝒅𝒚
+ 𝒖(𝒕)𝒚 = 𝒘(𝒕)
𝒅𝒕
General Solution of this case can be derived using Exact Differential Equations which is yet to
be studied in this course. Therefore, we borrow the General Solution:
𝒚(𝒕) = 𝒆− ∫ 𝒖(𝒕) 𝒅𝒕 {𝑨 + 𝒘(𝒕) ∫ 𝒆∫ 𝒖(𝒕) 𝒅𝒕 𝒅𝒕}
For simplicity, we may use 𝒖 and 𝒘 instead of 𝒖(𝒕) and 𝒘(𝒕), respectively.
𝒚(𝒕) = 𝒆− ∫ 𝒖 𝒅𝒕 {𝑨 + 𝒘 ∫ 𝒆∫ 𝒖 𝒅𝒕 𝒅𝒕}
𝒚(𝒕) = 𝒆− ∫ 𝒖 𝒅𝒕 {𝑨 + 𝒘 ∫ 𝒆∫ 𝒖 𝒅𝒕 𝒅𝒕}
Definite Solution can be found by substituting initial condition.
𝒅𝒚
Numerical Example (p.# 485): Find the general solution of the equation 𝒅𝒕 + 𝟐𝒕𝒚 = 𝒕.
Solution: Comparing the general form of first order differential equation (variable coefficient &
variable constant) with given specific form:
𝒅𝒚
+ 𝒖(𝒕) × 𝒚 = 𝒘(𝒕)
𝒅𝒕
𝒅𝒚
+ 𝟐𝒕 × 𝒚 = 𝒕
𝒅𝒕
Here, 𝒖(𝒕) = 𝟐𝒕, 𝒘(𝒕) = 𝒕.
Using the formula for general solution:
𝒚(𝒕) = 𝒆− ∫ 𝒖(𝒕) 𝒅𝒕 {𝑨 + 𝒘(𝒕) ∫ 𝒆∫ 𝒖(𝒕) 𝒅𝒕 𝒅𝒕}
Solving the exponent ∫ 𝒖(𝒕) 𝒅𝒕, separately before substituting, will avoid clutter:
= ∫ 𝒖(𝒕) 𝒅𝒕
= ∫ 𝟐𝒕 𝒅𝒕
= 𝟐 ∫ 𝒕 𝒅𝒕
𝒕𝟐
= 𝟐( )+ 𝒄
𝟐
∫ 𝒖(𝒕) 𝒅𝒕 = 𝒕𝟐 + 𝒄𝟏
Substituting the value of exponent back into formula of General Solution:

− ∫ 𝒖(𝒕) 𝒅𝒕 𝒖(𝒕) 𝒅𝒕
𝒚(𝒕) = 𝒆 {𝑨 + 𝒘(𝒕) ∫ 𝒆∫ 𝒅𝒕}
−(𝒕𝟐 +𝒄𝟏 ) (𝒕𝟐 +𝒄𝟏 )
𝒚(𝒕) = 𝒆 {𝑨 + 𝒕 ∫ 𝒆 𝒅𝒕}
𝟐 −𝒄 𝟐 +𝒄
𝒚(𝒕) = 𝒆−𝒕 𝟏 {𝑨 + 𝒕 ∫ 𝒆𝒕 𝟏 𝒅𝒕}
𝟐 𝟐
𝒚(𝒕) = 𝒆−𝒕 𝒆−𝒄𝟏 {𝑨 + 𝒕 ∫ 𝒆𝒕 𝒆𝒄𝟏 𝒅𝒕}
𝟐 𝟐
𝒚(𝒕) = 𝒆−𝒕 𝒆−𝒄𝟏 {𝑨 + 𝒕 ∫ 𝒆𝒕 𝒆𝒄𝟏 𝒅𝒕}
𝟐 𝟐
𝒚(𝒕) = 𝒆−𝒕 𝒆−𝒄𝟏 {𝑨 + 𝒆𝒄𝟏 × 𝒕 × ∫ 𝒆𝒕 𝒅𝒕}
______________________________________________________________________________
𝒆𝒇(𝒙)
Recalling Natural Exponential Function Rule: ∫ 𝒆𝒇(𝒙) 𝒅𝒙 = 𝒇′ (𝒙)
𝟐
𝒆𝒕
Here 𝒇(𝒙) = 𝒕𝟐 , so 𝒇′ (𝒙) = 𝟐𝒕. Therefore, ∫ 𝒆𝒇(𝒙) 𝒅𝒙 =
𝟐𝒕
𝟐
𝟐 𝒆𝒕
𝒚(𝒕) = 𝒆−𝒕 𝒆−𝒄𝟏 {𝑨 + 𝒆𝒄𝟏 × 𝒕 × ( ) + 𝒄𝟐 }
𝟐𝒕
𝟐
−𝒕𝟐 −𝒄𝟏 𝒄𝟏
𝒆𝒕
𝒚(𝒕) = 𝒆 𝒆 {𝑨 + 𝒆 × ( ) + 𝒄𝟐 }
𝟐
𝟐
𝟐 𝟐 𝒆𝒕 𝟐
𝒚(𝒕) = 𝑨𝒆−𝒕 𝒆−𝒄𝟏 + 𝒆−𝒕 𝒆−𝒄𝟏 × 𝒆𝒄𝟏 × ( ) + 𝒆−𝒕 𝒆−𝒄𝟏 × 𝒄𝟐
𝟐
𝟐 𝟐 𝟐 𝟏 𝟐
𝒚(𝒕) = 𝑨𝒆−𝒕 𝒆−𝒄𝟏 + 𝒆−𝒄𝟏 𝒆𝒄𝟏 × 𝒆−𝒕 × 𝒆𝒕 ( ) + 𝒆−𝒕 𝒆−𝒄𝟏 × 𝒄𝟐
𝟐
𝟐 𝟏 𝟐
𝒚(𝒕) = 𝑨𝒆−𝒕 𝒆−𝒄𝟏 + 𝟏 × 𝟏 × ( ) + 𝒆−𝒕 × 𝒆−𝒄𝟏 𝒄𝟐
𝟐
𝟐 𝟏 𝟐
𝒚(𝒕) = 𝑨𝒆−𝒕 𝒆−𝒄𝟏 + ( ) + 𝒆−𝒕 × 𝒆−𝒄𝟏 𝒄𝟐
𝟐
−𝒄𝟏 −𝒕𝟐 𝟐 𝟏
𝒚(𝒕) = 𝑨𝒆 𝒆 + 𝒆−𝒄𝟏 𝒄𝟐 × 𝒆−𝒕 +
𝟐
𝟐 𝟐 𝟏
𝒚(𝒕) = 𝑨𝒆−𝒄𝟏 𝒆−𝒕 + 𝒆−𝒄𝟏 𝒄𝟐 × 𝒆−𝒕 +
𝟐
𝟐 𝟏
𝒚(𝒕) = (𝑨𝒆−𝒄𝟏 + 𝒄𝟑 ) × 𝒆−𝒕 +
𝟐
−𝒄𝟏 −𝒕𝟐
𝟏
𝒚(𝒕) = (𝑨𝒆 + 𝒄𝟑 ) × 𝒆 +
𝟐
−𝒕𝟐
𝟏
𝒚(𝒕) = (𝑩) × 𝒆 +
𝟐
𝟐 𝟏
𝒚(𝒕) = 𝑩𝒆−𝒕 +
𝟐
𝟐 𝟏
Interpretation: 𝒚(𝒕) = 𝑩𝒆−𝒕 + 𝟐 is the General Solution of the first order differential
𝒅𝒚
equation 𝒅𝒕
+ 𝟐𝒕𝒚 = 𝒕 .
TOPIC 109: NON-LINEAR DIFFERENTIAL EQUATION: EXACT DIFFERENTIAL EQUATIONS
Nonlinear differential equation can have power of dependent variable [here, (𝒚) ] greater/other
than (𝟏). Nonlinear differential equations can be solved using either of three (𝟑) methods:
I- Exact Differential Equation Method
II- Separable Variables Method
III- Bernoulli Equation Method
Caveat: Even though our immediate purpose is to use it to solve & linear differential equation, an
exact differential equation can be either linear or nonlinear by itself.
I – Exact Differential Equations

Given a function of two variables 𝑭(𝒚, 𝒕) = 𝒌, its total differential is:
𝒅{𝑭(𝒚, 𝒕)} = 𝒅(𝒌)
______________________________________________________________________________
𝜹𝑭 𝜹𝑭
( ) 𝒅𝒚 + ( ) 𝒅𝒕 = 𝟎
𝜹𝒚 𝜹𝒕
Since both variables (𝒚 &𝒕) are on one side of equation, the differential of other side (some
constant) will be equal to zero:
𝜹𝑭 𝜹𝑭
( ) 𝒅𝒚 + ( ) 𝒅𝒕 = ⏟ 𝟎
⏟𝜹𝒚 𝜹𝒕 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑙 𝑜𝑓
𝑡𝑜𝑡𝑎𝑙 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑙 𝑜𝑓 𝑎 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 (𝒌)
𝑡ℎ𝑒 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 {𝑭(𝒚,𝒕)}
Interpretation: The function { 𝑭(𝒚, 𝒕) = 𝒌} represents an Exact Differential Equation (EDE) as its
𝜹𝑭 𝜹𝑭
total differential satisfies the equation, { (𝜹𝒚) 𝒅𝒚 + ( 𝜹𝒕 ) 𝒅𝒕 = 𝟎}:
Development of Exactness Test: A test can be developed to check if a differential equation is
an Exact Differential Equation or not.
Borrowing the result of total differential of 𝑭(𝒚, 𝒕) = 𝒌:
𝒅{𝑭(𝒚, 𝒕)} = 𝟎
𝜹𝑭 𝜹𝑭
( ) 𝒅𝒚 + ( ) 𝒅𝒕 = 𝟎
𝜹𝒚 𝜹𝒕
𝜹𝑭 𝜹𝑭
( ) 𝒅𝒚 + ( ) 𝒅𝒕 = 𝟎
𝜹𝒚 𝜹𝒕
𝜹𝑭 𝜹𝑭
As (𝜹𝒚) and ( 𝜹𝒕 ) are given in every specific form of EDE, we develop condition around these
derivative terms.
Resorting to Young’s Theorem for a function; 𝑭(𝒙, 𝒚):
𝑭𝒙𝒚 = 𝑭𝒚𝒙
𝜹 𝜹𝑭 𝜹 𝜹𝑭
( )= ( )
𝜹𝒙 𝜹𝒚 𝜹𝒚 𝜹𝒙
However, in current case, Young’s Theorem is for function; 𝑭(𝒚, 𝒕):
𝑭𝒕𝒚 = 𝑭𝒚𝒕
( )= ( )
𝜹𝒕 𝜹𝒚 𝜹𝒚 𝜹𝒕
( ) = ( )
𝜹𝒕 𝜹𝒚 𝜹𝒚 𝜹𝒕
𝜹𝑭 𝜹𝑭
Substituting (𝜹𝒚) = 𝑴 and ( 𝜹𝒕 ) = 𝑵 in two of the following equations:
Young’s Theorem Exact Differential Equation
𝜹 𝜹𝑭 𝜹 𝜹𝑭 𝜹𝑭 𝜹𝑭
( ) = ( ) ( ) 𝒅𝒚 + ( ) 𝒅𝒕 = 𝟎
𝜹𝒕 𝜹𝒚 𝜹𝒚 𝜹𝒕 𝜹𝒚 𝜹𝒕
𝜹 𝜹
𝑴 = 𝑵
𝜹𝒕 𝜹𝒚 (𝑴) 𝒅𝒚 + (𝑵) 𝒅𝒕 = 𝟎
𝜹 𝜹
(𝑴) = (𝑵)
𝜹𝒕 𝜹𝒚 (𝑴)𝒅𝒚 + (𝑵)𝒅𝒕 = 𝟎
Interpretation:
Following Young’s theorem, the Exactness Test is Exact Differential Equation is molded
developed for Exact Differential Equation. for Exactness Test.
______________________________________________________________________________
TOPIC 110: EXACT DIFFERENTIAL EQUATIONS: NUMERICAL
Given a differential equation: 𝟐𝒚𝒕 𝒅𝒚 + 𝒚𝟐 𝒅𝒕 = 𝟎.

Comparing with the standard form of exact differential equation:
𝑴 𝒅𝒚 + 𝑵 𝒅𝒕 = 𝟎
𝟐𝒚𝒕 𝒅𝒚 + 𝒚𝟐 𝒅𝒕 = 𝟎
𝒅 𝒅
Here, 𝑴 = 𝟐𝒚𝒕 and 𝑵 = 𝒚𝟐 , we can employ Exactness Text: (𝑴) = (𝑵):
𝒅𝒕 𝒅𝒚
𝒅 𝒅
(𝑴) = (𝑵)
𝒅𝒕 𝒅𝒚
𝒅 𝒅
(𝟐𝒚𝒕) = (𝒚𝟐 )
𝒅𝒕 𝒅𝒚
𝒅 𝒅
𝟐𝒚 (𝒕) = (𝒚𝟐 )
𝒅𝒕 𝒅𝒚
𝟐𝒚(𝟏) = 𝟐𝒚
𝟐𝒚 = 𝟐𝒚
Interpretation: Since Exactness Test is met on the differential equation 𝟐𝒚𝒕 𝒅𝒚 + 𝒚𝟐 𝒅𝒕 = 𝟎, it is
Exact Differential Equation.
Solution (4-Step Procedure): Exact Differential Equation requires a 4-step procedure for its
solution:
Step – I: Writing the Preliminary function 𝑭(𝒚, 𝒕) in terms of derivative and integral w.r.t 𝒚, with a
merger of possible constant of integration and variable 𝒕:
𝑖𝑛𝑡𝑒𝑔𝑟𝑎𝑙
⏞ 𝜹{𝑭(𝒚, 𝒕)} 𝑖𝑛𝑡𝑒𝑔𝑟𝑎𝑡𝑖𝑜𝑛 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
𝑭(𝒚, 𝒕) = ∫ 𝒅𝒚 + ⏞
𝒄 + 𝒇(𝒕)
⏟
⏟ 𝜹𝒚 𝑡𝑒𝑟𝑚(𝑠)
𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑖𝑛𝑔 "𝒕"
Combining 𝒄 and 𝒇(𝒕) into one term 𝝍(𝒕) and substituting:
⏞ 𝜹{𝑭(𝒚, 𝒕)} 𝑖𝑛𝑡𝑒𝑔𝑟𝑎𝑡𝑖𝑜𝑛 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
𝑭(𝒚, 𝒕) = ∫ 𝒅𝒚 + ⏞
𝝍(𝒕)
⏟
⏟ 𝜹𝒚 𝑡𝑒𝑟𝑚(𝑠) 𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑖𝑛𝑔 "𝒕"
𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒
𝝍(𝒕) can have two components:
1) Additive constant(s) or terms containing 𝒕 only reduce to zero (𝟎) when function 𝑭 (𝒚, 𝒕 )
gets differentiated w.r.t. 𝒚.
𝜹{𝑭(𝒚,𝒕)}
2) Constant of integration generated due to ∫ 𝜹𝒚
𝒅𝒚.
𝜹{𝑭(𝒚,𝒕)}
Recalling that 𝜹𝒚
= 𝑴, therefore:
𝑭(𝒚, 𝒕) = ∫ 𝑴 𝒅𝒚 + 𝝍(𝒕)
Substituting value of 𝑴:
𝑭(𝒚, 𝒕) = ∫ 𝟐𝒚𝒕 𝒅𝒚 + 𝝍(𝒕)
𝑭(𝒚, 𝒕) = 𝟐𝒕 ∫ 𝒚 𝒅𝒚 + 𝝍(𝒕)
𝒚𝟐
𝑭(𝒚, 𝒕) = 𝟐𝒕 ( ) + 𝒄 + 𝝍(𝒕)
𝟐
______________________________________________________________________________
To avoid excessive terms, merge 𝒄 in 𝝍(𝒕), as 𝝍(𝒕) already contains an integration constant
from Step – I:
𝑭(𝒚, 𝒕) = 𝒚𝟐 𝒕 + 𝝍(𝒕)
𝑭(𝒚, 𝒕) = 𝒚𝟐 𝒕 + 𝝍(𝒕) A Preliminary Result
Step – II: Differentiate 𝑭(𝒚, 𝒕) w.r.t 𝒕:
𝜹{𝑭(𝒚, 𝒕)} 𝜹
= {𝒚𝟐 𝒕 + 𝝍(𝒕)}
𝜹𝒕 𝜹𝒕
𝜹{𝑭(𝒚, 𝒕)} 𝜹 𝜹
= 𝒚𝟐 { (𝒕)} + {𝝍(𝒕)}
𝜹𝒕 𝜹𝒕 𝜹𝒕
𝜹{𝑭(𝒚, 𝒕)}
= 𝒚𝟐 {𝟏} + 𝝍′ (𝒕)
𝜹𝒕
𝜹𝑭
= 𝒚𝟐 + 𝝍′(𝒕)
𝜹𝒕
𝜹𝑭
Since value of 𝝍′ (𝒕) is not known, so we try to substitute ( 𝜹𝒕 ) to find it.
𝜹𝑭 𝜹𝑭
We know that 𝜹𝒕
= 𝑵, therefore, we put it in 𝜹𝒕
= 𝒚𝟐 + 𝝍′(𝒕).
𝜹𝑭
= 𝒚𝟐 + 𝝍′ (𝒕)
𝜹𝒕
𝑵 = 𝒚𝟐 + 𝝍′(𝒕)
Further 𝑵 = 𝒚𝟐 , so putting this value in the equation above:
𝒚𝟐 = 𝒚𝟐 + 𝝍′ (𝒕)
𝒚𝟐 = 𝒚𝟐 + 𝝍′(𝒕)
𝟎 = 𝝍′(𝒕)
𝝍′ (𝒕) = 𝟎
Step – III: We have found the value of derivative 𝝍′ (𝒕) and not the value of original function 𝝍(𝒕).
Therefore, taking integral to restore the original function 𝝍(𝒕) from 𝝍′ (𝒕):
𝝍′ (𝒕) = 𝟎
⏟
𝑧𝑒𝑟𝑜 (𝟎) 𝑐𝑎𝑛 𝑏𝑒 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒
𝑜𝑓 𝑎 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 (𝒌)
Integrating on both sides w.r.t 𝒕:
∫ 𝝍′ (𝒕) 𝒅𝒕 = ⏟
𝒌
𝑎 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑖𝑠 𝑝𝑟𝑖𝑚𝑖𝑡𝑖𝑣𝑒 𝑜𝑓
𝑎 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒,𝝍′ (𝒕)=𝟎
𝒅
∫ {𝝍(𝒕)} 𝒅𝒕 = 𝒌
𝒅𝒕
𝝍(𝒕) = 𝒌
Step – IV: Substitute the result from Step – III in Step –I.
Result from Step – I: 𝑭(𝒚, 𝒕) = 𝒚𝟐 𝒕 + 𝝍(𝒕)
Result from Step – III: 𝝍(𝒕) = 𝒌
𝑭(𝒚, 𝒕) = 𝒚𝟐 𝒕 + 𝒌
Recalling definition of an Exact Differential Equation:
𝒅{𝑭(𝒚, 𝒕)} = 𝟎, implies 𝑭(𝒚, 𝒕) = 𝒄𝑭 (𝑎 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡)
𝟎 is differential of a constant (𝒄𝑭 ).
𝒄𝑭 = 𝒚𝟐 𝒕 + 𝒌
𝒄𝑭 = 𝒚𝟐 𝒕 + 𝒌
______________________________________________________________________________
𝒄𝑭 − 𝒌 = 𝒚𝟐 𝒕
𝒚𝟐 𝒕 = 𝒄𝑭 − 𝒌
Since difference of two constants is yet another constant: 𝒄𝑭 − 𝒌 = 𝒄′
𝒚𝟐 𝒕 = 𝒄′
𝒚𝟐 𝒕 = 𝒄′
This implicit function is synonymous to 𝑭(𝒚, 𝒕) = 𝒌 , and can be considered as the solution.
However, its conventional to find the time path, i.e. 𝒚(𝒕), therefore, we convert it into an explicit
function:
𝒄′
𝒚𝟐 =
𝒕
𝒄′
𝒚(𝒕) = √
𝒕
√𝒄′
𝒚(𝒕) =
√𝒕
𝒄
𝒚(𝒕) = 𝟏 [∵ √𝒄′ = 𝒄]
𝒕𝟐
𝟏
𝒚(𝒕) = 𝒄 𝒕−𝟐
Interpretation: General Solution of Exact Differential Equation 𝟐𝒚𝒕 𝒅𝒚 + 𝒚𝟐 𝒅𝒕 = 𝟎 is
𝟏
𝒚(𝒕) = 𝒄 𝒕−𝟐 .
TOPIC 111: EXACT DIFFERENTIAL EQUATIONS: INTEGRATING FACTOR
What if Exactness Test fails and we have an Inexact Differential Equation:

𝜹 𝜹
(𝑴) ≠ (𝑵)
𝜹𝒕 𝜹𝒚
Since Exactness Test depends of given values of 𝑴 and 𝑵:
𝜹 𝜹
(𝑴) ⋛ (𝑵)
𝜹𝒕 𝜹𝒚
As per definitional requirement, 𝑳. 𝑯. 𝑺 of should be exact total differential of the given function
𝑭(𝒚, 𝒕) = 𝒌. However, in Inexact Differential Equation, this condition is violated:
(𝑴)𝒅𝒚
⏟ + (𝑵)𝒅𝒕 = ⏟ 𝟎
𝑵𝒐𝒕 𝒆𝒒𝒖𝒂𝒍 𝑡𝑜 𝑡𝑜𝑡𝑎𝑙 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑙 𝑜𝑓
𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑙 𝑜𝑓 𝑡ℎ𝑒 𝑎 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 (𝒌)
𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 {𝑭(𝒚,𝒕)}
We can try to change the values of 𝑴 and 𝑵 to get the equality in Exactness Test. This can be
done simply by multiplying a term (𝑭𝑰𝒏𝒕 ) throughout the equation.
(𝑭𝑰𝒏𝒕 ) × {(𝑴)𝒅𝒚 + (𝑵)𝒅𝒕} = (𝑭𝑰𝒏𝒕 ) × 𝟎
This will have a dual effect:
1. Change values of 𝑴 and 𝑵.
2. Equality, in differential equation, remains; i.e. 𝑳. 𝑯. 𝑺 = 𝑹. 𝑯. 𝑺
{(𝑭𝑰𝒏𝒕 ) × (𝑴)} 𝒅𝒚 + ⏟
⏟ {(𝑭𝑰𝒏𝒕 ) × (𝑵)} 𝒅𝒕 = ⏟ 𝟎
𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝑴 𝑐ℎ𝑎𝑛𝑔𝑒𝑠 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝑵 𝑐ℎ𝑎𝑛𝑔𝑒𝑠 𝑒𝑞𝑢𝑎𝑙𝑖𝑡𝑦
𝑏𝑦 𝑎 𝑓𝑎𝑐𝑡𝑜𝑟 (𝑭𝒊𝒏𝒕 ) 𝑏𝑦 𝑎 𝑓𝑎𝑐𝑡𝑜𝑟 (𝑭𝒊𝒏𝒕 ) 𝑟𝑒𝑚𝑎𝑖𝑛𝑠
______________________________________________________________________________
This term is called Integrating Factor (𝑭𝑰𝒏𝒕 ) as it’s a ‘multiplicative factor’ that converts 𝑳. 𝑯. 𝑺 of
𝜹𝑭 𝜹𝑭
the differential equation 𝑭(𝒚, 𝒕) into Exact Differential 𝒅{𝑭(𝒚, 𝒕)} = (𝜹𝒚) 𝒅𝒚 + ( 𝜹𝒕 ) 𝒅𝒕. After which,
the differential equation becomes integrable – hence, Integrating Factor.
Integrating Factor, if any, can be found by using two methods:
I – Integrating Factor by Trial & Error Method.
II – Integrating Factor by Formula.
I – Integrating Factor by Trial & Error Method:

Integrating Factor can be found by experimenting different values of Integrating Factor in the
Inexact Differential Equation and selecting one that makes Exactness Test Valid on post-
multiplication differential equation – Trial and Error method:
Numerical Example: Given the differential equation; 𝟐𝒕 𝒅𝒚 + 𝒚 𝒅𝒕 = 𝟎.
We employ Exactness Test:
𝜹 𝜹
(𝑴) = (𝑵)
𝜹𝒕 𝜹𝒚
Comparing the given differential equation: 𝟐𝒕 𝒅𝒚 + 𝒚 𝒅𝒕 = 𝟎
with standard form of differential equation: (𝑴)𝒅𝒚 + (𝑵)𝒅𝒕 = 𝟎.
𝟐𝒕
⏟ 𝒅𝒚 + ⏟ 𝒚 𝒅𝒕 = 𝟎
𝑴 𝑵
Here, 𝑴 = 𝟐𝒕 and 𝑵 = 𝒚, so substituting these values:
𝜹 𝜹
(𝟐𝒕) ⋛ (𝒚)
𝜹𝒕 𝜹𝒚
𝟐>𝟏
Exactness test doesn’t get satisfied, so we have an Inexact Differential Equation.
To convert it into Exact Differential Equation, we employ Trial and Error method: multiply
throughout by ′𝒚′ :
𝒚(𝟐𝒕 𝒅𝒚 + 𝒚 𝒅𝒕) = 𝒚 × 𝟎
𝟐𝒕𝒚 𝒅𝒚 + 𝒚𝟐 𝒅𝒕 = 𝟎
To check if it is converted into an Exact Differential Equation, employ Exactness Test again:
𝟐𝒕𝒚
⏟ 𝒅𝒚 + 𝒚 ⏟𝟐 𝒅𝒕 = 𝟎
𝑴 𝑵
After introducing the potential integrating factor, 𝑴 = 𝟐𝒕𝒚 and 𝑵 = 𝒚𝟐 , so substituting these values
again:
𝜹 𝜹
(𝟐𝒕𝒚) ⋛ (𝒚𝟐 )
𝜹𝒕 𝜹𝒚
𝟐𝒚 = 𝟐𝒚
′
Interpretation: After introducing ′𝒚 as a factor, Exactness Test is met. So, we can call it
integrating factor (𝑭𝒊𝒏𝒕 = 𝒚) of the Inexact Differential Equation 𝟐𝒕 𝒅𝒚 + 𝒚 𝒅𝒕 = 𝟎. The Exact
Differential Equation form of it is 𝟐𝒕𝒚 𝒅𝒚 + 𝒚𝟐 𝒅𝒕 = 𝟎.
Note: If ′𝒚′ doesn’t work, then ′𝒕′ can be experimented as a potential integrating factor. Higher
powers of ′𝒚′ and ′𝒕′ (𝒚𝟐 , 𝒕𝟐 , 𝒚𝟑 , 𝒕𝟑 , 𝒚𝟒 , 𝒕𝟒 , … ) can be used if both ′𝒚′ and ′𝒕′ fail to serve as
Integrating Factor.
II – Integrating Factor by Formula:

In addition to Trial and Error method, (𝑭𝒊𝒏𝒕 ), if any, can also be found by using these rules of
formula:
𝟏 𝜹𝑴 𝜹𝑵
Rule 1. If 𝑵 ( 𝜹𝒕 − 𝜹𝒚 ) = 𝒇(𝒚) alone, then 𝒆∫ 𝒇(𝒚)𝒅𝒚 is the integrating factor (𝑭𝒊𝒏𝒕 ).
______________________________________________________________________________
𝟏 𝜹𝑵 𝜹𝑴
Rule 2. If ( − ) = 𝒈(𝒕) alone, then 𝒆∫ 𝒈(𝒕)𝒅𝒕 is the integrating factor (𝑭𝒊𝒏𝒕 ).
𝑴 𝜹𝒚 𝜹𝒕
Numerical Example: Given 𝟓𝒚𝒕 𝒅𝒚 + (𝟓𝒚𝟐 + 𝟖𝒕) 𝒅𝒕 = 𝟎, We employ Exactness Test:
𝜹 𝜹
(𝑴) = (𝑵)
𝜹𝒕 𝜹𝒚
Comparing the given differential equation: 𝟓𝒚𝒕 𝒅𝒚 + (𝟓𝒚𝟐 + 𝟖𝒕) 𝒅𝒕 = 𝟎
with standard form of differential equation: (𝑴)𝒅𝒚 + (𝑵)𝒅𝒕 = 𝟎.
𝟐
𝟓𝒚𝒕
⏟ 𝒅𝒚 + (𝟓𝒚⏟ + 𝟖𝒕) 𝒅𝒕 = 𝟎
𝑴 𝑵
𝟐
Here, 𝑴 = 𝟓𝒚𝒕 and 𝑵 = (𝟓𝒚 + 𝟖𝒕), so substituting these values:
𝜹 𝜹
(𝟓𝒚𝒕) ⋛ (𝟓𝒚𝟐 + 𝟖𝒕)
𝜹𝒕 𝜹𝒚
𝟓𝒚 ≠ 𝟏𝟎𝒚

Applying Rule 1:
If ( − ) = 𝒇(𝒚) alone, then 𝒆∫ 𝒇(𝒚)𝒅𝒚 is the integrating factor (𝑭𝒊𝒏𝒕 ):
𝑵 𝜹𝒕 𝜹𝒚
𝟏 𝜹(𝟓𝒚𝒕) 𝜹(𝟓𝒚𝟐 + 𝟖𝒕)
= × { − }
(𝟓𝒚𝟐 + 𝟖𝒕) 𝜹𝒕 𝜹𝒚
𝟏
= 𝟐
× {𝟓𝒚 − 𝟏𝟎𝒚}
(𝟓𝒚 + 𝟖𝒕)
𝟏
= 𝟐
× (−𝟓𝒚)
(𝟓𝒚 + 𝟖𝒕)
−𝟓𝒚
= 𝟓𝒚𝟐 +𝟖𝒕
≠ 𝒇(𝒚) alone.
Since Rule 1 could not be satisfied, we resort to Rule 2:
If ( − ) = 𝒈(𝒕) alone, then 𝒆∫ 𝒈(𝒕)𝒅𝒕 is the integrating factor (𝑭𝒊𝒏𝒕 ).
= ( − )
𝟏 𝜹(𝟓𝒚𝟐 + 𝟖𝒕) 𝜹(𝟓𝒚𝒕)
= { − }
𝟓𝒚𝒕 𝜹𝒚 𝜹𝒕
𝟏
= {𝟏𝟎𝒚 − 𝟓𝒚}
𝟓𝒚𝒕
𝟏
= (𝟓𝒚)
𝟓𝒚𝒕
𝟏
= 𝒕
= 𝒈(𝒕) alone.
Since Rule 2 is met, the integrating factor (𝑭𝒊𝒏𝒕 ) is 𝒆∫ 𝒈(𝒕)𝒅𝒕 :
𝑭𝒊𝒏𝒕 = 𝒆∫ 𝒈(𝒕)𝒅𝒕
𝟏
∫( )𝒅𝒕
𝑭𝒊𝒏𝒕 = 𝒆 𝒕
𝑭𝒊𝒏𝒕 = 𝒆(𝒍𝒏|𝒕|+𝒄)
Constant of integration is dropped 𝒄 for the sake of convenience:
𝑭𝒊𝒏𝒕 = 𝒆(𝒍𝒏|𝒕|)
𝑭𝒊𝒏𝒕 = 𝒆𝒍𝒏|𝒕|
______________________________________________________________________________
𝑭𝒊𝒏𝒕 = 𝒕
Interpretation: Integrating factor (𝑭𝒊𝒏𝒕 ) for the Inexact Differential Equation 𝟓𝒚𝒕 𝒅𝒚 +
(𝟓𝒚𝟐 + 𝟖𝒕) 𝒅𝒕 = 𝟎 is ′𝒕′ . Multiplying it throughout the given equation, it becomes exact differential
equation: 𝟓𝒚𝒕𝟐 𝒅𝒚 + (𝟓𝒚𝟐 𝒕 + 𝟖𝒕𝟐 ) 𝒅𝒕 = 𝟎. Test of Exact can be applied to verify the answer.
Caveat: Once the Inexact Differential Equation is converted into Exact Differential Equation, it
can be solved using the 4-step procedure already explained before.
______________________________________________________________________________
Lesson 26
ECONOMIC APPLICATIONS OF EXACT DIFFERENTIAL EQUATIONS
TOPIC 112: PRODUCTION FUNCTION USING EXACT DIFFERENTIAL EQUATIONS
A firm uses 𝑳 number of workers and 𝑲 number of machines as two inputs in its production
process. The marginal products of the two inputs are given as:
𝟐 𝟏
𝑳 𝟑 𝑲 𝟑
𝑴𝑷𝑲 = 𝟐( ) and 𝑴𝑷𝑳 = 𝟒( )
𝑲 𝑳
Express the total product function of this firm assuming it stays on the same isoquant. What kind
of production function obtains?
Solution: A Production process is mathematized via a production function:
𝑸(𝑲, 𝑳) = 𝒌
Its Total Differential:
𝒅{𝑸(𝑲, 𝑳)} = 𝟎
𝜹𝑸 𝜹𝑸
( ) 𝒅𝑲 + ( ) 𝒅𝑳 = 𝟎
𝜹𝑲 𝜹𝑳
( 𝑴𝑷𝑲 )𝒅𝑲 + ( 𝑴𝑷𝑳 )𝒅𝑳 = 𝟎
𝟐 𝟏
𝑳 𝟑 𝑲 𝟑
{ 𝟐 ( ) } 𝒅𝑲 + { 𝟒 ( ) } 𝒅𝑳 = 𝟎
𝑲 𝑳

(𝑴) 𝒅𝒚 + (𝑵) 𝒅𝒕 = 𝟎
𝟐 𝟏
𝑳 𝟑 𝑲 𝟑
{ 𝟐 ( ) } 𝒅𝑲
⏟ + { 𝟒 ( ) } 𝒅𝑳
⏟ =𝟎
𝑲 𝒅𝒚
𝑳 𝒅𝒕
⏟ ⏟
𝑴 𝑵
𝟐 𝟏
𝑳 𝟑 𝑲 𝟑
It implies: 𝑴 = 𝟐 ( ) , 𝑵 = 𝟒 ( ) , 𝒅𝒚 = 𝒅𝑲 and 𝒅𝒕 = 𝒅𝑳.
𝑲 𝑳
𝜹 𝜹
Exactness Test: 𝜹𝒕 (𝑴) = 𝜹𝒚 (𝑵)
𝟐 𝟏
𝜹 𝑳 𝟑 𝜹 𝑲 𝟑
{𝟐 ( ) } = {𝟒 ( ) }
𝜹𝑳 𝑲 𝜹𝑲 𝑳
𝟐 𝟏
𝟏 𝟑 𝜹 𝟐 𝟏 𝟑 𝜹 𝟏
𝟐( ) {(𝑳)𝟑 } = 𝟒 ( ) {(𝑲)𝟑 }
𝑲 𝜹𝑳 𝑳 𝜹𝑲
𝟐 𝟏
𝟏 𝟑 𝟐 𝟐 𝜹 𝟏 𝟑 𝟏 𝟏 𝜹
𝟐 ( ) { × (𝑳)𝟑−𝟏 × (𝑳)} = 𝟒 ( ) { × (𝑲)𝟑−𝟏 × (𝑲)}
𝑲 𝟑 𝜹𝑳 𝑳 𝟑 𝜹𝑲
𝟐 𝟏
𝟐 𝟏 𝟑 𝟏 𝟏 𝟏 𝟑 𝟐
𝟐 ( ) ( ) {(𝑳)−𝟑 × 𝟏} = 𝟒 ( ) ( ) {(𝑲)−𝟑 × 𝟏}
𝟑 𝑲 𝟑 𝑳
𝟒 𝟏 𝟏 𝟒 𝟏 𝟏
( 𝟐) ( 𝟏) = ( 𝟏) ( 𝟐)
𝟑 𝟑 𝟑
𝑲𝟑 𝑳𝟑 𝑳 𝑲𝟑
______________________________________________________________________________
𝟒 𝟏 𝟒 𝟏
( 𝟐 𝟏) = ( 𝟐 𝟏)
𝟑 𝟑
𝑲𝟑 𝑳𝟑 𝑲𝟑 𝑳𝟑
Exactness test is met, so the given equation is an Exact Differential Equation.
solution:
Step – I: Writing the function 𝑸(𝑲, 𝑳) in terms of derivative and integral w.r.t 𝑲, with a merger of
possible constant of integration and variable 𝑳:
⏞ 𝜹{𝑸(𝑲, 𝑳)} 𝑖𝑛𝑡𝑒𝑔𝑟𝑎𝑡𝑖𝑜𝑛 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
𝑸(𝑲, 𝑳) = ∫ 𝒅𝑲 + ⏞𝒄 + 𝒇(𝑳)
⏟
⏟ 𝜹𝑲
𝑡𝑒𝑟𝑚(𝑠)
𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑖𝑛𝑔 "𝑳"
Combining 𝒄 and 𝒇(𝑳) into one term 𝝍(𝑳) and substituting:
⏞ 𝑖𝑛𝑡𝑒𝑔𝑟𝑎𝑡𝑖𝑜𝑛 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
𝜹{𝑸(𝑲, 𝑳)} ⏞
𝑸(𝑲, 𝑳) = ∫ 𝒅𝑲 + 𝝍(𝑳)
⏟
⏟ 𝜹𝑲
𝑡𝑒𝑟𝑚(𝑠) 𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑖𝑛𝑔 "𝑳"
𝝍(𝑳) can have two components:
1) Additive constant(s) or terms containing 𝑳 only reduce to zero (𝟎) when function
𝑸 (𝑲, 𝑳 ) gets differentiated w.r.t. 𝑲.
𝜹{𝑸(𝑲,𝑳)}
2) Constant of integration generated due to ∫ 𝜹𝑲
𝒅𝑲.
𝜹{𝑸(𝑲,𝑳)}
Recalling that = 𝑴, therefore:
𝜹𝑲
𝑸(𝑲, 𝑳) = ∫ 𝑴 𝒅𝒚 + 𝝍(𝑳)
𝟐
𝑳 𝟑
𝑸(𝑲, 𝑳) = ∫ {𝟐 ( ) } 𝒅𝑲 + 𝝍(𝑳)
𝑲
𝟐
𝟐 𝟏 𝟑
𝑸(𝑲, 𝑳) = 𝟐(𝑳)𝟑 ∫ ( ) 𝒅𝑲 + 𝝍(𝑳)
𝑲
𝟐 𝟐
𝑸(𝑲, 𝑳) = 𝟐(𝑳)𝟑 ∫(𝑲)−𝟑 𝒅𝑲 + 𝝍(𝑳)
Solving the integral while merging the constant of integration into 𝝍(𝑳).
𝟐
𝟐 (𝑲)−𝟑+𝟏
𝑸(𝑲, 𝑳) = 𝟐(𝑳)𝟑 { }+ 𝝍(𝑳)
𝟐
−𝟑 + 𝟏
𝟏
𝟐 (𝑲)𝟑
𝑸(𝑲, 𝑳) = 𝟐(𝑳)𝟑 { }+ 𝝍(𝑳)
𝟏
𝟑
𝟐 𝟑 𝟏
𝑸(𝑲, 𝑳) = 𝟐 (𝑳)𝟑 × × (𝑲)𝟑 + 𝝍(𝑳)
𝟏
𝟐 𝟏
𝑸(𝑲, 𝑳) = 𝟔 (𝑳)𝟑 (𝑲)𝟑 + 𝝍(𝑳)
______________________________________________________________________________
𝟐 𝟏
𝑸(𝑲, 𝑳) = 𝟔(𝑳)𝟑 (𝑲)𝟑 + 𝝍(𝑳)
Step – II: Differentiate 𝑸(𝑲, 𝑳) w.r.t 𝑳 {to evaluate 𝝍(𝑳)}:
𝜹{𝑸(𝑲, 𝑳)} 𝜹 𝟐 𝟏
= {𝟔(𝑳)𝟑 (𝑲)𝟑 + 𝝍(𝑳)}
𝜹𝑳 𝜹𝑳
𝜹{𝑸(𝑲, 𝑳)} 𝟏 𝜹 𝟐 𝜹
= (𝑲)𝟑 × {𝟔(𝑳)𝟑 } + {𝝍(𝑳)}
𝜹𝑳 𝜹𝑳 𝜹𝑳
𝜹{𝑸} 𝟏 𝟐 𝟐 𝜹
= 𝟔 × (𝑲)𝟑 { × (𝑳)𝟑−𝟏 × (𝑳)} + 𝝍′ (𝑳)
𝜹𝑳 𝟑 𝜹𝑳
𝜹𝑸 𝟐 𝟏 𝟏 𝜹
= 𝟔 × × (𝑲)𝟑 {(𝑳)−𝟑 × (𝑳)} + 𝝍′ (𝑳)
𝜹𝑳 𝟑 𝜹𝑳
𝜹𝑸
Since we already know; 𝜹𝑳
= 𝑵 , therefore 𝑳. 𝑯. 𝑺:
𝟏 𝟏
𝑵 = 𝟒 × (𝑲)𝟑 { 𝟏
× 𝟏} + 𝝍′ (𝑳)
𝑳𝟑
𝟏
𝑲 𝟑
Moreover, 𝑵 = 𝟒 (𝑳) , so we substitute:
𝟏 𝟏
𝑲 𝟑 (𝑲)𝟑
𝟒 ( ) = 𝟒 𝟏 + 𝝍′ (𝑳)
𝑳
𝑳𝟑
𝟏 𝟏
𝑲 𝟑 𝑲 𝟑
𝟒 ( ) = 𝟒 ( ) + 𝝍′ (𝑳)
𝑳 𝑳
𝟎 = 𝝍′ (𝑳)
𝝍′ (𝑳) = 𝟎
Step – III: We have found the value of derivative 𝝍′ (𝑳) and not the value of original function 𝝍(𝑳).
Therefore, taking integral to restore the original function 𝝍(𝑳) from 𝝍′ (𝑳):
𝝍′ (𝑳) = 𝟎
⏟
Integrating on both sides w.r.t 𝑳:
∫ 𝝍′ (𝑳) 𝒅𝑳 = ⏟
𝒌
𝑎 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒,𝝍′ (𝑳)=𝟎
𝒅
{𝝍(𝑳)} 𝒅𝑳 = 𝒌
∫
𝒅𝑳
𝝍(𝑳) = 𝒌
𝟐 𝟏
Result from Step – I: 𝑸(𝑲, 𝑳) = 𝟔(𝑳)𝟑 (𝑲)𝟑 + 𝝍(𝑳)
Result from Step – III: 𝝍(𝑳) = 𝒌
𝟐 𝟏
𝑸(𝑲, 𝑳) = 𝟔(𝑳)𝟑 (𝑲)𝟑 + 𝒌
𝒅{𝑸(𝑲, 𝑳)} = 𝟎, implies 𝑸(𝑲, 𝑳) = 𝒄𝑸 (𝑎 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡)
𝟎 is differential of a constant (𝒄𝑸 ).
𝟐 𝟏
𝒄𝑸 = 𝟔(𝑳)𝟑 (𝑲)𝟑 + 𝒌
______________________________________________________________________________
𝟐 𝟏
𝒄𝑸 = 𝟔(𝑳)𝟑 (𝑲)𝟑 + 𝒌
𝟐 𝟏
𝒄𝑸 − 𝒌 = 𝟔(𝑳)𝟑 (𝑲)𝟑
𝟐 𝟏
𝟔(𝑳)𝟑 (𝑲)𝟑 = 𝒄𝑸 − 𝒌
Since difference of two constants is yet another constant: 𝒄𝑸 − 𝒌 = 𝒄
𝟐 𝟏
𝟔(𝑳)𝟑 (𝑲)𝟑 = 𝒄
𝟏 𝟐
𝟔(𝑲)𝟑 (𝑳)𝟑 = 𝒄
Developing Functional expression on 𝑹. 𝑯. 𝑺:
𝟏 𝟐
𝑸(𝑲, 𝑳) = 𝟔(𝑲)𝟑 (𝑳)𝟑 − 𝒄
This expression is synonymous to Cobb-Douglas production function, with the exception of
addition term ′𝒄′ .
In absence of either or both of inputs (𝑲, 𝑳), the output will be zero:
𝟏 𝟐 𝟏 𝟐 𝟏 𝟐
𝑸( 𝟎 , 𝑳) = 𝟔( 𝟎 )𝟑 (𝑳)𝟑 − 𝒄 𝑸(𝑲, 𝟎 ) = 𝟔(𝑲)𝟑 ( 𝟎 )𝟑 − 𝒄 𝑸( 𝟎 , 𝟎 ) = 𝟔( 𝟎 )𝟑 ( 𝟎 )𝟑 − 𝒄
𝟏 𝟐 𝟏 𝟐 𝟏 𝟐
𝑸(𝟎, 𝑳) = 𝟔(𝟎)𝟑 (𝑳)𝟑 − 𝒄 𝑸(𝑲, 𝟎) = 𝟔(𝑲)𝟑 (𝟎)𝟑 − 𝒄 𝑸(𝟎, 𝟎) = 𝟔(𝟎)𝟑 (𝟎)𝟑 − 𝒄
𝑸(𝟎, 𝑳) = −𝒄 ≠ 𝟎 𝑸(𝟎, 𝑳) = −𝒄 ≠ 𝟎 𝑸(𝟎, 𝟎) = −𝒄 ≠ 𝟎
Which is a paradox. 𝑸 should be equal to zero if either or both 𝑲 or 𝑳 are zero.
Therefore, 𝒄 should be equated to 𝟎, for making ‘economic’ sense of Production Function.
𝟏 𝟐
𝑸(𝑲, 𝑳) = 𝟔(𝟎)𝟑 (𝑳)𝟑 − 𝒄
𝟏 𝟐
𝑸(𝑲, 𝑳) = 𝟔(𝟎)𝟑 (𝑳)𝟑 − 𝟎
𝟏 𝟐
𝑸(𝑲, 𝑳) = 𝟔(𝟎)𝟑 (𝑳)𝟑
𝟐 𝟏
𝑳 𝟑 𝑲 𝟑
Interpretation: General Solution of Exact Differential Equation 𝟐 (𝑲) 𝒅𝑲 + 𝟒 ( 𝑳 ) 𝒅𝑳 = 𝟎 is
𝟏 𝟐
𝑸(𝑲, 𝑳) = 𝟔(𝟎)𝟑 (𝑳)𝟑 . It makes economic sense of a Cobb-Douglas Production Function.
TOPIC 113: UTILITY FUNCTION USING EXACT DIFFERENTIAL EQUATIONS
A simple utility function of a consumer buying two goods 𝒙 and 𝒚 is denoted by the differential
equation 𝟒𝒙𝒚 𝒅𝒙 + 𝒙𝟐 𝒅𝒚 = 𝟎. Express this consumer’s total utility function using integrating
factors.
Solution: Given Utility function, in general form is:
𝑼(𝒙, 𝒚) = 𝒌
( 𝟒𝒙𝒚 )𝒅𝒙 + ( 𝒙𝟐 ) 𝒅𝒚 = 𝟎
(𝑴) 𝒅𝒚 + (𝑵) 𝒅𝒕 = 𝟎
⏟ + (⏟𝒙𝟐 ) 𝒅𝒚
(⏟𝟒𝒙𝒚 ) 𝒅𝒙 ⏟ =𝟎
𝑴 𝒅𝒚 𝒅𝒕
𝑵
𝟐
It implies: 𝑴 = 𝟒𝒙𝒚, 𝑵 = 𝒙 , 𝒅𝒚 = 𝒅𝒙 and 𝒅𝒕 = 𝒅𝒚.
𝜹 𝜹
Exactness Test: 𝜹𝒚 (𝑴) = 𝜹𝒙 (𝑵)
𝜹 𝜹 𝟐
(𝟒𝒙𝒚) = (𝒙 )
𝜹𝒚 𝜹𝒙
______________________________________________________________________________
𝜹 𝜹 𝟐
𝟒𝒙 { (𝒚)} = (𝒙 )
𝜹𝒚 𝜹𝒙
𝟒𝒙{𝟏} = 𝟐𝒙
𝟒𝒙 ≠ 𝟐𝒙
Applying Rule 1:
If 𝑵 ( 𝜹𝒕 − 𝜹𝒚 ) = 𝒇(𝒙) alone, then 𝒆∫ 𝒇(𝒙)𝒅𝒙 is the integrating factor (𝑭𝒊𝒏𝒕 ):
𝟏 𝜹(𝟒𝒙𝒚) 𝜹(𝒙𝟐 )
= × { − }
(𝒙𝟐 ) 𝜹𝒚 𝜹𝒙
𝟏
= 𝟐 × {𝟒𝒙 − 𝟐𝒙}
(𝒙 )
𝟏
= 𝟐 × 𝟐𝒙
(𝒙 )
𝟏 𝜹𝑴 𝜹𝑵 𝟐
( − ) = = 𝒇(𝒙) alone
𝑵 𝜹𝒚 𝜹𝒙 𝒙
Since Rule 1 is met, the integrating factor (𝑭𝒊𝒏𝒕 ) is 𝒆∫ 𝒇(𝒙)𝒅𝒙:

𝑭𝒊𝒏𝒕 = 𝒆∫ 𝒇(𝒙)𝒅𝒙
𝟐
∫( )𝒅𝒙
𝑭𝒊𝒏𝒕 = 𝒆 𝒙
𝑭𝒊𝒏𝒕 = 𝒆 𝟐 (𝒍𝒏|𝒙|)+𝒄
Constant of integration is dropped 𝒄 for the sake of convenience:
(𝒍𝒏|𝒙 𝟐 |)
𝑭𝒊𝒏𝒕 = 𝒆
𝟐
𝑭𝒊𝒏𝒕 = 𝒆(𝒍𝒏|𝒙 |)
𝑭𝒊𝒏𝒕 = 𝒙𝟐
Interpretation: Integrating factor (𝑭𝒊𝒏𝒕 ) for the Inexact Differential Equation 𝟒𝒙𝒚 𝒅𝒙 + 𝒙𝟐 𝒅𝒚 = 𝟎
is (𝒙𝟐 ). Multiplying it throughout the given equation, it becomes exact differential equation:
𝟒𝒙 𝟑 𝒚 𝒅𝒙 + 𝒙 𝟒 𝒅𝒚 = 𝟎.
After introducing the integrating factor (𝒙𝟐 ), 𝑴 = 𝟒𝒙𝟑 𝒚 and 𝑵 = 𝒙𝟒 , so substituting these values
𝜹 𝜹
in Exactness Test again: 𝜹𝒚 (𝑴) = 𝜹𝒙 (𝑵)
𝜹 𝜹 𝟒
(𝟒𝒙𝟑 𝒚) = (𝒙 )
𝜹𝒚 𝜹𝒙
𝜹 𝜹 𝟒
𝟒𝒙𝟑 { (𝒚)} = (𝒙 )
𝜹𝒚 𝜹𝒙
𝟒𝒙𝟑 {𝟏} = 𝟒𝒙𝟑
𝟒𝒙𝟑 = 𝟒𝒙𝟑
Exactness test is met, so the given equation is an Exact Differential Equation which can be solved
using 4-Step Procedure.
solution:
Step – I: Writing the function 𝑼(𝒙, 𝒚) in terms of derivative and integral w.r.t 𝒙, with a merger of
possible constant of integration and variable 𝒚:
______________________________________________________________________________
⏞ 𝜹{𝑼(𝒙, 𝒚)} 𝑖𝑛𝑡𝑒𝑔𝑟𝑎𝑡𝑖𝑜𝑛 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
𝑼(𝒙, 𝒚) = ∫ 𝒅𝒙 + ⏞𝒄 + 𝒇(𝒚)
⏟
⏟ 𝜹𝒙
𝑡𝑒𝑟𝑚(𝑠)
𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑖𝑛𝑔 "𝒚"
Combining 𝒄 and 𝒇(𝒚) into one term 𝝍(𝒚) and substituting:
𝑖𝑛𝑡𝑒𝑔𝑟𝑎𝑡𝑖𝑜𝑛 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
⏞ 𝜹{𝑼(𝒙, 𝒚)}
𝑼(𝒙, 𝒚) = ∫ 𝒅𝒙 + ⏞
𝝍(𝒚)
⏟
⏟ 𝜹𝒙
𝑡𝑒𝑟𝑚(𝑠) 𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑖𝑛𝑔 "𝒚"
𝝍(𝒚) can have two components:
1) Additive constant(s) or terms containing 𝒚 only reduce to zero (𝟎) when function
𝑼(𝒙, 𝒚 ) gets differentiated w.r.t. 𝒙.
𝜹{𝑼(𝒙,𝒚)}
2) Constant of integration generated due to ∫ 𝜹𝒙
𝒅𝒙.
𝜹{𝑼(𝒙,𝒚)}
𝜹𝒙
𝑼(𝒙, 𝒚) = ∫ 𝑴 𝒅𝒙 + 𝝍(𝒚)
𝑼(𝒙, 𝒚) = ∫(𝟒𝒙𝟑 𝒚) 𝒅𝒙 + 𝝍(𝒚)
𝑼(𝒙, 𝒚) = 𝟒𝒚 ∫(𝒙𝟑 ) 𝒅𝒙 + 𝝍(𝒚)

Solving the integral while merging the constant of integration into 𝝍(𝒚).
(𝒙)𝟑+𝟏
𝑼(𝒙, 𝒚) = 𝟒𝒚 { } + 𝝍(𝒚)
𝟑+𝟏
𝒙𝟒
𝑼(𝒙, 𝒚) = 𝟒𝒚 ( ) + 𝝍(𝒚)
𝟒
𝑼(𝒙, 𝒚) = 𝒙𝟒 𝒚 + 𝝍(𝒚)
Step – II: Differentiate 𝑼(𝒙, 𝒚) w.r.t 𝒚 {to evaluate 𝝍(𝒚)}:
𝜹{𝑼(𝒙, 𝒚)} 𝜹 𝟒
= {𝒙 𝒚 + 𝝍(𝒚)}
𝜹𝒚 𝜹𝒚
𝜹{𝑼(𝒙, 𝒚)} 𝜹 𝜹
= 𝒙𝟒 (𝒚) + {𝝍(𝒚)}
𝜹𝒚 𝜹𝒚 𝜹𝒚
𝜹{𝑼}
= 𝒙𝟒 (𝟏) + 𝝍′ (𝒚)
𝜹𝒚
𝜹𝑼
= 𝒙𝟒 + 𝝍′ (𝒚)
𝜹𝒚
𝜹𝑼
Since we already know; 𝜹𝒚
= 𝑵 , therefore 𝑳. 𝑯. 𝑺:
𝑵 = 𝒙𝟒 + 𝝍′ (𝒚)
Moreover, 𝑵 = 𝒙𝟒 , so we substitute:
𝒙𝟒 = 𝒙𝟒 + 𝝍′ (𝒚)
𝟎 = 𝝍′ (𝒚)
𝝍′ (𝒚) = 𝟎
______________________________________________________________________________
Step – III: We have found the value of derivative 𝝍′ (𝒚) and not the value of original function 𝝍(𝒚).
Therefore, taking integral to restore the original function 𝝍(𝒚) from 𝝍′ (𝒚):
𝝍′ (𝒚) = 𝟎
⏟
Integrating on both sides w.r.t 𝒚:
∫ 𝝍′ (𝒚) 𝒅𝒚 = ⏟
𝒌
𝑎 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒,𝝍′ (𝒚)=𝟎
𝒅
∫ {𝝍(𝒚)} 𝒅𝒚 = 𝒌
𝒅𝒚
𝝍(𝒚) = 𝒌
Result from Step – I: 𝑼(𝒙, 𝒚) = 𝒙𝟒 𝒚 + 𝝍(𝒚)
Result from Step – III: 𝝍(𝒚) = 𝒌
𝑼(𝒙, 𝒚) = 𝒙𝟒 𝒚 + 𝒌
𝒅{𝑼(𝒙, 𝒚)} = 𝟎, implies 𝑼(𝒙, 𝒚) = 𝒄𝑼 (𝑎 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡)
𝟎 is differential of a constant (𝒄𝑼 ).
𝒄𝑼 = 𝒙𝟒 𝒚 + 𝒌
𝒄𝑼 = 𝒙𝟒 𝒚 + 𝒌
𝒄𝑼 − 𝒌 = 𝒙𝟒 𝒚
𝒙𝟒 𝒚 = 𝒄𝑼 − 𝒌
Since difference of two constants is yet another constant: 𝒄𝑼 − 𝒌 = 𝒄
𝒙𝟒 𝒚 = 𝒄
𝒙𝟒 𝒚 = 𝒄
Developing Functional expression on 𝑹. 𝑯. 𝑺:
𝑼(𝒙, 𝒚) = 𝒙𝟒 𝒚 − 𝒄
Interpretation: General Solution of Exact Differential Equation 𝟒𝒙𝟑 𝒚 𝒅𝒙 + 𝒙𝟒 𝒅𝒚 = 𝟎 is
𝑼(𝒙, 𝒚) = 𝒙𝟒 𝒚 − 𝒄 . It makes economic sense of a Utility Function.
TOPIC 114: POINT ELASTICITY FORMULA USING EXACT DIFFERENTIAL EQUATIONS
The point elasticity of demand a firm is faced with is 𝑬𝒅 = −𝟐. Use the exact differential equation
approach to find the demand function of this firm.
Solution: Since no differential equation is given, we develop one from given data:
𝑬𝒅 = −𝟐
𝒅𝑸𝒅 𝑸𝒅
÷ = −𝟐
𝒅𝑷 𝑷
𝒅𝑸𝒅 𝑷
× = −𝟐
𝒅𝑷 𝑸𝒅
𝒅𝑸𝒅 𝑸𝒅
= −𝟐
𝒅𝑷 𝑷
Cross Multiplying, we get:
______________________________________________________________________________
𝒅𝑸𝒅 𝒅𝑷
= −𝟐
𝑸𝒅 𝑷
𝒅𝑸𝒅 𝒅𝑷
+𝟐
=𝟎
𝑸𝒅 𝑷
𝟏 𝟐
( ) 𝒅𝑸𝒅 + ( ) 𝒅𝑷 = 𝟎
𝑸𝒅 𝑷
(𝑴) 𝒅𝒚 + (𝑵) 𝒅𝒕 = 𝟎
𝟏 𝟐
( ) 𝒅𝑸
⏟𝒅 + ( ) 𝒅𝑷 ⏟ =𝟎
⏟𝑸𝒅 ⏟𝑷 𝒅𝒕
𝒅𝒚
𝑴 𝑵
𝟏 𝟐
It implies: 𝑴 = 𝑸𝒅
, 𝑵= 𝑷
, 𝒅𝒚 = 𝒅𝑸𝒅 and 𝒅𝒕 = 𝒅𝑷.
𝜹 𝜹
Exactness Test: 𝜹𝑷
(𝑴) = 𝜹𝑸 (𝑵)
𝒅
𝜹 𝟏 𝜹 𝟏
( ) = ( )
𝜹𝑷 𝑸𝒅 𝜹𝑸𝒅 𝑷
𝟎=𝟎
Exactness test gets satisfied, so we have an Exact Differential Equation.
After verification of Exactness test, given Exact Differential Equation can be solved using 4-Step
Procedure.
solution:
Step – I: Writing the function 𝑭(𝑸𝒅 , 𝑷) in terms of derivative and integral w.r.t 𝑸𝒅 , with a merger
of possible constant of integration and variable 𝑷:
⏞ 𝑖𝑛𝑡𝑒𝑔𝑟𝑎𝑡𝑖𝑜𝑛 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
𝜹{𝑭(𝑸𝒅 , 𝑷)}
𝑭(𝑸𝒅 , 𝑷) = ∫ 𝒅𝑸𝒅 + ⏞𝒄 + 𝒇(𝑷)
⏟
⏟ 𝜹𝑸𝒅 𝑡𝑒𝑟𝑚(𝑠)
𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑖𝑛𝑔 "𝑷"
Combining 𝒄 and 𝒇(𝑷) into one term 𝝍(𝑷) and substituting:
⏞ 𝜹{𝑭(𝑸 , 𝑷)} 𝑖𝑛𝑡𝑒𝑔𝑟𝑎𝑡𝑖𝑜𝑛 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
𝒅 ⏞
𝑭(𝑸𝒅 , 𝑷) = ∫ 𝒅𝑸𝒅 + 𝝍(𝑷)
⏟
⏟ 𝜹𝑸𝒅 𝑡𝑒𝑟𝑚(𝑠) 𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑖𝑛𝑔 "𝑷"
𝝍(𝑷) can have two components:
1) Additive constant(s) or terms containing 𝑷 only reduce to zero (𝟎) when function
𝑭 (𝑸𝒅 , 𝑷 ) gets differentiated w.r.t. 𝑸𝒅 .
𝜹{𝑭(𝑸𝒅 ,𝑷)}
2) Constant of integration generated due to ∫ 𝜹𝑸𝒅
𝒅𝑸𝒅 .
𝜹{𝑭(𝑸𝒅 ,𝑷)}
𝜹𝑸𝒅
𝑭(𝑸𝒅 , 𝑷) = ∫ 𝑴 𝒅𝑸𝒅 + 𝝍(𝑷)

𝟏
𝑭(𝑸𝒅 , 𝑷) = ∫ ( ) 𝒅𝑸𝒅 + 𝝍(𝑷)
𝑸𝒅
______________________________________________________________________________
Solving the integral while merging the constant of integration into 𝝍(𝑷).
𝑭(𝑸𝒅 , 𝑷) = 𝒍𝒏|𝑸𝒅 | + 𝝍(𝑷)
Step – II: Differentiate 𝑭(𝑸𝒅 , 𝑷) w.r.t 𝑷 {to evaluate 𝝍(𝑷)}:
𝜹{𝑭(𝑸𝒅 , 𝑷)} 𝜹
= {𝒍𝒏|𝑸𝒅 | + 𝝍(𝑷)}
𝜹𝑷 𝜹𝑷
𝜹{𝑭(𝑸𝒅 , 𝑷)} 𝜹 𝜹
= (𝒍𝒏|𝑸𝒅 |) + {𝝍(𝑷)}
𝜹𝑷 𝜹𝑷 𝜹𝒚
𝜹{𝑼}
= 𝟎 + 𝝍′ (𝑷)
𝜹𝒚
𝜹𝑼
= 𝝍′ (𝑷)
𝜹𝒚
𝜹𝑼
Since we already know; = 𝑵 , therefore 𝑳. 𝑯. 𝑺:
𝜹𝒚
𝑵 = 𝝍′ (𝑷)
𝟐
Moreover, 𝑵 = 𝑷 , so we substitute:
𝟐
= 𝝍′ (𝑷)
𝑷
𝟐
𝝍′ (𝑷) =
𝑷
′ (𝑷)
Step – III: We have found the value of derivative 𝝍 and not the value of original function 𝝍(𝑷).
Therefore, taking integral to restore the original function 𝝍(𝑷) from 𝝍′ (𝑷):
Integrating on both sides w.r.t 𝑷:
𝟐
∫ 𝝍′ (𝑷) 𝒅𝑷 = ∫ ( ) 𝒅𝑷
𝑷
𝒅 𝟏
∫ {𝝍(𝑷)} 𝒅𝑷 = 𝟐 × ∫ ( ) 𝒅𝑷
𝒅𝑷 𝑷
𝝍(𝑷) = 𝟐 × 𝒍𝒏|𝑷| + 𝒄𝟏
Result from Step – I: 𝑭(𝑸𝒅 , 𝑷) = 𝒍𝒏|𝑸𝒅 | + 𝝍(𝑷)
Result from Step – III: 𝝍(𝑷) = 𝝍(𝑷) = 𝟐 × 𝒍𝒏|𝑷| + 𝒄𝟏
𝑭(𝑸𝒅 , 𝑷) = 𝒍𝒏|𝑸𝒅 | + 𝟐 × 𝒍𝒏|𝑷| + 𝒄𝟏
𝒅{𝑭(𝑸𝒅 , 𝑷)} = 𝟎, implies 𝑭(𝑸𝒅 , 𝑷) = 𝒄𝑫 (𝑎 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡)
𝟎 is differential of a constant (𝒄𝑫 ).
𝒄𝑫 = 𝒍𝒏|𝑸𝒅 | + 𝟐 × 𝒍𝒏|𝑷| + 𝒄𝟏
𝒄𝑫 − 𝒄𝟏 = 𝒍𝒏|𝑸𝒅 | + 𝒍𝒏|𝑷 𝟐 |
𝒄𝟐 = 𝒍𝒏|𝑸𝒅 | + 𝒍𝒏|𝑷𝟐 |
𝒍𝒏|𝑸𝒅 × 𝑷𝟐 | = 𝒄𝟐
Raising to exponent on both sides, to get rid of natural logarithm.
𝟐
𝒆𝒍𝒏|𝑸𝒅 ×𝑷 | = 𝒆𝒄𝟐
𝑸𝒅 × 𝑷𝟐 = 𝒄
𝒄
𝑸𝒅 = 𝟐
𝑷
______________________________________________________________________________
𝟏 𝟐 𝒄
Interpretation: The Differential Equation (𝑸 ) 𝒅𝑸𝒅 + (𝑷) 𝒅𝑷 = 𝟎 has General solution 𝑸𝒅 = 𝑷𝟐. In
𝒅
𝒄
economic terms, the solution 𝑸𝒅 = 𝑷𝟐 represents a demand function; 𝑸𝒅 = 𝒇(𝑷).
TOPIC 115: NON-LINEAR DIFFERENTIAL EQUATION: SEPARABLE VARIABLES
The Concept: If 𝒇(𝒚, 𝒕) 𝒅𝒚 + 𝒈(𝒚, 𝒕) 𝒅𝒕 = 𝟎 has the convenient property that function 𝒇 and 𝒈 can
be written in terms of 𝒚 and 𝒕 only. i.e. 𝒇(𝒚) 𝒅𝒚 + 𝒈(𝒕) 𝒅𝒕 = 𝟎
Variables 𝒚 and 𝒕 are said to be separable, as 𝒇(𝒚, 𝒕) is consolidated into 𝒇(𝒚) and 𝒈(𝒚, 𝒕) is
consolidated into 𝒈(𝒕).
𝒚 is mathematically separated from 𝒕.
Solve using (simple) integration.
Example # 1: (p. 492): Solve the following differential equation:
𝟑𝒚𝟐 𝒅𝒚 − 𝒕 𝒅𝒕 = 𝟎
Rearranging we get: 𝟑𝒚𝟐 𝒅𝒚 = 𝒕 𝒅𝒕
Introducing integral on both sides:
∫ 𝟑𝒚𝟐 𝒅𝒚 = ∫ 𝒕 𝒅𝒕
𝒕𝟐
𝒚𝟑 + 𝒄𝟏 = + 𝒄𝟐
𝟐
𝒕𝟐
𝒚𝟑 = + 𝒄𝟐 − 𝒄𝟏
𝟐
𝒕𝟐
𝒚𝟑 = 𝟐 + 𝒄′ , [∵ 𝒄𝟐 − 𝒄𝟏 = 𝒄′ ].
𝟏
𝒕𝟐 𝟑
𝒚 = ( + 𝒄′ )
𝟐
In functional form of a time-path:
𝟏
𝒕𝟐 𝟑
𝒚(𝒕) = ( + 𝒄)
𝟐
Example # 2: (p. 493): 𝟐𝒕 𝒅𝒚 + 𝒚 𝒅𝒕 = 𝟎 is a differential equation.
Rearranging it, we get:
𝟐𝒕 𝒅𝒚 = −𝒚 𝒅𝒕,
Needs more rearrangement, therefore cross-multiplying;
𝒅𝒚 𝒅𝒕
=−
𝒚 𝟐𝒕
𝒅𝒚 𝟏 𝒅𝒕
=−
𝒚 𝟐 𝒕
Introducing integrals on both sides:
𝒅𝒚 𝟏 𝒅𝒕
∫ = − ∫
𝒚 𝟐 𝒕
𝟏
𝒍𝒏|𝒚| + 𝒄𝟏 = − 𝒍𝒏|𝒕| + 𝒄𝟐
𝟐
𝟏
𝒍𝒏|𝒚| + 𝒍𝒏|𝒕| = 𝒄𝟐 − 𝒄𝟏
𝟐
𝟏
𝒍𝒏|𝒚| + 𝒍𝒏 |𝒕𝟐 | = 𝒄′
______________________________________________________________________________
𝟏
𝒍𝒏 |𝒚 × 𝒕𝟐 | = 𝒄′
Raising to exponent on both sides, to get rid of natural logarithm:
𝟏
𝒍𝒏|𝒚×𝒕𝟐 | ′
𝒆 = 𝒆𝒄
′
𝒆𝒄 = 𝒄 : exponent of a constant 𝒄′ shall be yet another constant 𝒄.
𝟏
𝒚 × 𝒕𝟐 = 𝒄
𝟏
𝒚 = 𝒄𝒕−𝟐
In functional form of a time-path:
𝟏
𝒚(𝒕) = 𝒄𝒕−𝟐
TOPIC 116: DYNAMICS OF CAPITAL USING SEPARABLE VARIABLES METHOD
Let 𝑿 = 𝑿(𝒕) denote the national product, 𝑲 = 𝑲(𝒕) the capital stock, and 𝑳 = 𝑳(𝒕) the number of
workers in a country at time 𝒕. Suppose that, for all 𝒕 ≥ 𝟎 (𝒂) 𝑿 = √𝑲√𝑳, (𝒃) 𝑲̇ = 𝟎. 𝟒𝑿, (𝒄) 𝑳 =
𝒆𝟎.𝟎𝟒𝒕. Derive from these equations a single differential equation for 𝑲 = 𝑲(𝒕) and find the solution
of that equation when 𝑲(𝟎) = 𝟏𝟎𝟎𝟎𝟎.
𝑿 = √𝑲√𝑳; is a Cobb-Douglas production function.
𝑲̇ = 𝟎. 𝟒𝑿; shows that investment is proportional to output
𝑳 = 𝒆𝟎.𝟎𝟒𝒕 ; depicts the exponential growth of labor.
Solution: Starting with the equation of rate of change of capital; 𝑲̇ = 𝟎. 𝟒𝑿
𝒅𝑲
= 𝟎. 𝟒𝑿
𝒅𝒕
Borrowing value of 𝑿 from equation (𝒂): 𝑲̇ = 𝟎. 𝟒𝑿
𝒅𝑲
= 𝟎. 𝟒 × √𝑲√𝑳
𝒅𝒕
𝒅𝑲
= 𝟎. 𝟒 × √𝑲√ 𝑳
𝒅𝒕
Borrowing value of 𝑳 from equation (𝒄): 𝑳 = 𝒆𝟎.𝟎𝟒𝒕
𝒅𝑲
= 𝟎. 𝟒 × √𝑲√ 𝒆𝟎.𝟎𝟒𝒕
𝒅𝒕
𝟏 𝒅𝑲
× = 𝟎. 𝟒 × √𝒆𝟎.𝟎𝟒𝒕
√𝑲 𝒅𝒕
𝟏 𝒅𝑲
𝑲−𝟐 × = 𝟎. 𝟒 × √(𝒆𝟎.𝟎𝟐𝒕 )𝟐
𝒅𝒕
𝟏 𝒅𝑲
𝑲−𝟐 × = 𝟎. 𝟒 × 𝒆𝟎.𝟎𝟐𝒕
𝒅𝒕
Cross-multiplying for separable variable situation:
𝟏
(𝑲−𝟐 ) × 𝒅𝑲 = ⏟
𝟎. 𝟒 (𝒆𝟎.𝟎𝟐𝒕 )𝒅𝒕
⏟
𝒊𝒏 𝒕𝒆𝒓𝒎𝒔 𝒐𝒇
𝑖𝑛 𝑡𝑒𝑟𝑚𝑠 𝑜𝑓
(𝑲) 𝑜𝑛𝑙𝑦 (𝒕) 𝒐𝒏𝒍𝒚
𝟏
∫ (𝑲−𝟐 ) 𝒅𝑲 = ∫ 𝟎. 𝟒 (𝒆𝟎.𝟎𝟐𝒕 ) 𝒅𝒕
______________________________________________________________________________
𝟏
𝑲−𝟐+𝟏 𝒆𝟎.𝟎𝟐𝒕
( ) + 𝒄𝑲 = 𝟎. 𝟒 ( ) + 𝒄𝒕
𝟏 𝒅
−𝟐 + 𝟏 (𝟎. 𝟎𝟐𝒕)
𝒅𝒕
𝟏
𝑲𝟐 𝒆𝟎.𝟎𝟐𝒕
( ) = 𝟎. 𝟒 ( ) + 𝒄𝒕 − 𝒄𝑲
𝟏 𝟎. 𝟎𝟐
𝟐
𝟏 𝟎. 𝟒
𝟐𝑲𝟐 = (𝒆𝟎.𝟎𝟐𝒕 ) + 𝒄𝟏
𝟎. 𝟎𝟐
𝟏 𝟎. 𝟒
𝑲(𝒕)𝟐 = × 𝟐 (𝒆𝟎.𝟎𝟐𝒕 ) + 𝒄𝟏 × 𝟐
𝟎. 𝟎𝟐
𝟏 𝟎. 𝟒
𝑲(𝒕)𝟐 = ( ) (𝒆𝟎.𝟎𝟐𝒕 ) + 𝒄𝟐
𝟎. 𝟎𝟐 × 𝟐
√𝑲(𝒕) = (𝟏𝟎)(𝒆𝟎.𝟎𝟐𝒕 ) + 𝒄𝟐
Introducing the initial condition; 𝒕 = 𝟎, in the equation above:
√ 𝑲(𝟎) = (𝟏𝟎){𝒆𝟎.𝟎𝟐(𝟎) } + 𝒄𝟐
√ 𝟏𝟎𝟎𝟎𝟎 = (𝟏𝟎)(𝟏) + 𝒄𝟐
𝟏𝟎𝟎 = 𝟏𝟎 + 𝒄𝟐
𝟏𝟎𝟎 − 𝟏𝟎 = 𝒄𝟐
𝒄𝟐 = 𝟗𝟎
Substituting the value of 𝒄𝟐 in √𝑲(𝒕) = (𝟏𝟎)(𝒆𝟎.𝟎𝟐𝒕 ) + 𝒄𝟐 .
√𝑲(𝒕) = (𝟏𝟎)(𝒆𝟎.𝟎𝟐𝒕 ) + 𝟗𝟎
𝟐
𝑲(𝒕) = {𝟏𝟎(𝒆𝟎.𝟎𝟐𝒕 ) + 𝟗𝟎}
𝟐
𝑲(𝒕) = [𝟏𝟎{(𝒆𝟎.𝟎𝟐𝒕 ) + 𝟗}]
𝟐
𝑲(𝒕) = 𝟏𝟎𝟐 (𝒆𝟎.𝟎𝟐𝒕 + 𝟗)
𝟐
𝑲(𝒕) = 𝟏𝟎𝟎 (𝒆𝟎.𝟎𝟐𝒕 + 𝟗)
𝟏
Interpretation: Non-linear differential equation for capital (𝑲−𝟐 ) × 𝒅𝑲 = 𝟎. 𝟒 (𝒆𝟎.𝟎𝟐𝒕 )𝒅𝒕 has
𝟐
solution 𝑲(𝒕) = 𝟏𝟎𝟎 (𝒆𝟎.𝟎𝟐𝒕 + 𝟗) which represents the time-path of capital (𝑲) over time (𝒕).
______________________________________________________________________________
Lesson 27
ECONOMIC APPLICATIONS OF EXACT DIFFERENTIAL EQUATIONS (CONTINUED)
TOPIC 117: NON-LINEAR DIFFERENTIAL EQUATION: UNDERSTANDING BERNOULLI’S

EQUATION
A certain type of nonlinear differential equation, attributed to Jacob Bernoulli who was a Swiss
Mathematician in 𝟏𝟕𝒕𝒉 century (𝟏𝟔𝟓𝟒 − 𝟏𝟕𝟎𝟓).
𝒅𝒚
If the differential equation = 𝒉(𝒚, 𝒕) has the specific nonlinear form:
𝒅𝒕
𝒅𝒚
+ 𝑹𝒚 = 𝑻𝒚𝒎
𝒅𝒕
Salient Feature of this form:
𝒅𝒚 𝒎
+ 𝑹
⏟ 𝒚
⏟ = ⏟
𝑻 × 𝒚⏟
𝒅𝒕 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 (𝒚) 𝑎𝑝𝑝𝑒𝑎𝑟𝑠 𝑜𝑛 𝐿.𝐻.𝑆 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 (𝒚) 𝑎𝑝𝑝𝑒𝑎𝑟𝑠 𝑜𝑛 𝑅.𝐻.𝑆
𝑜𝑓 (𝒕) 𝑤𝑖𝑡ℎ 𝑝𝑜𝑤𝑒𝑟 (𝟏) 𝑜𝑓 (𝒕) 𝑤𝑖𝑡ℎ 𝑝𝑜𝑤𝑒𝑟
𝑜𝑡ℎ𝑒𝑟 𝑡ℎ𝑎𝑛 𝟎 𝑜𝑟 𝟏
𝑹: function of 𝒕.
𝑻: function of 𝒕.
𝒎: (𝒎 = 𝟎) ∨ (𝒎 = 𝟏).
With these conditions fulfilled we get a Bernoulli Differential Equation.
Caveat: what if 𝒎 = 𝟎 or 𝒎 = 𝟏?
𝒎=𝟎 𝒎=𝟏
𝒅𝒚 𝒅𝒚
+ 𝑹𝒚 = 𝑻𝒚 𝟎 + 𝑹𝒚 = 𝑻𝒚 𝟏
𝒅𝒕 𝒅𝒕
𝒅𝒚 𝒅𝒚
+ 𝑹𝒚 = 𝑻 × 𝟏 + 𝑹𝒚 = 𝑻𝒚
𝒅𝒕 𝒅𝒕
𝒅𝒚 𝒅𝒚
+𝑹𝒚 = 𝑻 + (𝑹 − 𝑻) 𝒚 = 𝟎
𝒅𝒕 𝒅𝒕
Differential equation doesn’t remain Differential equation doesn’t remain
nonlinear, rather becomes linear nonlinear, rather becomes linear
differential equation. differential equation.
Solution: Bernoulli Equation (nonlinear differential equation) can be reduced to a linear

differential equation and be solved as such.
𝒅𝒚
Reduction to linear form: Standard form of Bernoulli Equation: 𝒅𝒕 + 𝑹𝒚 = 𝑻 𝒚𝒎
______________________________________________________________________________
Since 𝒚𝒎 is the cause of nonlinearity, we reduce it to 1 by self-dividing:

𝒅𝒚
( + 𝑹𝒚) 𝑻 𝒚𝒎
𝒅𝒕
= 𝒎
𝒚𝒎 𝒚
𝟏 𝒅𝒚 𝒚 𝑻 𝒚𝒎
× +𝑹 𝒎 = 𝒎
𝒚𝒎 𝒅𝒕 𝒚 𝒚
𝒅𝒚
𝒚−𝒎 × + 𝑹(𝒚 × 𝒚−𝒎 ) = 𝑻
𝒅𝒕
𝒅𝒚
(𝒚−𝒎 ) +𝑹 ⏟𝒚𝟏−𝒎 =𝑻
⏟ 𝒅𝒕
𝑺𝒖𝒃𝒔𝒕𝒊𝒕𝒖𝒕𝒊𝒐𝒏−𝑰
𝑺𝒖𝒃𝒔𝒕𝒊𝒕𝒖𝒕𝒊𝒐𝒏−𝑰𝑰
Boxed terms, if substituted will free equation from nonlinear term of 𝒚.
Substitution – I: Substitution – II: 𝒛 depends on 𝒚, which further depends on 𝒕: 𝒛 =

Introducing a shorthand 𝒇(𝒚) and 𝒚 = 𝒈(𝒕); combining the two function: 𝒛 = 𝒇{𝒚(𝒕)}
notation 𝒛 for 𝒚𝟏−𝒎 : We get a composite function, which if differentiated, calls for Chain
𝒛 = 𝒚𝟏−𝒎 𝒅𝒛 𝒅𝒛 𝒅𝒚
Rule: 𝒅𝒕 = 𝒅𝒚 × 𝒅𝒕
Nonlinear term of 𝒚 will 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑡𝑒
be substituted by linear 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑙 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑙
𝑡𝑒𝑟𝑚 𝑜𝑓 𝑛𝑒𝑤 𝒛=𝒚𝟏−𝒎 𝑡𝑒𝑟𝑚 𝑜𝑓 𝑛𝑒𝑤
term of 𝒛. 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛 𝒛=𝒇(𝒕) 𝑤.𝑟.𝑡 𝒚 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛 𝒚=𝒇(𝒕)
⏞
𝒅𝒛 ⏞𝒅𝒛 ⏞𝒅𝒚
= ( ) × ( )
𝒅𝒕 𝒅𝒚 𝒅𝒕
𝟏−𝒎
Differentiate 𝒛 = 𝒚 w.r.t 𝒚:
𝒅𝒛 𝒅 𝒅
= (𝒚𝟏−𝒎 ) = (𝟏 − 𝒎) × (𝒚𝟏−𝒎−𝟏 ) × (𝒚)
𝒅𝒚 𝒅𝒚 𝒅𝒚
𝒅𝒛
= (𝟏 − 𝒎) × (𝒚−𝒎 )
𝒅𝒚
𝒅𝒛 𝒅𝒛 𝒅𝒛 𝒅𝒚
Substituting value of 𝒅𝒚
in chain rule; 𝒅𝒕
= 𝒅𝒚
× 𝒅𝒕
𝒅𝒛 𝒅𝒚
= (𝟏 − 𝒎) × (𝒚−𝒎 ) ×
𝒅𝒕 𝒅𝒕
𝟏 𝒅𝒛 𝒅𝒚
× = (𝒚−𝒎 )
(𝟏 − 𝒎) 𝒅𝒕 𝒅𝒕
______________________________________________________________________________
Combining the Substitution – I & Substitution – II:

𝒅𝒚
Now we can surrogate 𝒛 instead of 𝒚 in (𝒚−𝒎 ) +𝑹 𝒚𝟏−𝒎
⏟ =𝑻
⏟ 𝒅𝒕
𝑺𝒖𝒃𝒔𝒕𝒊𝒕𝒖𝒕𝒊𝒐𝒏−𝑰𝑰 𝑺𝒖𝒃𝒔𝒕𝒊𝒕𝒖𝒕𝒊𝒐𝒏−𝑰
𝟏 𝒅𝒛
× +𝑹 𝒛
⏟ =𝑻
(𝟏 − 𝒎) 𝒅𝒕
⏟ 𝑺𝒖𝒃𝒔𝒕𝒊𝒕𝒖𝒕𝒊𝒐𝒏−𝑰
𝟏 𝒅𝒛
× + 𝑹𝒛 = 𝑻
(𝟏 − 𝒎) 𝒅𝒕
𝒅𝒛 + 𝑹𝒛(𝟏 − 𝒎)𝒅𝒕
=𝑻
(𝟏 − 𝒎)𝒅𝒕
𝒅𝒛 + 𝑹𝒛(𝟏 − 𝒎)𝒅𝒕 = 𝑻 × (𝟏 − 𝒎)𝒅𝒕
𝒅𝒛 + 𝑹𝒛(𝟏 − 𝒎)𝒅𝒕 − 𝑻(𝟏 − 𝒎)𝒅𝒕 = 𝟎
𝒅𝒛 + {(𝟏 − 𝒎) × (𝑹𝒛 − 𝑻)}𝒅𝒕 = 𝟎
Objective of replacing a nonlinear term 𝒚𝒎 with linear term 𝒛(𝟏) .
𝒛 𝑖𝑛𝑠𝑡𝑒𝑎𝑑
𝑜𝑓 𝒚
𝒅𝒛 + {(𝟏 − 𝒎) × (𝑹 ⏟ ⏞
𝒛 − 𝑻)} 𝒅𝒕 = 𝟎
𝑙𝑖𝑛𝑒𝑎𝑟
𝑡𝑒𝑟𝑚
TOPIC 118: NON-LINEAR DIFFERENTIAL EQUATION: BERNOULLI’S EQUATION

NUMERICALLY
𝒅𝒚
Solve the equation + 𝒕𝒚 = 𝟑𝒕𝒚𝟐.
𝒅𝒕
Solution: It is Bernoulli equation as we compare it with standard from:
𝒅𝒚
+ 𝑹𝒚 = 𝑻𝒚 𝒎
𝒅𝒕
𝒅𝒚
+⏟ ⏟ 𝒚⏟𝟐
𝒕 𝒚 = 𝟑𝒕
𝒅𝒕 𝑹 𝑻 𝑚=2
Here; 𝑹 = 𝒕, 𝑻 = 𝟑𝒕, 𝒎 = 𝟐.
Substitution to replace 𝒛 instead 𝒚𝟏−𝒎 :
𝒛 = 𝒚𝟏−𝒎
𝒛 = 𝒚𝟏−𝟐
𝒛 = 𝒚−𝟏
Recalling the linearized form of Bernoulli Equation, in terms of 𝒛 instead of 𝒚:
𝒅𝒛 + {(𝟏 − 𝒎) × (𝑹𝒛 − 𝑻)}𝒅𝒕 = 𝟎
Substituting the given values of 𝒎, 𝑹, and 𝑻.
𝒅𝒛 + {(𝟏 − 𝟐) × (𝒕𝒛 − 𝟑𝒕)}𝒅𝒕 = 𝟎
𝒅𝒛 + {(−𝟏) × 𝒕(𝒛 − 𝟑)}𝒅𝒕 = 𝟎
______________________________________________________________________________
𝒅𝒛 + (−𝒛𝒕 + 𝟑𝒕)𝒅𝒕 = 𝟎
𝒅𝒚
Transforming into standard form of first order differential equation + 𝒖(𝒕) × 𝒚 = 𝒘(𝒕):
𝒅𝒕
{𝒅𝒛 + (−𝒛𝒕 + 𝟑𝒕)𝒅𝒕} 𝟎
=
𝒅𝒕 𝒅𝒕
𝒅𝒛
+ (−𝒕)𝒛 + 𝟑𝒕 = 𝟎
𝒅𝒕 𝒚
𝒅⏞
𝒛
+ (−𝒕)
⏟ 𝒛 = −𝟑𝒕
⏟
𝒅𝒕 𝒖(𝒕) 𝒘(𝒕)
Using the General Solution of FODE-variable coefficient & variable term:
𝒚 (𝒕) = 𝒆− ∫ 𝒖 𝒅𝒕 (𝑨 + ∫ 𝒘 𝒆 ∫ 𝒖 𝒅𝒕 𝒅𝒕)
Solving the ∫ 𝒖 𝒅𝒕 separately to avoid the clutter:
∫ 𝒖 𝒅𝒕 = ∫(−𝒕) 𝒅𝒕
𝑑𝑟𝑜𝑝𝑝𝑖𝑛𝑔 𝒄
𝑓𝑜𝑟 𝑒𝑎𝑠𝑒
⏞ 𝒕𝟐
∫ 𝒖 𝒅𝒕 = −
𝟐
Returning the value of ∫ 𝒖 𝒅𝒕 in general solution equation:
𝒕𝟐 𝒕𝟐
−(− ) −
𝟐
𝒛(𝒕) = 𝒆 (𝑨 + ∫ −𝟑𝒕 𝒆 𝟐 𝒅𝒕)
𝒕𝟐 𝒕𝟐
(𝟐) (− 𝟐 )
𝒛(𝒕) = 𝒆 {𝑨 + 𝟑 ∫(−𝒕) × 𝒆 𝒅𝒕}
𝒕𝟐
Since here 𝒇(𝒙) = − , 𝒇′ (𝒙) = −𝒕:
𝟐
𝒆𝒇(𝒙)
𝒕𝟐 𝒇′ (𝒙)
(𝟐)
⏞ 𝒕𝟐
⏞ (− )
𝒛(𝒕) = 𝒆 𝑨 + 𝟑 ∫ −𝒕 × 𝒆 𝟐 𝒅𝒕
⏟
∫{𝒇′ (𝒙)×𝒆𝒇(𝒙) }𝒅𝒙
{ }
𝒕𝟐 𝒕𝟐
(𝟐) (− 𝟐 )
𝒛(𝒕) = 𝒆 {𝑨 + 𝟑 𝒆⏟ }
𝒆𝒇(𝒙)
𝒕𝟐 𝒕𝟐 𝒕𝟐
(𝟐) ( 𝟐 ) (− 𝟐 )
𝒛(𝒕) = 𝑨𝒆 +𝟑𝒆 𝒆
𝒕𝟐
(𝟐)
𝒛(𝒕) = 𝑨𝒆 +𝟑
Time-path of 𝒛 in terms of time, however, given information was in terms of 𝒚. So, we restore
the 𝒚 to find its time-path 𝒚(𝒕).
𝒛 = 𝒚−𝟏
After substitution:
𝒕𝟐
(𝟐)
𝒚−𝟏 = 𝑨𝒆 +𝟑
______________________________________________________________________________
𝒕 𝟐
𝟏 ( )
= 𝑨𝒆 𝟐 + 𝟑
𝒚
𝟏
𝒕𝟐
=𝒚
( )
𝑨𝒆 𝟐 +𝟑
𝟏
𝒚(𝒕) = 𝒕𝟐
( )
+𝟑 𝑨𝒆 𝟐
Interpretation: Using Bernoulli equation method, we substituted 𝒛 in place of 𝒚 to linearize the
nonlinear differential equation. Then used procedure of solving FODE to find time-path (however,
of 𝒛). Finally, we restored 𝒚 from 𝒛 and found time-path 𝒚(𝒕).
TOPIC 119: AVERAGE COST FUNCTION USING EXACT DIFFERENTIAL EQUATIONS
The unit cost of a production firm is 𝑨𝑪, while the number of units produced within the firm is 𝑸
such that the following relationship exists:
𝒅(𝑨𝑪) 𝑨𝑪
+ = (𝑨𝑪)𝟑
𝒅𝑸 𝑸
𝒅(𝑨𝑪) 𝟏
+ ( ) 𝑨𝑪 = (𝑨𝑪)𝟑
𝒅𝑸 𝑸
Solution: It is Bernoulli equation as we compare it with standard from:

𝒅𝒚
+ 𝑹𝒚 = 𝑻𝒚 𝒎
𝒅𝒕
𝒅(𝑨𝑪) 𝟏
+ ( ) 𝑨𝑪
⏟ =⏟ 𝟏⏟(𝑨𝑪) 𝟑
⏟𝒅𝑸 ⏟𝑸 𝒚 𝑻 𝑦𝑚
𝒅𝒚 𝑹
𝒅𝒕
𝒅𝒚 𝒅(𝑨𝑪) 𝟏
Here; 𝒚 = 𝑨𝑪, 𝒕 = 𝑸, 𝒅𝒕
= 𝒅𝑸
, 𝑹 = 𝑸, 𝑻 = 𝟏, 𝒎 = 𝟑.
Substitution to replace 𝒛 instead (𝑨𝑪)𝟏−𝒎 :
𝒛 = (𝑨𝑪)𝟏−𝟑
𝒛 = (𝑨𝑪)𝟏−𝟑
𝒛 = (𝑨𝑪)−𝟐
Recalling the linearized form of Bernoulli Equation, in terms of 𝒛 instead of (𝑨𝑪):
𝒅𝒛 + {(𝟏 − 𝒎) × (𝑹𝒛 − 𝑻)}𝒅𝑸 = 𝟎
Substituting the given values of 𝒎, 𝑹, and 𝑻.
𝟏
𝒅𝒛 + [(𝟏 − 𝟑) × {( ) 𝒛 − 𝟏}] 𝒅𝑸 = 𝟎
𝑸
𝟏
𝒅𝒛 + [−𝟐 {( ) 𝒛𝒅𝑸 + 𝒅𝑸}] = 𝟎
𝑸
𝟏
𝒅𝒛 − 𝟐 ( ) 𝒛𝒅𝑸 − 𝟐𝒅𝑸 = 𝟎
𝑸
𝒅𝒛
Dividing by 𝒅𝑸 throughout, to get the standard term of 𝒅𝑸:
𝟏
𝒅𝒛 𝟐 (𝑸) 𝒛 𝒅𝑸 𝟐𝒅𝑸 𝟎
− − =
𝒅𝑸 𝒅𝑸 𝒅𝑸 𝒅𝑸
______________________________________________________________________________
𝒅𝒛 𝟏
+ (−𝟐) ( ) 𝒛 − 𝟐 = 𝟎
𝒅𝑸 𝑸
𝒅𝒚
Transforming into standard form of first order differential equation 𝒅𝒕 + 𝒖(𝒕) × 𝒚 = 𝒘(𝒕):
𝒅𝒛 𝟏
+ {(−𝟐) ( )} ⏟
𝒛= ⏟
𝟐
𝒅𝑸
⏟ ⏟ 𝑸 𝑦 𝑤(𝑡)=𝑤(𝑄)
𝑑𝑦 𝑑𝑦 𝑢(𝑡)=𝑢(𝑄)
=
𝑑𝑡 𝑑𝑄
𝒅𝒚 𝒅𝒛 𝟏
Here; = , 𝒚 = 𝒛, 𝒕 = 𝑸, 𝒖(𝒕) = 𝒖(𝑸) = (−𝟐) ( ), 𝒘(𝒕) = 𝒘(𝑸) = 𝟐.
𝒅𝒕 𝒅𝑸 𝑸
Using the General Solution of FODE-variable coefficient & variable term:
𝒚 (𝒕) = 𝒆− ∫ 𝒖 𝒅𝒕 (𝑨 + ∫ 𝒘 𝒆 ∫ 𝒖 𝒅𝒕 𝒅𝒕)
Here the symbols are economic in nature, 𝒚 = 𝑨𝑪 & 𝒕 = 𝑸.
𝑨𝑪 (𝑸) = 𝒆− ∫ 𝒖(𝑸) 𝒅𝑸 (𝑨 + ∫ 𝒘(𝑸) 𝒆 ∫ 𝒖(𝑸) 𝒅𝑸 𝒅𝑸)
Formula General Solution is also expressed in economic variable:
𝒅𝒛 𝟏
− {(−𝟐) ( )} ⏟
𝒛= ⏟
𝟐
𝒅𝑸
⏟ ⏟ 𝑸 𝑦 𝑤(𝑄)
𝑑𝑦 𝑢(𝑄)
𝑑𝑄
Solving the ∫ 𝒖(𝑸) 𝒅𝑸 separately to avoid the clutter:
𝟏
∫ 𝒖(𝑸) 𝒅𝑸 = ∫(−𝟐) ( ) 𝒅𝑸
𝑸
𝟏
∫ 𝒖(𝑸) 𝒅𝑸 = (−𝟐) ∫ ( ) 𝒅𝑸
𝑸
𝑑𝑟𝑜𝑝𝑝𝑖𝑛𝑔 𝒄
𝑓𝑜𝑟 𝑒𝑎𝑠𝑒
∫ 𝒖(𝑸) 𝒅𝑸 = (−𝟐) × ⏞
𝒍𝒏|𝑸|
∫ 𝒖(𝑸) 𝒅𝑸 = (−𝟐) × 𝒍𝒏|𝑸|

Returning the value of ∫ 𝒖(𝑸) 𝒅𝑸 in general solution equation:
𝑨𝑪 (𝑸) = 𝒆− ∫ 𝒖(𝑸) 𝒅𝑸 (𝑨 + ∫ 𝒘(𝑸) 𝒆 ∫ 𝒖(𝑸) 𝒅𝑸 𝒅𝑸)
𝒛 (𝑸) = 𝒆− (−𝟐)×𝒍𝒏|𝑸| (𝑨 + ∫ 𝟐 𝒆 (−𝟐)×𝒍𝒏|𝑸| 𝒅𝑸)
𝒛(𝑸) = 𝒆(𝟐×𝒍𝒏|𝑸|) (𝑨 + ∫ 𝟐 × 𝒆(−𝟐)×𝒍𝒏|𝑸| 𝒅𝑸)

𝟐 −𝟐 |
𝒛(𝑸) = 𝒆𝒍𝒏|𝑸 | (𝑨 + 𝟐 × ∫ 𝒆𝒍𝒏|𝑸 𝒅𝑸)
𝒛(𝑸) = 𝑸𝟐 (𝑨 + 𝟐 × ∫ 𝑸−𝟐 𝒅𝑸)

𝒐𝒎𝒊𝒕𝒕𝒊𝒏𝒈
𝒊𝒏𝒕𝒆𝒈𝒓𝒂𝒕𝒊𝒐𝒏
𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕
⏞𝑸−𝟏
𝒛(𝑸) = 𝑸𝟐 × 𝑨 + 𝑸𝟐 × 𝟐 ×
(−𝟏)
𝟐
𝒛(𝑸) = 𝑸𝟐 × 𝑨 + × 𝑸𝟐 × 𝑸−𝟏
(−𝟏)
______________________________________________________________________________
𝒛(𝑸) = 𝑨𝑸𝟐 − 𝟐 × 𝑸𝟏
𝒛(𝑸) = 𝑨𝑸𝟐 − 𝟐𝑸
It is the solution of 𝒛 in terms of 𝑸, however, we started the analysis in terms of (𝑨𝑪); returning
value of 𝒛 in terms of (𝑨𝑪): 𝒛(𝑸) = 𝒛 = (𝑨𝑪)−𝟐
(𝑨𝑪)−𝟐 = 𝑨𝑸𝟐 − 𝟐𝑸
(𝑨𝑪)−𝟐 = 𝑨𝑸𝟐 − 𝟐𝑸
𝟏
−
(𝑨𝑪) = (𝑨𝑸𝟐 − 𝟐𝑸) 𝟐
𝟏
𝑨𝑪(𝑸) = 𝟏
(𝑨𝑸𝟐 − 𝟐𝑸)𝟐
𝟏
Interpretation: 𝑨𝑪(𝑸) = 𝟏 is the General solution of the Bernoulli type differential
(𝑨𝑸𝟐 −𝟐𝑸)𝟐
𝒅(𝑨𝑪) 𝟏
equation 𝒅𝑸
+ (𝑸) 𝑨𝑪 = (𝑨𝑪)𝟑 .
𝟏
𝑨𝑪(𝑸) = 𝟏 depicts a negative relationship between average cost and output, provided
(𝑨𝑸𝟐 −𝟐𝑸)𝟐
𝑨 > 𝟎.
In economic sense, 𝑨𝑪 is declining over the range of 𝑸. This shows that 𝑨𝑪 must be predominantly
in the form of fixed cost (𝑭𝑪). This usually happens in case of Natural monopoly, where the 𝑭𝑪
constitute a large share of overall cost and declining trend of its average (𝑨𝑭𝑪) dominates the
shape of 𝑨𝑪 curve. A long-run situation is plotted here:
______________________________________________________________________________
TOPIC 120: QUALITATIVE ANALYSIS: PHASE DIAGRAM
There is a Qualitative aspect of a time-path – Phase Diagram.

Phase Diagrams convey the dynamic properties of differential equations.
Let 𝒚(𝒕) denote an autonomous differential equation, so that 𝒚′ (𝒕) is just a function of 𝒚 and
independent of 𝒕.
𝒅𝒚
Mathematically its functional form is: 𝒅𝒕
= 𝒇(𝒚)
Similar but not the Same

Differential Equation Phase Diagram Time-Path
(Autonomous) Differential
Equation
𝒅𝒚 𝒅𝒚 𝒚 = 𝒇(𝒕)
= 𝒇(𝒚, 𝒕) = 𝒇(𝒚)
𝒅𝒕 𝒅𝒕
______________________________________________________________________________
𝒅𝒚 𝒅𝒚
= 𝒚−𝟕 ⇒ + (−𝟏)
⏟ 𝒚 = −𝟕
⏟
𝒅𝒕 𝒅𝒕 𝒂 𝑏
Before 𝒙 – intercept After 𝒙 – intercept

Phase line is below 𝒙 – axis Phase line is above 𝒙 – axis
𝒅𝒚 𝒅𝒚
( < 𝟎) ( > 𝟎)
𝒅𝒕 𝒅𝒕
𝒅𝒚 𝒅𝒚
( < 𝟎) ⇒ 𝒚 ↓ ⇒ 𝑚𝑜𝑣𝑒 𝒍𝒆𝒇𝒕𝒘𝒂𝒓𝒅𝒔 ( > 𝟎) ⇒ 𝒚 ↑ ⇒ 𝑚𝑜𝑣𝑒 𝒓𝒊𝒈𝒉𝒕𝒘𝒂𝒓𝒅𝒔
𝒅𝒕 𝒅𝒕
Repulsion from 𝒙 – intercept (Repellor): Divergence/Dynamic Unstability
Unstable Intertemporal Equilibrium
Verification: To check if qualitative (graphical) analysis has correct diagnosis regarding the time-
path, we recall the qualitative method.
𝒅𝒚 𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡 𝑖𝑓 (𝒂>𝟎) 𝒅𝒚
+ 𝒂𝒚 = 𝒃 → + (−𝟕)
⏟ 𝒚=⏟ 𝟒
𝒅𝒕 𝒅𝒕 𝒂<𝟎
⏟ 𝑏
𝒅𝒊𝒗𝒆𝒓𝒈𝒆𝒏𝒕
Hence verified that qualitative method has the correct diagnosis.
𝒅𝒚 𝒚 𝒅𝒚 𝟏
=𝟒− ⇒ + ( )𝒚 = ⏟
𝟒
𝒅𝒕 𝟐 𝒅𝒕 ⏟𝟐 𝑏
𝑎
______________________________________________________________________________
Before 𝒙 – intercept After 𝒙 – intercept

Phase line is above 𝒙 – axis Phase line is below 𝒙 – axis
𝒅𝒚 𝒅𝒚
( > 𝟎) ( < 𝟎)
𝒅𝒕 𝒅𝒕
𝒅𝒚 𝒅𝒚
( > 𝟎) ⇒ 𝒚 ↑ ⇒ 𝑚𝑜𝑣𝑒 𝒓𝒊𝒈𝒉𝒕𝒘𝒂𝒓𝒅𝒔 ( < 𝟎) ⇒ 𝒚 ↓ ⇒ 𝑚𝑜𝑣𝑒 𝒍𝒆𝒇𝒕𝒘𝒂𝒓𝒅𝒔
𝒅𝒕 𝒅𝒕
Attraction towards 𝒙 – intercept (Attractor): Convergence/ Dynamic Stability
Stable Intertemporal Equilibrium OR Steady-State Equilibrium
Verification: To check if qualitative (graphical) analysis has correct diagnosis regarding the time-
path, we recall the qualitative method.
𝒅𝒚 𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡 𝑖𝑓 (𝒂>𝟎) 𝒅𝒚 𝟏
+ 𝒂𝒚 = 𝒃 → + ( ) 𝒚=⏟ 𝟒
𝒅𝒕 𝒅𝒕 ⏟𝟐 𝑏
𝒂>𝟎
⏟
𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒕
Hence verified that qualitative method has the correct diagnosis.
TOPIC 121: SOLOW GROWTH MODEL: QUALITATIVE ANALYSIS
Attributed to Robert M. Solow – a Nobel Laureate.

Model was built on Domar’s growth model questioning the need for ‘delicate
balancing’.
Given the qualitative result 𝒌̇ = 𝒔 𝝓(𝒌) − 𝝀𝒌 we can qualitatively analyze the Solow growth model.
𝒌̇
⏟ = ⏟𝒔 𝝓(𝒌)
⏟ − 𝝀𝒌
⏟
𝑟𝑎𝑡𝑒 𝑜𝑓 𝑴𝑷𝑺 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛 𝐝𝐞𝐩𝐫𝐞𝐜𝐢𝐚𝐭𝐢𝐨𝐧
(𝑐ℎ𝑎𝑛𝑔𝑒 𝑜𝑓) ⏟ ( 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 )
𝑐𝑎𝑝𝑖𝑡𝑎𝑙 (𝐢𝐧𝐯𝐞𝐬𝐭𝐦𝐞𝐧𝐭)
______________________________________________________________________________
⏟̇ = 𝒔
𝒌 ⏟𝝓(𝒌) − 𝝀𝒌 ⏟̇ = 𝒔
𝒌 ⏟𝝓(𝒌) − 𝝀𝒌 ⏟̇ = 𝒔
𝒌 ⏟𝝓(𝒌) − 𝝀𝒌
⏟
𝑘̇ >0 𝑠 𝜙(𝑘)>𝜆𝑘 ⏟
𝑘̇ =0 𝑠 𝜙(𝑘)=𝜆𝑘 ⏟
𝑘̇ <0 𝑠 𝜙(𝑘)<𝜆𝑘
𝒌↑ 𝒌𝟏 𝑡𝑜 ⏟∗
𝒌 𝒌 𝑎𝑙𝑟𝑒𝑎𝑑𝑦 𝑎𝑡 ⏟∗
𝒌 𝒌↓ 𝒌𝟐 𝑡𝑜 ⏟∗
𝒌
𝒔𝒕𝒆𝒂𝒅𝒚 𝒔𝒕𝒂𝒕𝒆 𝒔𝒕𝒆𝒂𝒅𝒚 𝒔𝒕𝒂𝒕𝒆 𝒔𝒕𝒆𝒂𝒅𝒚 𝒔𝒕𝒂𝒕𝒆
( ) ( ) ( )
( 𝒆𝒒𝒖𝒊𝒍𝒊𝒃𝒓𝒊𝒖𝒎 ) ( 𝒆𝒒𝒖𝒊𝒍𝒊𝒃𝒓𝒊𝒖𝒎 ) ( 𝒆𝒒𝒖𝒊𝒍𝒊𝒃𝒓𝒊𝒖𝒎 )
This equation shows that investment-depreciation gap leads to investment. This causes capital
to grow and remove the gap. Investment (& hence capital) shall decrease if there is depreciation
is greater than investment.
Solow Growth Model:
______________________________________________________________________________
Solow Growth Model: Phase Diagram
In second panel of the Solow Growth Model, phase diagram shows an 𝒙 – intercept.
Before 𝒙 – intercept (𝒌∗ ) After 𝒙 – intercept (𝒌∗ )
Phase line is above 𝒙 – axis Phase line is below 𝒙 – axis
𝒅𝒌 𝒅𝒌
( > 𝟎) ( < 𝟎)
𝒅𝒕 𝒅𝒕
𝒅𝒌 𝒅𝒌
( > 𝟎) ⇒ 𝒌 ↑ ⇒ 𝑚𝑜𝑣𝑒 𝒓𝒊𝒈𝒉𝒕𝒘𝒂𝒓𝒅𝒔 ( < 𝟎) ⇒ 𝒌 ↓ ⇒ 𝑚𝑜𝑣𝑒 𝒍𝒆𝒇𝒕𝒘𝒂𝒓𝒅𝒔
𝒅𝒕 𝒅𝒕
Attraction towards 𝒙 – intercept (Attractor): Convergence/ Dynamic Stability
Stable Intertemporal Equilibrium OR Steady-State Equilibrium suggested by Robert
Solow in this Growth Model.
TOPIC 122: SOLOW GROWTH MODEL: QUANTITATIVE ANALYSIS
After understanding the qualitative analysis of the model, we conduct a quantitative analysis by
assuming a linearly homogeneous Cobb-Douglas function:
______________________________________________________________________________
𝑸 = 𝒇(𝑲, 𝑳) 𝑸 = 𝑲𝜶 𝑳𝟏−𝜶
Taking 𝑳 as common: 𝑸 = 𝑲𝜶 𝑳 𝑳−𝜶
𝑲 𝑲𝜶
𝑸 = 𝑳 {𝒇 ( , 𝟏)} 𝑸=𝑳 𝜶
𝑳 𝑳
𝑲
As ( 𝑳 = 𝒌), 𝒇 in terms of 𝑲 and 𝑳 while 𝝓 in 𝑲 𝜶
𝑸 = 𝑳( )
terms of 𝒌. 𝑳
𝑸 = 𝑳{𝝓(𝒌)} 𝑸 = 𝑳𝒌𝜶
Comparing the two results:
𝜙(𝑘)
𝑸=𝑳 𝒌 ⏞𝜶 ⇒ 𝝓(𝒌) = 𝒌𝜶
First Order Differential Equation of Solow Growth Model in terms of 𝒌 is:
𝒌̇ = 𝒔 𝝓(𝒌) − 𝝀𝒌
𝒌̇ = 𝒔 𝝓(𝒌)
⏟ − 𝝀𝒌
𝑘𝛼
𝒌̇ = 𝒔𝒌𝜶 − 𝝀𝒌
𝒌̇ + 𝝀𝒌 = 𝒔𝒌𝜶
𝒅𝒌
+ 𝝀𝒌 = 𝒔𝒌𝜶
𝒅𝒕
We have a First Order Differential of nonlinear nature due to 𝒌𝜶 with 𝜶 ≠ 1. This makes a
Bernoulli equation:
𝒅𝒚
( + 𝑹𝒚 = 𝑻𝒚𝒎 )
𝒅𝒕
𝑚≠1
𝒅𝒌 ⏞
+⏟𝝀𝒌⏟=⏟ ⏟ 𝜶
𝒔𝒌
⏟
𝒅𝒕 𝑅 𝑦 𝑇 𝑦
𝑑𝑦
𝑑𝑡
Solution: Bernoulli Equation (nonlinear differential equation) can be reduced to a linear
differential equation and be solved as such.
𝒅𝒚 𝒅𝒌
Here; 𝒅𝒕 = 𝒅𝒕 , 𝒚 = 𝒌, 𝑹 = 𝝀, 𝑻 = 𝒔, 𝒎 = 𝜶.
Reduction to Linear Form: Standard form of Bernoulli Equation:
𝒅𝒚 𝑙𝑖𝑛𝑒𝑎𝑟𝑖𝑧𝑎𝑡𝑖𝑜𝑛
+ 𝑹𝒚 = 𝑻𝒚𝒎 → 𝒅𝒛 + [(𝟏 − 𝒎)𝑹𝒛 − (𝟏 − 𝒎)𝑻]𝒅𝒕 = 𝟎
𝒅𝒕
Linearization of Solow’s Growth Differential Equation can also be done by recalling the linearized
form of Bernoulli’s equation and putting the values:
𝒅𝒛 + [(𝟏 − 𝒎)𝑹𝒛 − (𝟏 − 𝒎)𝑻]𝒅𝒕 = 𝟎
𝒅𝒛 + [(𝟏 − 𝒎 ) 𝑹 𝒛 − (𝟏 − 𝒎 ) 𝑻 ]𝒅𝒕 = 𝟎
𝒅𝒛 + [(𝟏 − 𝜶 ) 𝝀 𝒛 − (𝟏 − 𝜶 ) 𝒔 ]𝒅𝒕 = 𝟎
𝒅𝒛 + [(𝟏 − 𝜶)𝝀𝒛 − (𝟏 − 𝜶)𝒔]𝒅𝒕 = 𝟎
⏟
𝐿𝑖𝑛𝑒𝑎𝑟𝑖𝑧𝑒𝑑 𝑓𝑜𝑟𝑚 𝑜𝑓 𝑆𝑜𝑙𝑜𝑤 ′ 𝑠 𝐺𝑟𝑜𝑤𝑡ℎ
𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑙 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛
Dividing throughout by 𝒅𝒕 and rearranging to develop the standard form of First Order
Differential equation.
𝒅𝒛
+ (𝟏 − 𝜶)(𝝀 𝒛 − 𝒔) = 𝟎
𝒅𝒕
𝒅𝒛
+ (𝟏 − 𝜶)𝝀 𝒛 − (𝟏 − 𝜶)𝒔 = 𝟎
𝒅𝒕
______________________________________________________________________________
𝒅𝒛
+ (𝟏 − 𝜶)𝝀 𝒛 = (𝟏 − 𝜶)𝒔
𝒅𝒕
𝒅𝒚
Comparable with 𝒅𝒕 + 𝒂 𝒚 = 𝒃
𝒂 = (𝟏 − 𝜶)𝝀 and 𝒃 = (𝟏 − 𝜶)𝒔
𝒅𝒛
(𝟏 − 𝜶)𝝀 𝒛 = ⏟
+⏟ (𝟏 − 𝜶)𝒔
𝒅𝒕
𝒂 𝒃
Definite Solution:
𝒃 𝒃
𝒚(𝒕) = {𝒚(𝟎) − } 𝒆−𝒂𝒕 +
𝒂 𝒂
(𝟏 − 𝜶)𝒔 −(𝟏−𝜶)𝝀𝒕 (𝟏 − 𝜶)𝒔
𝒛(𝒕) = {𝒛(𝟎) − }𝒆 +
(𝟏 − 𝜶)𝝀 (𝟏 − 𝜶)𝝀
𝒔 𝒔
𝒛(𝒕) = {𝒛(𝟎) − } 𝒆−(𝟏−𝜶)𝝀𝒕 +
𝝀 𝝀
Time-path of 𝒛. However, the original function was 𝒌.
𝒔 𝒔
Therefore, recalling 𝒛 = 𝒌𝟏−𝜶 and substituting in 𝒛(𝒕) = {𝒛(𝟎) − 𝝀} 𝒆−(𝟏−𝜶)𝝀𝒕 + 𝝀
𝒔 𝒔
𝒌𝟏−𝜶 = [{𝒌(𝟎)}𝟏−𝜶 − ] 𝒆−(𝟏−𝜶)𝝀𝒕 +
𝝀 𝝀
Here 𝒌(𝟎) is the initial value of capital-labor ratio.
Dynamic Stability:
𝐴𝑟𝑏𝑖𝑡𝑟𝑎𝑟𝑦 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
𝑖𝑛 𝑡𝑒𝑟𝑚𝑠 𝑜𝑓 𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙
𝑖𝑛𝑖𝑡𝑖𝑎𝑙 𝑐𝑜𝑛𝑑𝑖𝑡𝑖𝑜𝑛 𝑒𝑥𝑝𝑟𝑒𝑠𝑠𝑖𝑜𝑛
⏞ 𝒔 −(𝟏−𝜶)𝝀𝒕 𝒔
𝒌𝟏−𝜶 = [{𝒌(𝟎)}𝟏−𝜶 − ] ⏞
𝒆 +
⏟ 𝝀 ⏟
𝝀
𝑘𝑐 𝑘𝑝
−(1−𝛼)𝜆𝒕
{𝒆
⏟ } explains the dynamic stability of time path.
𝑇𝑖𝑚𝑒 𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡
𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙 𝑒𝑥𝑝𝑟𝑒𝑠𝑠𝑖𝑜𝑛
For dynamic stability, 𝒆−(1−𝛼)𝜆𝒕 → 𝟎 , as 𝒕 → ∞
Unraveling the Exponent (𝟏 − 𝜶)𝝀 > 𝟎:
As already known, that (𝟎 < 𝜶 < 𝟏) and 𝝀 > 𝟎 (assumed in the model).
(>0,<1) (>0)
(𝟏
⏟− 𝜶 ⏞ ) ⏞
𝝀 >𝟎
⏟ (>0,<1)
>0
𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙
𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑐𝑎𝑛𝑑 𝑜𝑓 𝑑𝑒𝑐𝑎𝑦
𝑒𝑥𝑝𝑜𝑒𝑛𝑡𝑖𝑎𝑙 𝑑𝑒𝑐𝑎𝑦 𝑡𝑒𝑟𝑚
⏞ (>0)
𝟏−𝜶
⏞ 𝒔 ⏞ 𝒔
𝒌 = [{𝒌(𝟎)}𝟏−𝜶 − ] 𝒆−(𝟏−𝜶)𝝀𝒕 +
⏟ 𝝀 𝝀
→𝟎 𝑎𝑠 𝒕→∞
𝒔
𝒌𝟏−𝜶 → 𝒌 as 𝑡 → ∞
Since we need the Original value of 𝒛, we shift its power to R.H.S
𝟏
𝒔 𝟏−𝜶
𝒌(𝒕) → (𝝀) as 𝑡 → ∞.
Interpretation: Capital-labor (𝒌) ratio assumes a constant value at equilibrium. Which varies
directly w.r.t. 𝒔 and indirectly w.r.t. 𝝀.
Appendix: Linearization via Derivation:
However, linearization can be derived via following steps:
______________________________________________________________________________
Since 𝒌𝜶 is the cause of nonlinearity, we reduce it to 1 by self-dividing:

𝒅𝒌
( + 𝝀𝒌) 𝒔 𝒌𝜶
𝒅𝒕 = 𝜶
𝒌𝜶 𝒌
𝟏 𝒅𝒌 𝒌 𝒔 𝒌𝜶
× + 𝝀 =
𝒌𝜶 𝒅𝒕 𝒌𝜶 𝒌𝜶
𝒅𝒌
𝒌−𝜶 × + 𝝀(𝒌 × 𝒌−𝜶 ) = 𝒔
𝒅𝒕
𝒅𝒛
(𝒌−𝜶 ) +𝝀 𝒌𝟏−𝜶
⏟ =𝒔
⏟ 𝒅𝒕
Boxed terms, if substituted will free equation from nonlinear term of 𝒛.
Substitution – I: Substitution – II: 𝒛 depends on 𝒌, which further depends on 𝒕: 𝒛 =

Introducing a shorthand 𝒇(𝒌) and 𝒌 = 𝒈(𝒕); combining the two function: 𝒛 = 𝒇{𝒌(𝒕)}
notation 𝒛 for 𝒌𝟏−𝜶 : We get a composite function, which if differentiated, calls for Chain
𝒛 = 𝒌𝟏−𝜶 𝒅𝒛 𝒅𝒛 𝒅𝒌
Rule: 𝒅𝒕 = 𝒅𝒌 × 𝒅𝒕
𝟏−𝜶
⏟
𝒛 = 𝒌⏟ 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑡𝑒
𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑙 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑙
𝑙𝑖𝑛𝑒𝑎𝑟 𝑛𝑜𝑛𝑙𝑖𝑛𝑒𝑎𝑟
𝑡𝑒𝑟𝑚 𝑡𝑒𝑟𝑚 𝑡𝑒𝑟𝑚 𝑜𝑓 𝑛𝑒𝑤 𝒛=𝒌𝟏−𝜶 𝑡𝑒𝑟𝑚 𝑜𝑓 𝑛𝑒𝑤
𝟏−𝜶 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛 𝒛=𝒇(𝒕) 𝑤.𝑟.𝑡 𝒌 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛 𝒌=𝒇(𝒕)
Nonlinear term 𝒌 will
be substituted by linear ⏞
𝒅𝒛 ⏞𝒅𝒛 ⏞𝒅𝒌
term of 𝒛. = ( ) × ( )
𝒅𝒕 𝒅𝒌 𝒅𝒕
Differentiate 𝒛 = 𝒌𝟏−𝜶 w.r.t 𝒌:
𝒅𝒛 𝒅 𝒅
= (𝒌𝟏−𝜶 ) = (𝟏 − 𝜶) × (𝒌𝟏−𝜶−𝟏 ) × (𝒌)
𝒅𝒌 𝒅𝒌 𝒅𝒌
𝒅𝒛
= (𝟏 − 𝜶) × (𝒌−𝜶 )
𝒅𝒌
𝒅𝒛 𝒅𝒛 𝒅𝒛 𝒅𝒌
Substituting value of 𝒅𝒌
in chain rule; 𝒅𝒕
= 𝒅𝒌
× 𝒅𝒕
𝒅𝒛 𝒅𝒌
= (𝟏 − 𝜶) × (𝒌−𝜶 ) ×
𝒅𝒕 𝒅𝒕
𝟏 𝒅𝒛 𝒅𝒌
× = (𝒌−𝜶 )
(𝟏 − 𝜶) 𝒅𝒕 𝒅𝒕
______________________________________________________________________________
Combining the Substitution – I & Substitution – II:

𝒅𝒌
Now we can surrogate 𝒛 instead of 𝒌 in (𝒌−𝒎 ) +𝝀 𝒌𝟏−𝜶
⏟ =𝒔
⏟ 𝒅𝒕
𝟏 𝒅𝒛
× +𝝀 𝒛
⏟ =𝒔
(𝟏 − 𝜶) 𝒅𝒕
⏟ 𝑺𝒖𝒃𝒔𝒕𝒊𝒕𝒖𝒕𝒊𝒐𝒏−𝑰
𝟏 𝒅𝒛
× + 𝝀𝒛 = 𝒔
(𝟏 − 𝜶) 𝒅𝒕
𝒅𝒛 + 𝝀𝒛(𝟏 − 𝜶)𝒅𝒕
=𝒔
(𝟏 − 𝜶)𝒅𝒕
𝒅𝒛 + 𝝀𝒛(𝟏 − 𝜶)𝒅𝒕 = 𝒔 × (𝟏 − 𝜶)𝒅𝒕
𝒅𝒛 + 𝝀𝒛(𝟏 − 𝜶)𝒅𝒕 − 𝒔(𝟏 − 𝜶)𝒅𝒕 = 𝟎
𝒅𝒛 + {(𝟏 − 𝜶) × (𝝀𝒛 − 𝒔)}𝒅𝒕 = 𝟎
Objective of replacing a nonlinear term 𝒌𝜶 with linear term 𝒛(𝟏)
𝒛 𝑖𝑛𝑠𝑡𝑒𝑎𝑑
𝑜𝑓 𝒌
𝒅𝒛 + {(𝟏 − 𝜶) × (𝝀 ⏟ ⏞
𝒛 − 𝒔)} 𝒅𝒕 = 𝟎
𝑙𝑖𝑛𝑒𝑎𝑟
𝑡𝑒𝑟𝑚
Numerical Evidence
𝒌(𝟎) = 𝟏𝟎𝟎, 𝒔 = 𝟎. 𝟐, 𝜶 = 𝟎. 𝟕, 𝝀 = 𝟐. 𝟓%(= 𝟎. 𝟎𝟐𝟓)
𝒔 𝒔 𝒔
𝒕 𝒆−(𝟏−𝜶)𝝀𝒕 [{𝒌(𝟎)}𝟏−𝜶 − 𝝀] 𝒆−(𝟏−𝜶)𝝀𝒕 𝒌𝟏−𝜶 = [{𝒌(𝟎)}𝟏−𝜶 − ] 𝒆−(𝟏−𝜶)𝝀𝒕 +
𝝀 𝝀
0 1 -4.019 3.981
1 0.993 -3.989 4.011
100 0.472 -1.898 6.102
200 0.223 -0.897 7.103
300 0.105 -0.424 7.576
400 0.050 -0.200 7.800
500 0.024 -0.095 7.905
600 0.011 -0.045 7.955
700 0.005 -0.021 7.979
800 0.002 -0.010 7.990
900 0.001 -0.005 7.995
998 0.001 -0.002 7.998
Exponential 𝒔
decay 𝒌𝒄 fading to 𝟎 as 𝒕 → ∞ Nearing 𝒌∗ = 𝝀 = 𝟖
Exponential Decay causes 𝒌𝒄 to decay which converges the 𝒌𝟏−𝜶 (𝒕) to 𝒌𝒑 (= 𝒌∗ ) = 𝟖.
______________________________________________________________________________
𝟏
𝟏−𝜶 𝒔 𝒔 𝟏−𝜶
𝒌 → or 𝒌 →
𝒌
(𝝀) as 𝒕 → ∞.
In economic terms, (𝒌∗ )
represents the Steady-State equilibrium suggested in Solow’s Growth
Model. So, the model shows Convergence towards Steady-State equilibrium.
______________________________________________________________________________
Lesson 28
SECOND ORDER DIFFERENTIAL EQUATIONS
TOPIC 123: HIGHER (SECOND) ORDER DIFFERENTIAL EQUATION
In economic analysis, Rate of change of the rate of change can be question regarding economic
variables e.g. income, profits etc.
Such situations can be mathematized in second (higher) order differential equations; e.g.
𝒅𝟐 𝒚
= 𝒌𝒚
𝒅𝒕𝟐
𝒅 𝒅
× (𝒚)
⏟ = 𝒌𝒚
⏟
𝒅𝒕 ⏟
𝒅𝒕 𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡
𝑟𝑎𝑡𝑒 𝑜𝑓 𝑟𝑎𝑡𝑒 𝑜𝑓
𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒
𝑐ℎ𝑎𝑛𝑔𝑒 𝑐ℎ𝑎𝑛𝑔𝑒
Since derivatives can have any order higher than second order, we generalize the higher order
differential equations:
Linear Differential Equation with 𝒏𝒕𝒉 Order:
Leibniz Notation:
𝒅𝒏 𝒚 𝒅𝒏−𝟏 𝒚 𝒅𝒏−𝟐 𝒚 𝒅𝒚
(⏟ 𝒏 ) + 𝒂
⏟𝟏 × (⏟ 𝒏−𝟏 ) + 𝒂
⏟𝟐 × (⏟ 𝒏−𝟐 ) + ⋯ + 𝒂𝒏−𝟏 × (⏟) + 𝒂
⏟𝒏 (𝒚)
⏟ = ⏟
𝒃
𝒅𝒕 𝒅𝒕 𝒅𝒕 𝒅𝒕
1𝑠𝑡 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 2𝑛𝑑 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑛𝑡ℎ 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 𝑐𝑜𝑛𝑠𝑡𝑎
𝑛𝑡ℎ𝑂𝑟𝑑𝑒𝑟 (𝑛−1)𝑡ℎ (𝑛−2)𝑡ℎ 1𝑠𝑡𝑂𝑟𝑑𝑒𝑟
𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝐷𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒
𝐷𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑂𝑟𝑑𝑒𝑟 𝑂𝑟𝑑𝑒𝑟
𝐷𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝐷𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒
Lagrange Notation:
𝒚(𝒏) (𝒕) + 𝒂𝟏 𝒚(𝒏−𝟏) (𝒕) + 𝒂𝟐 𝒚(𝒏−𝟐) (𝒕) + ⋯ + 𝒂𝒏−𝟏 𝒚(𝟏) (𝒕) + 𝒂𝒏 𝒚 = 𝒃
Order: 𝒏𝒕𝒉 (due to 𝒏𝒕𝒉 derivative)
Degree: Linear/First degree (due to highest power ‘𝟏’ of the 𝒏𝒕𝒉 derivative)
Constant Coefficients (𝒂′ 𝑠) and constant term (𝒃).
Delimitations:
1. For simplicity variable coefficients and variable term are not considered.
2. This course deals with linear cases only.
3. Cross terms – product of function ′𝒚(𝒕)′ and its derivative 𝒚′ (𝒕) – are not included.
(𝒏 = 𝟐) ⇒ Second Order Differential Equation (SODE):
For (𝒏 = 𝟐), Standard Form of Second-Order Differential Equation reduces to:
𝒅𝟐 𝒚 𝒅𝒚
𝟐
+ 𝒂𝟏 ∙ + 𝒂𝟐 ∙ 𝒚 = 𝒃
𝒅𝒕 𝒅𝒕
𝒚′′ (𝒕) + 𝒂𝟏 ∙ 𝒚′ (𝒕) + 𝒂𝟐 ∙ 𝒚 = 𝒃
Possible Cases: Depending on the values of 𝒃, we get 2 possible broad cases:
Main Cases based on value of (𝒃)
Identification 𝒃=𝟎 𝒃≠𝟎
Terminology Homogenous Case Non-Homogenous Case
Components of 𝒚𝒑 = 𝟎 𝒚𝒑 ≠ 𝟎
Solution
Sub-Cases based on values of (𝒂𝟏 ) and/or (𝒂𝟐 )
Identification 𝒂𝟐 ≠ 𝟎, 𝒂𝟐 = 𝟎, 𝒂𝟐 = 𝟎, 𝒂𝟐 ≠ 𝟎, 𝒂𝟐 = 𝟎, 𝒂𝟐 = 𝟎,
(𝒂𝟏 ≠ 𝟎) 𝒂𝟏 ≠ 𝟎 𝒂𝟏 = 𝟎 (𝒂𝟏 ≠ 𝟎) 𝒂𝟏 ≠ 𝟎 𝒂𝟏 = 𝟎
{ ⋁ } { ⋁ }
(𝒂𝟏 = 𝟎) (𝒂𝟏 = 𝟎)
______________________________________________________________________________
Particular Solution (𝟎) (𝟎) 𝒚𝒑 = 𝒃 𝒃 𝒃

𝒚𝒑 = 𝒚𝒑 = 𝒕 𝒚𝒑 = 𝒂 𝒚𝒑 = 𝒂 𝒕 𝒚𝒑 = 𝟐 𝒕𝟐
𝒂𝟐 𝒂𝟏 (𝟎) 𝟐
(𝒚𝒑 ) 𝒕
𝟐 𝟏
TOPIC 124: SECOND-ORDER DIFFERENTIAL EQUATION: PARTICULAR INTEGRAL
Particular Integral (𝒚𝒑 ): Any particular solution (containing no arbitrary constants) of complete
equation. 𝒚𝒑 provides equilibrium value of 𝒚(𝒕), in intertemporal sense:
𝒚(𝒕)
⏟ = 𝒚𝒄
⏟ + 𝒚
⏟𝒑
𝑇𝑖𝑚𝑒 𝑃𝑎𝑡ℎ 𝐷𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑓𝑟𝑜𝑚 𝐸𝑞𝑢𝑖𝑙𝑖𝑏𝑟𝑖𝑢𝑚
𝑒𝑞𝑢𝑖𝑙𝑖𝑏𝑟𝑖𝑢𝑚 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝒚(𝒕)
𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝒚(𝒕)
Subcase – I: Since the particular integral (𝒚𝒑 ) can be any solution, i.e., any value of 𝒚 that
satisfies this nonhomogeneous equation, we should always try the simplest possible type;
namely, 𝒚 = 𝒂 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕. If 𝒚 = 𝒌, it follows that:
𝒚 = 𝒚(𝒕) = 𝒌
𝒚′ (𝒕) = 𝟎
𝒚′′ (𝒕) = 𝟎
Substituting the values of 𝒚(𝒕), 𝒚′ (𝒕) and 𝒚′′ (𝒕) in the general form 𝒚′′ (𝒕) + 𝒂𝟏 ∙ 𝒚′ (𝒕) + 𝒂𝟐 ∙
𝒚(𝒕) = 𝒃:
𝟎 + 𝒂𝟏 ∙ 𝟎 + 𝒂𝟐 ∙ 𝒌 = 𝒃
𝒂𝟐 ∙ 𝒌 = 𝒃
𝒃
𝒌=
𝒂𝟐
Since 𝒌 is a particular value of 𝒚, that satisfies the differential equation, we can call it Particular
Solution (𝒚𝒑 ):
𝒃
𝒚𝒑 =
𝒂𝟐
𝒃
𝒚𝒑 =
𝒂𝟐
Interpretation: Particular Solution has actually a constant value as 𝒃 and 𝒂𝟐 are constants and
so will be their ratio.
𝒃
Subcase – II: If 𝒂𝟐 = 𝟎, the formula of particular solution (𝒚𝒑 = 𝒂𝟐
) shall give undefined answer.
𝒃
i.e. (𝒚𝒑 = 𝟎
= ∞). Therefore, we need some other formula for (𝒚𝒑 ). For another formula, we need
another assumption regarding solution value of 𝒚(𝒕). Instead of considering constant a solution
𝒚, we try a nonconstant (variable); i.e. 𝒚 = 𝒌𝒕.
𝒚(𝒕) = 𝒌𝒕
𝒚′ (𝒕) = 𝒌
𝒚′′ (𝒕) = 𝟎
Substituting the values of 𝒚(𝒕), 𝒚′ (𝒕), 𝒚′′ (𝒕) and 𝒂𝟐 in the general form 𝒚′′ (𝒕) + 𝒂𝟏 ∙ 𝒚′ (𝒕) + 𝒂𝟐 ∙
𝒚(𝒕) = 𝒃:
𝟎 + 𝒂𝟏 ∙ 𝒌 + 𝟎 ∙ 𝒌𝒕 = 𝒃
𝒂𝟏 ∙ 𝒌 + 𝟎 = 𝒃
______________________________________________________________________________
𝒃
𝒌 =
𝒂𝟏
Since the assumed solution value of 𝒚 was 𝒚 = 𝒌 𝒕, we substitute the value of 𝒌 to get the
solution value of 𝒚:
𝒃
𝒚= 𝒕
𝒂𝟏
Since solution value is actually the Particular Solution (𝒚𝒑 ):
𝒃
𝒚𝒑 = ( ) 𝒕
𝒂𝟏
Interpretation: Particular Solution has actually a nonconstant value regardless of 𝒃 and 𝒂𝟏
𝒃
being constants. It is due to 𝒕 that gets multiplied with their ratio ( ).
𝒂𝟏
Subcase – III: If 𝒂𝟏 = 𝟎 and already 𝒂𝟐 = 𝟎, we check if Subcase – I and Subcase – II formula

of particular solution are applicable:
Subcase – I Subcase – II
𝒃 𝒃
𝒚𝒑 = 𝒚𝒑 = 𝒕
𝒂𝟏 𝒂𝟐
Substituting values of Subcase – III (𝒂𝟏 = 𝟎, 𝒂𝟐 = 𝟎)
𝒃 𝒃
𝒚𝒑 = = ∞ 𝒚𝒑 = 𝒕 = ∞
𝟎 𝟎
Therefore, the particular integral using both Subcase – III give an undefined answer.
We need another formula to find the value of 𝒚𝒑 .
Building on the assumptions 𝒚 = 𝒌 and 𝒚 = 𝒌𝒕, we use 𝒚 = 𝒌𝒕𝟐 in Subcase – III.
𝒚(𝒕) = 𝒌𝒕𝟐
𝒚′ (𝒕) = 𝟐𝒌𝒕
𝒚′′ (𝒕) = 𝟐𝒌
Substituting the values of 𝒚(𝒕), 𝒚′ (𝒕), 𝒚′′ (𝒕), 𝒂𝟏 and 𝒂𝟐 in the standard form: 𝒚′′ (𝒕) + 𝒂𝟏 ∙
𝒚′ (𝒕) + 𝒂𝟐 ∙ 𝒚(𝒕) = 𝒃:
𝟐𝒌 + 𝟎 ∙ 𝟐𝒌𝒕 + 𝟎 ∙ 𝒌𝒕 = 𝒃
𝟐𝒌 + 𝟎 + 𝟎 = 𝒃
𝟐𝒌 = 𝒃
𝒃
𝒌 =
𝟐
Since the assumed solution value of 𝒚 was 𝒚 = 𝒌 𝒕𝟐 , we substitute the value of 𝒌 to get the
𝒚 = 𝒌 𝒕𝟐
𝒃
𝒚𝒑 = ( ) 𝒕𝟐
𝟐
Interpretation: Particular Solution has actually a nonconstant value regardless of 𝒃 being
𝒃
constant. It is due to 𝒕𝟐 that gets multiplied with their ratio (𝟐).
______________________________________________________________________________
TOPIC 125: SECOND-ORDER DIFFERENTIAL EQUATION: PARTICULAR INTEGRAL

NUMERICAL
Second-Order Differential Equation: Particular Integral Numerical

Non-Homogenous Subcases
(𝒂𝟏 ≠ 𝟎)
Non-Homogenous Subcase – I: 𝒃 ≠ 𝟎, 𝒂𝟐 ≠ 𝟎, { ⋁ }.
(𝒂𝟏 = 𝟎)
Example: Consider a Second Order Differential Equation:
𝒚′′ (𝒕) + 𝒚′ (𝒕) − 𝟐𝒚 = −𝟏𝟎
Comparing with the given specific equation with the standard form:
𝒚′′ (𝒕) + 𝒂𝟏 𝒚′ (𝒕) + 𝒂𝟐 𝒚 = 𝒃
′′ (𝒕)
𝒚 ⏟ × 𝒚′ (𝒕) + (−𝟐)
+ (𝟏) ⏟ × 𝒚 = −𝟏𝟎
⏟
𝒂𝟏 𝒂𝟐 𝑏
𝑵𝒐𝒏−𝑯𝒐𝒎𝒐𝒈𝒆𝒏𝒐𝒖𝒔
𝑺𝒖𝒃𝒄𝒂𝒔𝒆−𝑰 𝑪𝒂𝒔𝒆
Here; ⏞
(𝒂𝟏 = 𝟏 ≠ 𝟎), (𝒂𝟐 = −𝟐 ≠ 𝟎), ⏞ 𝒃 = −𝟏𝟎 ≠ 𝟎 .
𝒃
Suitable Formula for Particular Solution: 𝒚𝒑 = 𝒂
𝟐
−𝟏𝟎
𝒚𝒑 =
−𝟐
𝒚𝒑 = 𝟓
Interpretation: 𝒚𝒑 = 𝟓 is particular solution of the Non-Homogenous Subcase – I of Second
Order Differential Equation 𝒚′′ (𝒕) + 𝒚′ (𝒕) − 𝟐𝒚 = −𝟏𝟎 .
Non-Homogenous Case – II: 𝒃 ≠ 𝟎, 𝒂𝟐 = 𝟎, 𝒂𝟏 ≠ 𝟎.
𝒚′′ (𝒕) + 𝒚′ (𝒕) = −𝟏𝟎
𝒚′′ (𝒕) + 𝒂𝟏 𝒚′ (𝒕) + 𝒂𝟐 𝒚 = 𝒃
𝑠𝑒𝑙𝑓
𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑒𝑑
⏟ × 𝒚′ (𝒕) + ⏞
𝒚′′ (𝒕) + (𝟏) (𝟎)
⏟ × 𝒚 = −𝟏𝟎
⏟
𝑺𝒖𝒃𝒄𝒂𝒔𝒆−𝑰𝑰 𝑪𝒂𝒔𝒆
Here; ⏞(𝒂𝟏 = 𝟏 ≠ 𝟎), (𝒂𝟐 = 𝟎), ⏞𝒃 = −𝟏𝟎 ≠ 𝟎 .
𝒃
Suitable Formula for Particular Solution: 𝒚𝒑 = 𝒂 𝒕
𝟏
−𝟏𝟎
𝒚𝒑 = ×𝒕
𝟏
𝒚𝒑 = −𝟏𝟎𝒕
Interpretation: 𝒚𝒑 = −𝟏𝟎𝒕 is particular solution of the Non-Homogenous Case – II of Second
Order Differential Equation 𝒚′′ (𝒕) + 𝒚′ (𝒕) = −𝟏𝟎 .
Non-Homogenous Case – III: 𝒃 ≠ 𝟎, 𝒂𝟐 = 𝟎, 𝒂𝟏 = 𝟎,
𝒚′′ (𝒕) = −𝟏𝟎
𝒚′′ (𝒕) + 𝒂𝟏 𝒚′ (𝒕) + 𝒂𝟐 𝒚 = 𝒃
______________________________________________________________________________
𝑠𝑒𝑙𝑓 𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑒𝑑
𝒚′′ (𝒕) + ⏞⏟ × 𝒚′ (𝒕) + (𝟎)

(𝟎) ⏟ × 𝒚 = −𝟏𝟎
⏟
𝑺𝒖𝒃𝒄𝒂𝒔𝒆−𝑰𝑰𝑰 𝑪𝒂𝒔𝒆
Here; ⏞
(𝒂𝟏 = 𝟎), (𝒂𝟐 = 𝟎), 𝒃 ⏞ = −𝟏𝟎 ≠ 𝟎 .
𝒃
Suitable Formula for Particular Solution: 𝒚𝒑 = 𝟐 𝒕𝟐
−𝟏𝟎
𝒚𝒑 = × 𝒕𝟐
𝟐
𝒚𝒑 = −𝟓𝒕𝟐
Interpretation: 𝒚𝒑 = −𝟓𝒕𝟐 is particular solution of the Non-Homogenous Case – III of Second
Order Differential Equation 𝒚′′ (𝒕) = −𝟏𝟎 .
Summary: Non-Homogenous Case
Subcase – I Subcase – II Subcase – III
′′ (𝒕) ′ (𝒕) ′′ (𝒕) ′ (𝒕)
𝒚 +𝒚 − 𝟐𝒚 = −𝟏𝟎 𝒚 +𝒚 = −𝟏𝟎 𝒚′′ (𝒕) = −𝟏𝟎
𝒃 𝒃 𝒃 𝟐
𝒚𝒑 = 𝒚𝒑 = 𝒕 𝒚𝒑 = 𝒕
𝒂𝟐 𝒂𝟏 𝟐
In Non-Homogenous case, 𝒃 = −𝟏𝟎
−𝟏𝟎 −𝟏𝟎 −𝟏𝟎 𝟐
−𝟐 𝟏 𝟐
𝒚𝒑 = 𝟓 𝒚𝒑 = −𝟏𝟎𝒕 𝒚𝒑 = −𝟓𝒕𝟐
HOMOGENOUS SUBCASES
Since formulae of all subcases of 𝒚𝒑 have 𝒃 in numerator:
Summary: Homogenous Case
𝒚′′ (𝒕) + 𝒚′ (𝒕) − 𝟐𝒚 = 𝟎 𝒚′′ (𝒕) + 𝒚′ (𝒕) = 𝟎 𝒚′′ (𝒕) = 𝟎
𝒃 𝒃 𝒃 𝟐
𝒂𝟐 𝒂𝟏 𝟐
In Homogenous case, 𝒃 = 𝟎
𝟎 𝟎 𝟎 𝟐
−𝟐 𝟏 𝟐
𝒚𝒑 = 𝟎 𝒚𝒑 = 𝟎 𝒚𝒑 = 𝟎
All subcases have a zero value of particular solution; i.e. (𝒚𝒑 = 𝟎).
TOPIC 126: SECOND-ORDER DIFFERENTIAL EQUATION: STANDARD FORM &

COMPLEMENTARY FUNCTION
Complementary Function (𝒚𝒄 ): General solution (containing arbitrary constants) of reduced

equation. 𝒚𝒄 provides deviation of time path 𝒚(𝒕) from equilibrium value, at every point in time.
𝒚(𝒕)
⏟ = 𝒚
⏟𝒄 + 𝒚
⏟𝒑
______________________________________________________________________________
This is why, for dynamic stability, we want 𝒚𝒄 term to disappear. So that 𝒚(𝒕) reaches 𝒚𝒑 : 𝒚(𝒕)→
𝒚𝒑 when 𝒚𝒄 → 𝟎.
Derivation of the Complementary Function (𝒚𝒄 ):

Exponential expressions (𝑨𝒆𝒓𝒕 ) served well in complementary functions for differential equations
as first order differential equations.
Experimenting with a comparison numerical instance:
𝒚(𝒕) = 𝑨𝒆𝒓𝒕 , For simplicity; 𝑨 = 𝟏 𝒚(𝒕) = 𝒕𝟐
𝒚(𝒕) = 𝒆𝒓𝒕 𝒚′ (𝒕) = 𝟐𝒕
𝒓𝒕
𝒚(𝒕) = 𝒓𝒆 𝒚′′ (𝒕) = 𝟐
𝟐 𝒓𝒕
𝒚(𝒕) = 𝒓 𝒆
Assuming a second order differential equation:
𝒚′′ (𝒕) + 𝒂𝟏 𝒚′ (𝒕) + 𝒂𝟐 𝒚(𝒕) = 𝟎
𝒓𝟐 𝒆𝒓𝒕 + 𝒂𝟏 𝒓𝒆𝒓𝒕 + 𝒂𝟐 𝒆𝒓𝒕 = 𝟎 𝟐 + 𝒂𝟏 𝟐𝒕 + 𝒂𝟐 𝒕𝟐 = 𝟎
Reduction of 𝑳. 𝑯. 𝑺 to zero (𝟎) is possible as Reduction of 𝑳. 𝑯. 𝑺 to zero (𝟎) is not possible
all variable expression 𝒆𝒓𝒕 are same. This, off as there are two different functions i.e. linear
course, will happen with the help of certain (𝟐𝒕) and quadratic (𝒕𝟐 ). Here, values of 𝒂𝟏
values of 𝒓, 𝒂𝟏 and 𝒂𝟐 . and 𝒂𝟐 cannot help in reducing of 𝑹. 𝑯. 𝑺. to
zero (𝟎).
Therefore, we can develop Characteristic Equation, by substituting the value of 𝒚(𝒕), 𝒚′ (𝒕) and
𝒚′′ (𝒕):
𝒚(𝒕) = 𝑨𝒆𝒓𝒕
𝒚′ (𝒕) = 𝒓𝑨𝒆𝒓𝒕
𝒚′′ (𝒕) = 𝒓𝟐 𝑨𝒆𝒓𝒕
𝒚′′ (𝒕) + 𝒂𝟏 𝒚′ (𝒕) + 𝒂𝟐 𝒚(𝒕) = 𝟎
𝒓𝟐 𝑨𝒆𝒓𝒕 + 𝒂𝟏 𝒓𝑨𝒆𝒓𝒕 + 𝒂𝟐 𝑨𝒆𝒓𝒕 = 𝟎
𝑨𝒆𝒓𝒕 (𝒓𝟐 + 𝒂𝟏 𝒓 + 𝒂𝟐 ) = 𝟎
𝒓𝟐 + 𝒂𝟏 𝒓 + 𝒂𝟐 = 𝟎
𝟐
Comparing 𝒂𝒙 + 𝒃𝒙 + 𝒄 = 𝟎 we get 𝒙 = 𝒓, 𝒂 = 𝟏, 𝒃 = 𝒂𝟏 , 𝒄 = 𝒂𝟐 .
−𝒃±√𝒃𝟐 −𝟒𝒂𝒄
Quadratic Formula: 𝒙 = 𝟐𝒂
becomes:
−𝒂𝟏 ± √𝒂𝟐𝟏 − 𝟒𝒂𝟐

𝒓=
𝟐
−𝒂𝟏 +√𝒂𝟐𝟏 −𝟒𝒂𝟐 −𝒂𝟏 −√𝒂𝟐𝟏 −𝟒𝒂𝟐
𝒓𝟏 = 𝟐
and 𝒓𝟐 = 𝟐
Two Relationships to verify the Roots (𝒓𝟏 , 𝒓𝟐 ):
Relationship – I: Sum of the two roots equals (−𝒂𝟏 ).
−𝒂𝟏 +√𝒂𝟐𝟏 −𝟒𝒂𝟐 −𝒂𝟏 −√𝒂𝟐𝟏 −𝟒𝒂𝟐 𝒂𝟏
√𝒂𝟐𝟏 −𝟒𝒂𝟐
𝒂𝟏
√𝒂𝟐𝟏 −𝟒𝒂𝟐
𝒂𝟏 𝒂𝟏
𝒓𝟏 + 𝒓𝟐 = ( 𝟐
)+( 𝟐
)⇒− 𝟐
+ 𝟐
+− 𝟐
− 𝟐
⇒− 𝟐
− 𝟐
⇒
−𝒂𝟏 −𝒂𝟏 −𝟐𝒂

𝟏
⇒ ⇒ −𝒂𝟏
𝟐 𝟐
Relationship – II: Product of the two roots equals (𝒂𝟐 ).
______________________________________________________________________________
𝟐
−𝒂𝟏 + √𝒂𝟐𝟏 − 𝟒𝒂𝟐 −𝒂𝟏 − √𝒂𝟐𝟏 − 𝟒𝒂𝟐 𝒂𝟏 𝟐 √𝒂𝟐𝟏 − 𝟒𝒂𝟐
𝒓𝟏 × 𝒓𝟐 = × ⇒ (− ) −
𝟐 𝟐 𝟐 𝟐
( ) ( ) ( )
𝒂𝟐𝟏 𝒂𝟐𝟏 − 𝟒𝒂𝟐 𝒂𝟐𝟏 − 𝒂𝟐𝟏 + 𝟒𝒂𝟐 𝟒𝒂𝟐
⇒ − ⇒ ⇒ ⇒ 𝒂𝟐
𝟒 𝟒 𝟒 𝟒
TOPIC 127: SECOND-ORDER DIFFERENTIAL EQUATION: DISTINCT REAL ROOTS
Characteristic Equation of second order differential equation is (𝒓𝟏 , 𝒓𝟐 ):

−𝒂𝟏 ± √𝒂𝟐𝟏 − 𝟒𝒂𝟐
𝒓=
𝟐
−𝒂𝟏 +√𝒂𝟐𝟏 −𝟒𝒂𝟐 −𝒂𝟏 −√𝒂𝟐𝟏 −𝟒𝒂𝟐
𝒓𝟏 = 𝟐
and 𝒓𝟐 = 𝟐
Nature of characteristic roots depends on value of discriminant (𝒂𝟐𝟏 − 𝟒𝒂𝟐 ):
It can have three possibilities: (𝒂𝟐𝟏 − 𝟒𝒂𝟐 ) ⋛ 𝟎
Case – I Case – II Case – III
𝟐 𝟐 𝟐
(𝒂𝟏 − 𝟒𝒂𝟐 ) > 𝟎 (𝒂𝟏 − 𝟒𝒂𝟐 ) = 𝟎 (𝒂𝟏 − 𝟒𝒂𝟐 ) < 𝟎
𝟐 𝟐
𝒂𝟏 > 𝟒𝒂𝟐 𝒂𝟏 = 𝟒𝒂𝟐 𝒂𝟐𝟏 < 𝟒𝒂𝟐
Distinct Real Roots Repeated Real Roots Complex Roots
Here, we consider the Distinct Real Roots (𝒂𝟐𝟏 − 𝟒𝒂𝟐 ) > 𝟎:
−𝒂𝟏 ± √𝒂𝟐𝟏 − 𝟒𝒂𝟐

𝒓=
𝟐
𝑎𝑑𝑑𝑒𝑑
−𝒂𝟏 ⏞ ±
⏟ √𝒂𝟐𝟏 − 𝟒𝒂𝟐
𝑠𝑢𝑏𝑡𝑟𝑎𝑐𝑡𝑒𝑑
𝒓𝟏 , 𝒓𝟐 = ⇒ 𝒓𝟏 ≠ 𝒓𝟐
𝟐
The complementary function in case of Distinct Real Roots shall have two and distinct
exponential terms:
𝒚𝒄 = 𝒚𝟏 + 𝒚𝟐
𝒚𝒄 = 𝑨𝟏 𝒆𝒓𝟏𝒕 + 𝑨𝟐 𝒆𝒓𝟐𝒕 (𝒓𝟏 ≠ 𝒓𝟐 )
Numerical Example: Solve the differential equation: 𝒚′′ (𝒕) + 𝒚′ (𝒕) − 𝟐𝒚 = −𝟏𝟎, if 𝒚(𝟎) = 𝟏𝟐 and
𝒚′ (𝟎) = −𝟐.
Comparing with the standard of second order differential equation:
𝒚′′ (𝒕) + 𝒂𝟏 𝒚′ (𝒕) + 𝒂𝟐 𝒚 = 𝒃
′′ (𝒕)
𝒚 ⏟ 𝒚′ (𝒕) + (−𝟐)
+ (𝟏) ⏟ 𝒚 = −𝟏𝟎⏟
𝒂𝟏 𝒂𝟐 𝒃
Nature of Roots: (𝒂𝟐𝟏 − 𝟒𝒂𝟐 ) ⋛ 𝟎
______________________________________________________________________________
{(𝟏)𝟐 − 𝟒(−𝟐)} ⋛ 𝟎
𝟗
⏟> 𝟎
𝐷𝑖𝑠𝑡𝑖𝑛𝑐𝑡
𝑅𝑒𝑎𝑙 𝑅𝑜𝑜𝑡𝑠
𝟗
⏟> 𝟎
Calculation of Roots:
−𝒂𝟏 ± √𝒂𝟐𝟏 − 𝟒𝒂𝟐
𝒓𝟏 , 𝒓𝟐 =
𝟐
−(𝟏) ± √(𝟏)𝟐 − 𝟒(−𝟐)
𝒓𝟏 , 𝒓𝟐 =
𝟐
−𝟏 ± √𝟗
𝒓𝟏 , 𝒓𝟐 =
𝟐
−𝟏 + 𝟑 −𝟏 − 𝟑
𝒓𝟏 = , 𝒓𝟐 =
𝟐 𝟐
𝒓𝟏 = 𝟏 or 𝒓𝟐 = −𝟐
These roots can be substituted in the formula of Complementary Function (𝒚𝒄 ):
Complementary Function: 𝒚𝒄 = 𝑨𝟏 𝒆 𝒓𝟏 𝒕 + 𝑨𝟐 𝒆 𝒓𝟐 𝒕
𝒚𝒄 = 𝑨𝟏 𝒆 𝟏 𝒕 + 𝑨𝟐 𝒆 −𝟐 𝒕
𝒚𝒄 = 𝑨𝟏 𝒆𝒕 + 𝑨𝟐 𝒆−𝟐𝒕
General Solution: 𝒚(𝒕) = 𝒚𝒄 + 𝒚𝒑
𝑨𝟏 𝒆𝒕 + 𝑨𝟐 𝒆−𝟐𝒕 + ⏟
𝒚(𝒕) = ⏟ 𝟓
𝒚𝒄 𝒚𝒑
Definite Solution: For definite solution, we substitute 𝒕 = 𝟎 in the General Solution; 𝒚(𝒕) =
𝑨𝟏 𝒆𝒕 + 𝑨𝟐 𝒆−𝟐𝒕 + 𝟓:
(𝟎)
⏟ = 𝑨𝟏 𝒆 + 𝑨𝟐 𝒆−𝟐(𝟎) + 𝟓
𝒚(𝟎)
𝐺𝑖𝑣𝑒𝑛
⏟ = 𝑨𝟏 (𝟏) + 𝑨𝟐 (𝟏) + 𝟓
𝒚(𝟎)
12
𝟏𝟐 = 𝑨𝟏 + 𝑨𝟐 + 𝟓
𝟕 = 𝑨𝟏 + 𝑨 𝟐
As there are two unknowns 𝑨𝟏 and 𝑨𝟐 , we need two (simultaneous) equations to evaluate them.
Second equation can be the derivative of 𝒚(𝒕), as 𝒚′ (𝟎) is given.
𝒚(𝒕) = 𝑨𝟏 𝒆𝒕 + 𝑨𝟐 𝒆−𝟐𝒕 + 𝟓
′ (𝒕)
𝒚 = 𝑨𝟏 𝒆𝒕 + 𝑨𝟐 (−𝟐)𝒆−𝟐𝒕 + 𝟎
′ (𝒕)
𝒚 = 𝑨𝟏 𝒆𝒕 − 𝟐𝑨𝟐 𝒆−𝟐𝒕
Introducing the initial condition 𝒕 = 𝟎 in 𝒚′(𝒕) = 𝑨𝟏 𝒆𝒕 − 𝟐𝑨𝟐 𝒆−𝟐𝒕 :
𝒚′ (𝟎) = 𝑨𝟏 𝒆(𝟎) − 𝟐𝑨𝟐 𝒆−𝟐(𝟎)
⏟
′ (𝟎)
𝒚
⏟ = 𝑨𝟏 (𝟏) − 𝟐𝑨𝟐 (𝟏)
(−2)
−𝟐 = 𝑨𝟏 − 𝟐𝑨𝟐
Solving simultaneously: 𝟕 = 𝑨𝟏 + 𝑨𝟐 and −𝟐 = 𝑨𝟏 − 𝟐𝑨𝟐
𝟕 = 𝑨𝟏 + 𝑨𝟐 ⇒ 𝟕 − 𝑨𝟐 = 𝑨𝟏
−𝟐 = 𝑨𝟏 − 𝟐𝑨𝟐 ⇒ −𝟐 = 𝟕 − 𝑨𝟐 − 𝟐𝑨𝟐
______________________________________________________________________________
𝑨𝟐 + 𝟐𝑨𝟐 = 𝟕 + 𝟐
𝟑𝑨𝟐 = 𝟗
𝑨𝟐 = 𝟑
For evaluating 𝑨𝟏 , put 𝑨𝟐 = 𝟑 in 𝟕 − 𝑨𝟐 = 𝑨𝟏
𝑨𝟏 = 𝟕 − 𝑨 𝟐
𝑨𝟏 = 𝟕 − 𝟑
𝑨𝟏 = 𝟒
Returning value of 𝑨𝟏 and 𝑨𝟐 in General Solution 𝒚(𝒕) = 𝑨𝟏 𝒆𝒕 + 𝑨𝟐 𝒆−𝟐𝒕 + 𝟓 to get the Definite
Solution:
𝒚(𝒕) = 𝟒 𝒆𝒕 + 𝟑 𝒆−𝟐𝒕 + 𝟓
Interpretation: Definite Solution of Second Order Differential Equation
′′ (𝒕) ′ (𝒕) 𝒕 −𝟐𝒕
𝒚 +𝒚 − 𝟐𝒚 = −𝟏𝟎 with case of Distinct Real Roots is 𝒚(𝒕) = 𝟒𝒆 + 𝟑𝒆 +𝟓.
______________________________________________________________________________
Lesson 29
SECOND ORDER DIFFERENTIAL EQUATIONS (CONTINUED 1)
TOPIC 128: SECOND-ORDER DIFFERENTIAL EQUATION: REPEATED REAL ROOTS

−𝒂𝟏 ± √𝒂𝟐𝟏 − 𝟒𝒂𝟐
𝒓=
𝟐
−𝒂𝟏 +√𝒂𝟐𝟏 −𝟒𝒂𝟐 −𝒂𝟏 −√𝒂𝟐𝟏 −𝟒𝒂𝟐
𝒓𝟏 = 𝟐
and 𝒓𝟐 = 𝟐
𝟐 𝟐 𝟐
(𝒂𝟏 − 𝟒𝒂𝟐 ) > 𝟎 (𝒂𝟏 − 𝟒𝒂𝟐 ) = 𝟎 (𝒂𝟏 − 𝟒𝒂𝟐 ) < 𝟎
𝒂𝟐𝟏 > 𝟒𝒂𝟐 𝒂𝟐𝟏 = 𝟒𝒂𝟐 𝒂𝟐𝟏 < 𝟒𝒂𝟐
Here, we consider the Repeated Real Roots (𝒂𝟐𝟏 − 𝟒𝒂𝟐 ) = 𝟎:

−𝒂𝟏 ± √𝟎
𝒓=
𝟐
−𝒂𝟏
𝒓𝟏 , 𝒓𝟐 = ⇒ 𝒓𝟏 = 𝒓𝟐 = 𝒓
𝟐
The complementary function in case of Repeated Real Roots shall have two but Repeated
exponential terms:
𝒚𝒄 = 𝑨𝟑 𝒆𝒓𝟏𝒕 + 𝑨𝟒 𝒕𝒆𝒓𝟐𝒕
Numerical Example: Solve the differential equation: 𝒚′′ (𝒕) + 𝟔𝒚′ (𝒕) + 𝟗𝒚 = 𝟐𝟕, if 𝒚(𝟎) = 𝟓 and
𝒚′ (𝟎) = −𝟓.
𝒚′′ (𝒕) + 𝒂𝟏 𝒚′ (𝒕) + 𝒂𝟐 𝒚 = 𝒃
′′ (𝒕)
𝒚 ⏟ 𝒚′ (𝒕) + (𝟗)
+ (𝟔) ⏟ 𝒚 = 𝟐𝟕
⏟
______________________________________________________________________________
{(𝟔)𝟐 − 𝟒(𝟗)} ⋛ 𝟎
𝟑𝟔 − 𝟑𝟔 = 𝟎
𝟑𝟔 − 𝟑𝟔 = 𝟎
𝟎
⏟> 𝟎
𝑅𝑒𝑝𝑒𝑎𝑡𝑒𝑑
−𝒂𝟏 ± √𝒂𝟐𝟏 − 𝟒𝒂𝟐
𝒓𝟏 , 𝒓𝟐 =
𝟐
−(𝟔) ± √(𝟔)𝟐 − 𝟒(𝟗)
𝒓𝟏 , 𝒓𝟐 =
𝟐
−𝟔 ± √𝟎
𝒓𝟏 , 𝒓𝟐 =
𝟐
−𝟔 −𝟔
𝒓𝟏 = , 𝒓𝟐 =
𝟐 𝟐
𝒓𝟏 = −𝟑 and 𝒓𝟐 = −𝟑
𝒓 = −𝟑
Complementary Function: 𝒚𝒄 = 𝑨𝟑 𝒆 𝒓 𝒕 + 𝑨𝟒 𝒕𝒆 𝒓 𝒕
𝒚𝒄 = 𝑨𝟑 𝒆 −𝟑 𝒕 + 𝑨𝟒 𝒕𝒆 −𝟑 𝒕
𝒚𝒄 = 𝑨𝟑 𝒆−𝟑𝒕 + 𝑨𝟒 𝒕𝒆−𝟑𝒕
𝑨𝟑 𝒆−𝟑𝒕 + 𝑨𝟒 𝒕𝒆−𝟑𝒕 + ⏟
𝒚(𝒕) = ⏟ 𝟑
𝒚𝒄 𝒚𝒑
Definite Solution: For definite solution, we substitute 𝒕 = 𝟎 in the General Solution; 𝒚(𝒕) =
𝑨𝟑 𝒆−𝟑𝒕 + 𝑨𝟒 𝒕𝒆−𝟑𝒕 + 𝟑:
⏟ = 𝑨𝟑 𝒆−𝟑(𝟎) + 𝑨𝟒 (𝟎)𝒆−𝟑(𝟎) + 𝟑
𝒚(𝟎)
⏟ = 𝑨𝟑 (𝟏) + 𝟎 + 𝟑
𝒚(𝟎)
5
𝟓 = 𝑨𝟑 + 𝟑
𝑨𝟑 = 𝟐
There is another unknown constant (𝑨𝟒 ). Value of 𝒚′ (𝟎) is given, so take derivative of 𝒚(𝒕) =
𝑨𝟑 𝒆−𝟑𝒕 + 𝑨𝟒 𝒕𝒆−𝟑𝒕 + 𝟑:
𝒚(𝒕) = 𝑨𝟑 𝒆−𝟑𝒕 + 𝑨𝟒 𝒕𝒆−𝟑𝒕 + 𝟑
𝒚′ (𝒕) = 𝑨𝟑 (−𝟑)𝒆−𝟑𝒕 + 𝑨𝟒 {(𝟏)𝒆−𝟑𝒕 + 𝒕(−𝟑)𝒆−𝟑𝒕 } + 𝟎
𝒚′ (𝒕) = −𝟑 𝑨 ⏟𝟑 𝒆−𝟑𝒕 + 𝑨𝟒 (𝒆−𝟑𝒕 − 𝟑𝒕𝒆−𝟑𝒕 )
𝑔𝑖𝑣𝑒𝑛
𝒚′ (𝒕) = −𝟑 𝑨
⏟𝟑 𝒆−𝟑𝒕 + 𝑨𝟒 𝒆−𝟑𝒕 − 𝟑𝑨𝟒 𝒕𝒆−𝟑𝒕
𝟐
𝒚′ (𝒕) = −𝟑(𝟐)𝒆−𝟑𝒕 + 𝑨𝟒 𝒆−𝟑𝒕 − 𝟑𝑨𝟒 𝒕𝒆−𝟑𝒕
Introducing initial condition 𝒕 = 𝟎 in 𝒚′ (𝒕) = −𝟑(𝟐)𝒆−𝟑𝒕 + 𝑨𝟒 𝒆−𝟑𝒕 − 𝟑𝑨𝟒 𝒕𝒆−𝟑𝒕 :
𝒚′ (𝟎) = −𝟑(𝟐)(𝟏) + 𝑨𝟒 (𝟏) − 𝟎
⏟
𝒚′ (𝟎) = −𝟔 + 𝑨𝟒
⏟
−5
−𝟓 = −𝟔 + 𝑨𝟒
______________________________________________________________________________
𝑨𝟒 = 𝟏
Returning value of 𝑨𝟑 and 𝑨𝟒 in General Solution 𝒚(𝒕) = 𝑨𝟑 𝒆−𝟑𝒕 + 𝑨𝟒 𝒕𝒆−𝟑𝒕 + 𝟑 to get the
Definite Solution:
𝒚(𝒕) = 𝟐 𝒆−𝟑𝒕 + 𝟏 𝒕𝒆−𝟑𝒕 + 𝟑
𝒚(𝒕) = 𝟐𝒆−𝟑𝒕 + 𝒕𝒆−𝟑𝒕 + 𝟑
′′ (𝒕) ′ (𝒕) −𝟑𝒕 −𝟑𝒕
𝒚 + 𝟔𝒚 + 𝟗𝒚 = 𝟐𝟕 with case of Repeated Real Roots is 𝒚(𝒕) = 𝟐𝒆 + 𝒕𝒆 +𝟑.
TOPIC 129: SECOND-ORDER DIFFERENTIAL EQUATION: COMPLEX ROOTS

−𝒂𝟏 ± √𝒂𝟐𝟏 − 𝟒𝒂𝟐
𝒓=
𝟐
−𝒂𝟏 +√𝒂𝟐𝟏 −𝟒𝒂𝟐 −𝒂𝟏 −√𝒂𝟐𝟏 −𝟒𝒂𝟐
𝒓𝟏 = and 𝒓𝟐 =
𝟐 𝟐

(𝒂𝟐𝟏 − 𝟒𝒂𝟐 ) > 𝟎 (𝒂𝟐𝟏 − 𝟒𝒂𝟐 ) = 𝟎 (𝒂𝟐𝟏 − 𝟒𝒂𝟐 ) < 𝟎
𝒂𝟐𝟏 > 𝟒𝒂𝟐 𝒂𝟐𝟏 = 𝟒𝒂𝟐 𝒂𝟐𝟏 < 𝟒𝒂𝟐
Here, we consider the Complex Roots (𝒂𝟐𝟏 − 𝟒𝒂𝟐 ) < 𝟎:
−𝒂𝟏 ± √𝒂𝟐𝟏 − 𝟒𝒂𝟐

𝒓𝟏 , 𝒓𝟐 =
𝟐
−𝒂𝟏 ± √(𝟒𝒂𝟐 − 𝒂𝟐𝟏 ) × (−𝟏)
𝒓𝟏 , 𝒓𝟐 =
𝟐
−𝒂𝟏 ± √(𝟒𝒂𝟐 − 𝒂𝟐𝟏 ) × √(−𝟏)
𝒓𝟏 , 𝒓𝟐 =
𝟐
−𝒂𝟏 ± √(𝟒𝒂𝟐 − 𝒂𝟐𝟏 ) × 𝒊
𝒓𝟏 , 𝒓𝟐 =
𝟐
𝟐
−𝒂𝟏 √𝟒𝒂𝟐 − 𝒂𝟏
𝒓𝟏 , 𝒓𝟐 = ± ×𝒊
𝟐 𝟐
√𝟒𝒂𝟐 −𝒂𝟐𝟏 √𝒂𝟐𝟏 −𝟒𝒂𝟐
−𝒂𝟏 −𝒂𝟏
𝒓𝟏 = 𝟐
+ 𝟐
× 𝒊 and 𝒓𝟐 = 𝟐
− 𝟐
×𝒊
Using shorthand notations 𝒉 and 𝝊.
𝒓𝟏 = 𝒉 + 𝝊𝒊 and 𝒓𝟐 = 𝒉 − 𝝊𝒊
In the form of Conjugate Complex Roots:
______________________________________________________________________________
𝒓𝟏 , 𝒓𝟐 = 𝒉
⏟ ± 𝝊𝒊
⏟
𝑟𝑒𝑎𝑙 𝑖𝑚𝑎𝑔𝑖𝑛𝑎𝑟𝑦
𝑝𝑎𝑟𝑡 𝑝𝑎𝑟𝑡
Borrowing these roots to put in the Complementary Function:
𝒚𝒄 = 𝒚𝟏 + 𝒚𝟐
𝒚𝒄 = 𝑨𝟓 𝒆𝒓𝟏𝒕 + 𝑨𝟔 𝒆𝒓𝟐𝒕
𝒚𝒄 = 𝑨𝟓 𝒆(𝒉+𝝊𝒊)𝒕 + 𝑨𝟔 𝒆(𝒉−𝝊𝒊)𝒕
𝒚𝒄 = 𝑨𝟓 𝒆𝒉𝒕+𝝊𝒊𝒕 + 𝑨𝟔 𝒆𝒉𝒕−𝝊𝒊𝒕
𝒚𝒄 = 𝑨𝟓 𝒆𝒉𝒕 𝒆𝝊𝒊𝒕 + 𝑨𝟔 𝒆𝒉𝒕 𝒆−𝝊𝒊𝒕
𝒚𝒄 = 𝒆𝒉𝒕 (𝑨𝟓 𝒆𝝊𝒊𝒕 + 𝑨𝟔 𝒆−𝝊𝒊𝒕 )
The complementary function in case of Complex Roots shall have two but Complex exponential
terms:
𝒚𝒄 = 𝒆𝒉𝒕 (𝑨𝟓 𝒆𝝊𝒊𝒕 + 𝑨𝟔 𝒆−𝝊𝒊𝒕 )
Involvement of imaginary component (𝒊) in the expression requires attention. An alternative form
in terms of trigonometric ratios is:
𝒚𝒄 = 𝒆𝒉𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝝊𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝝊𝒕 )
______________________________________________________________________________
Argand Diagram: Graphical Representation of a Complex number

𝑹𝟐 = 𝒉𝟐 + 𝒗𝟐 (Pythagorean Theorem)
𝑹 = √𝒉𝟐 + 𝒗𝟐
Numerical Example: Solve the differential equation: 𝒚′′ (𝒕) + 𝟐𝒚′ (𝒕) + 𝟏𝟕𝒚 = 𝟑𝟒, if 𝒚(𝟎) = 𝟑 and
𝒚′ (𝟎) = 𝟏𝟏.
𝒚′′ (𝒕) + 𝒂𝟏 𝒚′ (𝒕) + 𝒂𝟐 𝒚 = 𝒃
′′ (𝒕)
𝒚 ⏟ 𝒚′ (𝒕) + (𝟏𝟕)
+ (𝟐) ⏟ 𝒚 = 𝟑𝟒 ⏟
Here, 𝒂𝟏 = 𝟐, 𝒂𝟐 = 𝟏𝟕, 𝒃 = 𝟑𝟒
______________________________________________________________________________
−𝟔𝟒 ⋛ 𝟎
−𝟔𝟒 < 𝟎
⏟
𝐶𝑜𝑚𝑝𝑙𝑒𝑥
𝑅𝑜𝑜𝑡𝑠
−𝟔𝟒 < 𝟎
⏟
−𝒂𝟏 ± √𝒂𝟐𝟏 − 𝟒𝒂𝟐
𝒓𝟏 , 𝒓𝟐 =
𝟐
−(𝟐) ± √(𝟐)𝟐 − 𝟒(𝟏𝟕)
𝒓𝟏 , 𝒓𝟐 =
𝟐
−𝟐 ± √−𝟔𝟒
𝒓𝟏 , 𝒓𝟐 =
𝟐
−𝟐 ± √𝟔𝟒 × √−𝟏
𝒓𝟏 , 𝒓𝟐 =
𝟐
−𝟐 ± 𝟖 𝒊
𝒓𝟏 , 𝒓𝟐 =
𝟐
−𝟐 + 𝟖 𝒊 −𝟐 − 𝟖 𝒊
𝒓𝟏 = , 𝒓𝟐 =
𝟐 𝟐
𝟐 𝟖 𝟐 𝟖
𝒓𝟏 = − + 𝒊, 𝒓𝟐 = − − 𝒊
𝟐 𝟐 𝟐 𝟐
𝒓𝟏 = −𝟏 ⏟ +⏟ 𝟒 𝒊 , 𝒓𝟐 = −𝟏
⏟ −⏟ 𝟒𝒊
ℎ 𝝊 ℎ 𝝊
From the roots, we extract: 𝒉 = −𝟏 and 𝝊 = 𝟒.
Substituting in the formula of Complementary Function 𝒚𝒄 = 𝒆 𝒉 𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝝊 𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝝊 𝒕):
Complementary Function: 𝒚𝒄 = 𝒆 𝒉 𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝝊 𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝝊 𝒕 )
𝒚𝒄 = 𝒆 −𝟏 𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝟒 𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝟒 𝒕)
𝒚𝒄 = 𝒆−𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝟒𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝟒𝒕)
After developing complementary function, we get general solution by adding particular integral
𝒚𝒑 = 𝟐.
𝒚(𝒕) = 𝒆−𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝟒𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝟒𝒕) + 𝟐
After attaining the General Solution, we get Definite Solution by definitizing the arbitrary
constants 𝑨𝟓 and 𝑨𝟔 .
Introducing initial condition; 𝒕 = 𝟎 in the General Solution:
𝒚(𝟎) = 𝒆−(𝟎) {𝑨𝟓 𝐜𝐨𝐬 𝟒(𝟎) + 𝑨𝟔 𝐬𝐢𝐧 𝟒(𝟎)} + 𝟐
⏟ = 𝟏 × (𝑨𝟓 𝐜𝐨𝐬 𝟎 + 𝑨𝟔 𝐬𝐢𝐧 𝟎) + 𝟐
𝒚(𝟎)
⏟ = {𝑨𝟓 (𝟏) + 𝑨𝟔 (𝟎)} + 𝟐
𝒚(𝟎)
𝟑
𝟑 = (𝑨𝟓 ) + 𝟐
𝑨𝟓 = 𝟏
′ (𝟎),
To substitute 𝒚 we take the derivative of general solution 𝒚(𝒕):
𝒅 𝒅
{𝒚(𝒕)} = {𝒆−𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝟒𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝟒𝒕) + 𝟐}
𝒅𝒕 𝒅𝒕
______________________________________________________________________________
𝒅 −𝒕 𝒅
𝒚′ (𝒕) = {𝒆⏟ ⏟ (𝑨𝟓 𝐜𝐨𝐬 𝟒𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝟒𝒕)} + (𝟐)
𝒅𝒕 𝑔(𝑡) 𝒅𝒕
ℎ(𝑡)
𝒅 𝒅 𝒅
Product Rule of differentiation: {𝒇(𝒕)} = 𝒅𝒕 {𝒈(𝒕)} × 𝒉(𝒕) + 𝒅𝒕 {𝒉(𝒕)} × 𝒈(𝒕)
𝒅𝒕
𝒅 𝒅
𝒚′ (𝒕) = 𝒅𝒕 {(𝒆−𝒕 )} × (𝑨𝟓 𝐜𝐨𝐬 𝟒𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝟒𝒕) + 𝒅𝒕 {(𝑨𝟓 𝐜𝐨𝐬 𝟒𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝟒𝒕)} × (𝒆−𝒕 )
𝒅 𝒅
𝒚′ (𝒕) = (−𝟏)(𝒆−𝒕 ) × (𝑨𝟓 𝐜𝐨𝐬 𝟒𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝟒𝒕) + {𝒅𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝟒𝒕) + 𝒅𝒕 (𝑨𝟔 𝐬𝐢𝐧 𝟒𝒕)} × (𝒆−𝒕 )
𝒅 𝒅
𝒚′ (𝒕) = ( − 𝒆−𝒕 ) × (𝑨𝟓 𝐜𝐨𝐬 𝟒𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝟒𝒕) + {𝑨𝟓 (𝐜𝐨𝐬 𝟒𝒕) + 𝑨𝟔 (𝐬𝐢𝐧 𝟒𝒕)} × (𝒆−𝒕 )
𝒅𝒕 𝒅𝒕
′ (𝒕)
𝒚 = (𝒆 ) × ( − 𝑨𝟓 𝐜𝐨𝐬 𝟒𝒕 − 𝑨𝟔 𝐬𝐢𝐧 𝟒𝒕) + {𝑨𝟓 (−𝟒)(𝐬𝐢𝐧 𝟒𝒕) + 𝑨𝟔 𝟒(𝐜𝐨𝐬 𝟒𝒕)} × (𝒆−𝒕 )
−𝒕
𝒚′ (𝒕) = (𝒆−𝒕 ) × (−𝑨𝟓 𝐜𝐨𝐬 𝟒𝒕 − 𝑨𝟔 𝐬𝐢𝐧 𝟒𝒕) + {−𝟒𝑨𝟓 (𝐬𝐢𝐧 𝟒𝒕) + 𝟒𝑨𝟔 (𝐜𝐨𝐬 𝟒𝒕)} × (𝒆−𝒕 )
Introducing the initial condition (𝒕 = 𝟎) in this derivative:
𝒚′ (𝟎) = {𝒆−(𝟎) } × {−𝑨𝟓 𝐜𝐨𝐬 𝟒(𝟎) − 𝑨𝟔 𝐬𝐢𝐧 𝟒(𝟎)} + {−𝟒𝑨𝟓 𝐬𝐢𝐧 𝟒(𝟎) + 𝟒𝑨𝟔 𝐜𝐨𝐬 𝟒(𝟎)} × {𝒆−(𝟎) }
𝒚′ (𝟎) = (𝟏) × (−𝑨𝟓 𝐜𝐨𝐬 𝟎 − 𝑨𝟔 𝐬𝐢𝐧 𝟎) + {−𝟒𝑨𝟓 (𝐬𝐢𝐧 𝟎) + 𝟒𝑨𝟔 (𝐜𝐨𝐬 𝟎)} × (𝟏)
⏟
⏟′ (𝟎) = {−𝑨𝟓 (𝟏) − 𝑨𝟔 (𝟎)} − 𝟒𝑨𝟓 (𝟎) + 𝟒𝑨𝟔 (𝟏)

𝒚
11
𝟏𝟏 = −𝑨𝟓 − 𝟎 − 𝟎 + 𝟒𝑨𝟔
𝟏𝟏 = − 𝑨 ⏟𝟓 + 𝟒𝑨𝟔
𝑎𝑙𝑟𝑒𝑎𝑑𝑦
𝑓𝑜𝑢𝑛𝑑
𝟏𝟏 = − 𝑨
⏟𝟓 + 𝟒𝑨𝟔
1
𝟏𝟏 = −𝟏 + 𝟒𝑨𝟔
𝟏𝟐 = 𝟒𝑨𝟔
𝑨𝟔 = 𝟑
Returning the values of 𝑨𝟓 and 𝑨𝟔 in the general solution 𝒚(𝒕) = 𝒆−𝒕 ( 𝑨𝟓 𝐜𝐨𝐬 𝟒𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝟒𝒕) +
𝟐:
𝒚(𝒕) = 𝒆−𝒕 ( 𝟏 𝐜𝐨𝐬 𝟒𝒕 + 𝟑 𝐬𝐢𝐧 𝟒𝒕) + 𝟐
𝒚(𝒕) = ⏟𝒆−𝒕 (𝐜𝐨𝐬 𝟒𝒕 + 𝟑 𝐬𝐢𝐧 𝟒𝒕) + ⏟
𝟐
𝒚𝒄 𝒚𝒑
′′ (𝒕) ′ (𝒕) −𝒕 (𝐜𝐨𝐬
𝒚 + 𝟐𝒚 + 𝟏𝟕𝒚 = 𝟑𝟒 with case of Complex Roots is 𝒚(𝒕) = 𝒆 𝟒𝒕 + 𝟑 𝐬𝐢𝐧 𝟒𝒕) + 𝟐 .
TOPIC 130: SECOND-ORDER DIFFERENTIAL EQUATION: SUMMARY OF

COMPLEMENTARY FUNCTION
As we have developed the characteristic roots before:

−𝒂𝟏 ± √𝒂𝟐𝟏 − 𝟒𝒂𝟐
𝒓 = 𝒓𝟏 , 𝒓𝟐 =
𝟐
Discriminant (𝒂𝟐𝟏 − 𝟒𝒂𝟐 ) determines the nature of roots.
Possible Cases for Complementary Solution (𝒚𝒄 ) of (SODE)
Condition 𝒂𝟐𝟏 − 𝟒𝒂𝟐 > 𝟎, 𝒂𝟐𝟏 > 𝟒𝒂𝟐 𝒂𝟐𝟏 − 𝟒𝒂𝟐 = 𝟎, 𝒂𝟐𝟏 = 𝟒𝒂𝟐 𝒂𝟐𝟏 − 𝟒𝒂𝟐 < 𝟎, 𝒂𝟐𝟏 <
𝟒𝒂𝟐
Remarks +𝑣𝑒 Discriminant. 𝑧𝑒𝑟𝑜 Discriminant. −𝑣𝑒 Discriminant.
______________________________________________________________________________
Distinct Real Roots Repeated/Equal Real Complex Roots

Distinct: Roots Complex:
±(+𝑣𝑒 Discriminant) Repeated/Equal: ±(−𝑣𝑒 Discriminant)
⇒ 𝒓𝟏 ≠ 𝒓𝟐 ±(𝑧𝑒𝑟𝑜 Discriminant) 𝒓𝟏 = 𝒓𝟐
⇒ 𝒓𝟏 = 𝒓𝟐 ⇒ 𝒓
−𝒂
Root(s)’ −𝒂𝟏 ±√𝒂𝟐𝟏 −𝟒𝒂𝟐 𝒓𝟏 = 𝒓𝟐 = 𝒓 = 𝟐 𝟏 (Conjugate Roots)
Formula 𝒓𝟏 , 𝒓𝟐 = 𝒓𝟏 , 𝒓𝟐 = 𝒉 ± 𝒗𝒊
𝟐 −𝒂
Where 𝒉 = 𝟐 𝟏
√𝟒𝒂𝟐 −𝒂𝟐𝟏
𝒗= 𝟐
Verification 𝒓𝟏 + 𝒓𝟐 = −𝒂𝟏 𝒓𝟏 + 𝒓𝟐 = −𝒂𝟏 𝒓𝟏 + 𝒓𝟐 = −𝒂𝟏
𝒓𝟏 × 𝒓𝟐 = 𝒂𝟐 𝒓𝟏 × 𝒓𝟐 = 𝒂𝟐 𝒓𝟏 × 𝒓𝟐 = 𝒂𝟐
(𝒚𝒄 ) 𝒚𝒄 = 𝑨𝟏 𝒆𝒓𝟏𝒕 + 𝑨𝟐 𝒆𝒓𝟐𝒕 𝒚𝒄 = 𝑨𝟑 𝒆𝒓𝒕 + 𝑨𝟒 𝒕𝒆𝒓𝒕 𝒚𝒄 = 𝑨𝟓 𝒆(𝒉+𝒗𝒊)𝒕 +
Formula 𝑨𝟔 𝒆(𝒉−𝒗𝒊)𝒕
𝒚𝒄 = 𝒆𝒉𝒕 (𝑨𝟓 𝒆𝒗𝒊𝒕 +
𝑨𝟔 𝒆−𝒗𝒊𝒕 )
𝒚𝒄
= 𝒆𝒉𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝒗𝒕
+ 𝑨𝟔 𝐬𝐢𝐧 𝒗𝒕)
Dynamic 𝒓𝟏 & 𝒓𝟐 < 𝟎 𝒓<𝟎 𝒉<𝟎
Stability
TOPIC 131: SECOND-ORDER DIFFERENTIAL EQUATION: DYNAMIC STABILITY OF

DISTINCT REAL ROOTS
Recalling the under Distinct Real Roots case:

⏟𝟏 𝒆𝒓𝟏𝒕 + 𝑨𝟐 𝒆𝒓𝟐 𝒕 +
𝒚(𝒕) = 𝑨 𝒚
⏟𝒑
𝑦𝑐 𝑏 𝑏 𝑏 2
, 𝑡, 𝑡
𝑎2 𝑎1 2
𝑨𝟏 𝒆𝒓𝟏𝒕 + 𝑨𝟐 𝒆𝒓𝟐𝒕 +
𝒚(𝒕) = ⏟ 𝒚
⏟𝒑
𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑓𝑟𝑜𝑚 𝑒𝑞𝑢𝑖𝑙𝑖𝑏𝑟𝑖𝑢𝑚
𝑒𝑞𝑢𝑖𝑙𝑖𝑏𝑟𝑖𝑢𝑚
______________________________________________________________________________
𝑇𝑟𝑦 𝑡𝑜 𝑚𝑖𝑡𝑖𝑔𝑎𝑡𝑒 𝑖𝑡
⏞
⏟𝟏 𝒆𝒓𝟏𝒕 + 𝑨𝟐 𝒆𝒓𝟐𝒕
𝑨
𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑓𝑟𝑜𝑚
𝑎𝑟𝑏𝑖𝑡𝑟𝑎𝑟𝑦 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 𝑎𝑟𝑏𝑖𝑡𝑟𝑎𝑟𝑦 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡
𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒
⏞ 𝒓𝟏 ⏞
𝒕 ⏞ 𝒓𝟐 ⏞
𝒕
⏟ 𝑨𝟏 𝒆 + 𝑨𝟐 𝒆
𝑨𝟏 and 𝑨𝟐 are constant and their values depend on initial conditions. 𝒕 is independent variable
spanning (𝟎 − ∞). Deviation from equilibrium will depend on roots (𝒓𝟏 , 𝒓𝟐 ).
Hair Splitting:
−𝑣𝑒 −𝑣𝑒
⏞𝟏 (𝒕)
𝑨𝟏 𝒆 + 𝑨𝟐 𝒆 𝒓⏞𝟐 (𝒕)
𝒓
𝒓𝟏 & 𝒓𝟐 , determine the dynamic stability.

If both 𝒓𝟏 & 𝒓𝟐 are −𝒗𝒆, then the exponential terms (𝑨𝟏 𝒆𝒓𝟏𝒕 & 𝑨𝟐 𝒆𝒓𝟐𝒕 ) will fade away due to
exponential decay. Automatically their sum also fades away:
⏞𝟏 𝒕
𝒓 ⏞𝟐 𝒕
𝒓
{𝑨𝟏 × 𝒆⏟ + 𝑨𝟐 × 𝒆⏟ } → 𝟎 when 𝒕 → ∞, ∴ 𝒚𝒄 → 𝟎.
𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙 𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙
𝑑𝑒𝑐𝑎𝑦 𝑑𝑒𝑐𝑎𝑦
Caveat: Both 𝒓𝟏 and 𝒓𝟐 should be negative, else convergence may not take place.
______________________________________________________________________________
Numerical Example:
𝒚(𝒕) = 𝟒𝒆𝒕 + 𝟑𝒆𝟐𝒕 + 𝟓 𝒚(𝒕) = 𝟒𝒆𝒕 + 𝟑𝒆−𝟐𝒕 + 𝟓 𝒚(𝒕) = 𝟒𝒆−𝒕 + 𝟑𝒆−𝟐𝒕 + 𝟓
𝟒𝒆𝒕 + 𝟑𝒆𝟐𝒕 𝟒𝒆𝒕 + 𝟑𝒆−𝟐𝒕 𝟒𝒆−𝒕 + 𝟑𝒆−𝟐𝒕
𝑨𝟏 𝒆𝒓𝟏𝒕 𝑨𝟐 𝒆𝒓𝟐𝒕 𝒚 𝑨𝟏 𝒆𝒓𝟏𝒕 𝑨𝟐 𝒆𝒓𝟐 𝒕 𝒚 𝑨𝟏 𝒆𝒓𝟏𝒕 𝑨𝟐 𝒆𝒓𝟐𝒕 𝒚
+𝑣𝑒 +𝑣𝑒 𝒄 +𝑣𝑒 −𝑣𝑒 𝒄 −𝑣𝑒 −𝑣𝑒 𝒄
⏞𝒕
(𝟏) ⏞𝒕
(𝟐) ⏞𝒕
(𝟏) ⏞
(−𝟐)𝒕 ⏞
(−𝟏)𝒕 ⏞
(−𝟐)𝒕
𝒕 𝟒𝒆 𝟑𝒆 𝟒𝒆 𝟑𝒆 𝟒𝒆 𝟑𝒆
0 4 3 7 4 3 7 4 3 7
1.471 0.406
1 10.9 22.2 33.0 10.9 0.4060 11.3 5 0 1.8775
0.541 0.054
2 29.6 163.8 193.4 29.6 0.0549 29.6 3 9 0.5963
1210. 0.199 0.007
3 80.3 3 1290.6 80.3 0.0074 80.3 1 4 0.2066
8942. 0.073 0.001
4 218.4 9 9161.3 218.4 0.0010 218.4 3 0 0.0743
66079 0.027 0.000
5 593.7 .4 66673.1 593.7 0.0001 593.7 0 1 0.0271
1613. 48826 489878. 0.009 0.000
6 7 4.4 1 1613.7 0.0000 1613.7 9 0 0.0099
𝒓 𝒕 𝒓
(𝑨𝟏 𝒆 ) (𝑨𝟐 𝒆 ) (𝒚𝒄 )
𝟏 𝟐 𝒕 (𝑨𝟏 𝒆 )𝒓𝟏 𝒕 𝒓
(𝑨𝟐 𝒆 )𝟐 𝒕 (𝒚𝒄 ) 𝒓 𝟏 𝒕 𝒓 𝟐 𝒕
(𝑨𝟏 𝒆 ) (𝑨𝟐 𝒆 ) (𝒚𝒄 )
⇆
↑ ↑ ↑ ↑ ↓ ↑ ↓ ↓ ↓
Both Deviati 1st exponential term Deviati Both Deviati
exponential on increases, 2nd one on exponential on
terms increas decreases increas terms decreas
increase es es decrease es
Divergence Divergence Convergence
Dynamically Unstable Dynamically Unstable Dynamically Stable
Hence it is numerically confirmed that for Dynamic Stability, BOTH the roots must be
−𝒗𝒆: (𝒓𝟏 ∧ 𝒓𝟐 ) < 𝟎
______________________________________________________________________________
Lesson 30
SECOND ORDER DIFFERENTIAL EQUATIONS (CONTINUED 2)

REPEATED REAL ROOTS
Recalling the under Repeated Real Roots case:

⏟𝟑 𝒆𝒓𝒕 + 𝑨𝟒 𝒕𝒆𝒓𝒕 +
𝒚(𝒕) = 𝑨 𝒚
⏟𝒑
, 𝑡, 𝑡
𝑎2 𝑎1 2
𝑨𝟑 𝒆𝒓𝒕 + 𝑨𝟒 𝒕𝒆𝒓𝒕 +
𝒚(𝒕) = ⏟ 𝒚
⏟𝒑
⏞
𝑨𝟑 𝒆𝒓𝒕 + 𝑨𝟒 𝒕𝒆𝒓𝒕
⏟
𝑎𝑟𝑏𝑖𝑡𝑟𝑎𝑟𝑦 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 𝑎𝑟𝑏𝑖𝑡𝑟𝑎𝑟𝑦 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡
𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒
⏞ 𝒓 ⏞
𝒕 ⏞ ⏞
⏟ 𝑨𝟑 𝒆 + 𝑨𝟒 ⏞
𝒕 𝒆𝒓 𝒕

𝑨𝟑 and 𝑨𝟒 are constant and their values depend on initial conditions. 𝒕 is independent variable
spanning (𝟎 − ∞). Deviation from equilibrium will depend on root (𝒓).
Hair Splitting:
⏞
𝒓 (𝒕) ⏞
𝒓 (𝒕)
𝑨𝟑 𝒆 + 𝑨𝟒 𝒕𝒆
𝒓, determine the dynamic stability.
If 𝒓 is −𝒗𝒆, then the exponential terms (𝑨𝟏 𝒆𝒓𝟏𝒕 & 𝑨𝟐 𝒕𝒆𝒓𝟐𝒕 ) will fade away due to exponential decay.
Though, the second term (𝑨𝟐 𝒕 𝒆𝒓𝟐𝒕 ) has an additional term 𝒕 being multiplied, yet it doesn’t
reverse the dynamics of this exponential term. Eventually, sum of the two exponential terms also
fades away:
𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑐𝑎𝑛𝑑
(0,1,2,…∞)
−𝒗𝒆 ⏞ −𝒗𝒆
⏞
𝒓 𝒕
𝑨𝟑 × 𝒆⏟ + 𝑨𝟒 × ⏟ ⏞
𝒕 ⏞
𝒓 𝒕
𝒆⏟ → 𝟎 when 𝒕 → ∞, ∴ 𝒚𝒄 → 𝟎.
𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙 𝑁𝑒𝑢𝑡𝑟𝑎𝑙 𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙
{ 𝑑𝑒𝑐𝑎𝑦 𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑒𝑟 𝑑𝑒𝑐𝑎𝑦 }
Numerical Example:
𝒚(𝒕) = 𝟐𝒆𝟑𝒕 + 𝒕𝒆𝟑𝒕 + 𝟑 𝒚(𝒕) = 𝟐𝒆−𝟑𝒕 + 𝒕𝒆−𝟑𝒕 + 𝟑
𝟐𝒆𝟑𝒕 + 𝒕𝒆𝟑𝒕 𝟐𝒆−𝟑𝒕 + 𝒕𝒆−𝟑𝒕
𝑨𝟑 𝒆𝒓𝒕 𝑨𝟒 𝒕𝒆𝒓𝒕 𝒚𝒄 𝑨𝟑 𝒆𝒓𝒕 𝑨𝟒 𝒕𝒆𝒓𝒕 𝒚𝒄
+𝑣𝑒 +𝑣𝑒 −𝑣𝑒 −𝑣𝑒
⏞ ⏞ ⏞ ⏞
𝒕 𝟐𝒆 (𝟑) 𝒕 𝒕𝒆 (𝟑) 𝒕 𝟐𝒆(−𝟑)𝒕 𝒕𝒆(−𝟑)𝒕
0 2 0 2 2 0 2
1 40.2 20.1 60.3 0.0996 0.0498 0.149361
2 806.9 806.9 1613.7 0.0050 0.0050 0.009915
3 16206.2 24309.3 40515.4 0.0002 0.0004 0.000617
4 325509.6 651019.2 976528.7 0.00001 0.00002 0.000037
5 6538034.7 16345086.9 22883121.6 0.0000006 0.000002 0.000002
(𝑨𝟑 𝒆𝒓𝒕 )↑ (𝑨𝟒 𝒕𝒆𝒓𝒕 )↑ (𝒚𝒄 )↑ (𝑨𝟑 𝒆𝒓𝒕 )↓ (𝑨𝟒 𝒕𝒆𝒓𝒕 )↓ (𝒚𝒄 )↓
⇆
______________________________________________________________________________
Both exponential terms Deviation Both exponential terms Deviation

increase increases decrease decreases
Divergence Convergence
Dynamically Unstable Dynamically Stable
Hence it is numerically confirmed that for Dynamic Stability, the repeated root (𝒓) must
be −𝒗𝒆: 𝒓 < 𝟎

COMPLEX ROOTS
Recalling the under Complex Roots case:

⏟𝒉𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝝊𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝝊𝒕) +
𝒚(𝒕) = 𝒆 𝒚
⏟𝒑
, 𝑡, 𝑡
𝑎2 𝑎1 2
Hair Splitting:
⏟𝒉𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝝊𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝝊𝒕) +
𝒚(𝒕) = 𝒆 𝒚
⏟𝒑
⏞
⏟𝒉𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝝊𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝝊𝒕)
𝒆
𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑐𝑎𝑛𝑑
⏞𝑎𝑟𝑏𝑖𝑡𝑟𝑎𝑟𝑦 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 𝑎𝑟𝑏𝑖𝑡𝑟𝑎𝑟𝑦 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡
𝒉𝒕
𝒆⏟ ( ⏞𝟓
𝑨 𝐜𝐨𝐬 𝝊 ⏞
𝒕 + ⏞𝟔
𝑨 𝐬𝐢𝐧 𝝊 ⏞
𝒕 )
𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑒𝑟
⏟
⏟
The multiplicand (𝑨𝟓 𝐜𝐨𝐬 𝝊𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝝊𝒕) has the knack of having a wave-like graph as shown
above. This shall have uniform amplitude above and below the equilibrium. So, convergence shall
not take place. However, multiplier 𝒆𝒉𝒕 can mitigate the amplitude of the graph if it’s value declines
with time (𝒕): i.e. 𝒆𝒉𝒕 → 𝟎 as 𝒕 → ∞.
______________________________________________________________________________
Numerical Example:
𝒚(𝒕)
𝒕 (𝐜𝐨𝐬
𝒆 𝟒𝒕 + 𝟑 𝐬𝐢𝐧 𝟒𝒕) + 𝟐 (𝟎)𝒕
𝒆 (𝐜𝐨𝐬 𝟒𝒕 + 𝟑 𝐬𝐢𝐧 𝟒𝒕) 𝒆−𝒕 (𝐜𝐨𝐬 𝟒𝒕 + 𝟑 𝐬𝐢𝐧 𝟒𝒕) + 𝟐
+𝟐
𝒆(𝟏)𝒕 (𝐜𝐨𝐬 𝟒𝒕 𝒆(𝟎)𝒕 (𝐜𝐨𝐬 𝟒𝒕 𝒆(−𝟏)𝒕 (𝐜𝐨𝐬 𝟒𝒕
+ 𝟑 𝐬𝐢𝐧 𝟒𝒕) + 𝟑 𝐬𝐢𝐧 𝟒𝒕) + 𝟑 𝐬𝐢𝐧 𝟒𝒕)
𝒉𝒕
𝒆 (𝑨𝟓 𝐜𝐨𝐬 𝝊𝒕
𝒚𝒄
(𝑨𝟓 𝐜𝐨𝐬 𝝊𝒕
𝒚 𝒆𝒉𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝝊𝒕
𝒚𝒄
+ 𝑨𝟔 𝐬𝐢𝐧 𝝊𝒕) 𝒆𝒉𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝝊𝒕) 𝒄 + 𝑨𝟔 𝐬𝐢𝐧 𝝊𝒕)
+𝑣𝑒 −𝑣𝑒
⏞ (𝐜𝐨𝐬 𝟒𝒕 𝑧𝑒𝑟𝑜 (𝐜𝐨𝐬 𝟒𝒕 ⏞ (𝐜𝐨𝐬 𝟒𝒕
𝒕 𝒆 (𝟏)𝒕 + 𝟑 𝐬𝐢𝐧 𝟒𝒕) 𝒆
⏞ 𝒕
(𝟎)
+ 𝟑 𝐬𝐢𝐧 𝟒𝒕) 𝒆 (𝟎)𝒕 + 𝟑 𝐬𝐢𝐧 𝟒𝒕)
0 1 1 𝟏 1 1 1 1 1 𝟏
2.7 -2.9 −𝟕. 𝟗 0.3678 -2.9 −𝟏. 𝟎𝟕𝟓𝟕
1 1 -2.9 -2.9 8
7.4 2.8 20.9 0.1353 2.8 0.3820
2 1 2.8 2.8 4
20.1 -0.8 -15.4 0.0497 -0.0381
3 1 -0.8 -0.8 9 -0.8
54.6 -1.8 -99.4 0.0183 -0.0334
4 1 -1.8 -1.8 2 -1.8
148.4 3.1 467.0 0.0067 0.0212
5 1 3.1 𝟑. 𝟏 4 3.1
403.4 -2.3 -924.9 0.0024 -0.0057
6 1 -2.3 -2.3 8 -2.3
1096.6 -0.1 -164.4 0.0009 -0.0001
7 1 -0.1 -0.1 1 -0.1
2981.0 2.5 7418.1 0.0003 0.0008
8 1 2.5 2.5 4 2.5
8103.1 -3.1 −𝟐𝟓𝟏𝟒𝟔. 𝟑 0.0001 −𝟎. 𝟎𝟎𝟎𝟒
9 1 -3.1 𝟑. 𝟏 2 -3.1
1 22026. 1.6 𝟑𝟒𝟓𝟒𝟔. 𝟑 0.0000 𝟎. 𝟎𝟎𝟎𝟎𝟕
0 5 1 1.6 1.6 5 1.6
(𝒆𝒉𝒕 )↑ +𝒗𝒆 +𝒗𝒆 +𝒗𝒆 +𝒗𝒆 (𝒆𝒉𝒕 )↓ +𝒗𝒆 +𝒗𝒆
( ) ( ) (𝒆𝒉𝒕 ) ( ) ( ) ( ) ( )
⇆
−𝒗𝒆 −𝒗𝒆 −𝒗𝒆 −𝒗𝒆 −𝒗𝒆 −𝒗𝒆

↓↑ ↑ ↓↑ ↓↑ ↓↑ ↓
Increa con Dampe
Fluctuati Explosi Fluctuati Unifo Decre Fluctuati
sing sta d
on ve on rm asing on
nt
Explosive Fluctuation Uniform Fluctuation Damped Fluctuation
& Divergent & Neither Divergent & Convergent
nor Convergent
______________________________________________________________________________
Dynamically Unstable Neither Dynamically Stable

Hence it is numerically confirmed that for Dynamic Stability, in Complex root; (𝒉 < 𝟎).
TOPIC 134: ALTERNATIVE FORMS OF COMPLEX ROOTS
Complex Roots are:

𝒓𝟏 , 𝒓𝟐 = 𝒉 ± 𝝊𝒊
However, two more alternative forms exit: 𝒓𝟏 , 𝒓𝟐 = 𝑹(𝒄𝒐𝒔𝜽 ± 𝒊 𝒔𝒊𝒏𝜽) = 𝑹𝒆±𝒊𝜽
Therefore, three forms are there:
1. Cartesian Form (𝒉 ± 𝒗𝒊): Based on Quadrant – I of cartesian coordinates, with 𝒉 on 𝒙-
axis and 𝝊 on 𝒚-axis.
𝟑 𝟑√𝟑
Examples: 𝟐
+ 𝟐 𝒊
𝟒(√𝟑 + 𝒊)
2. Polar Form {𝑹(𝒄𝒐𝒔𝜽 ± 𝒊 𝒔𝒊𝒏𝜽)}: Based on Polar Functions which have angles (𝜽)
involved in them. Aka Triangular Form.
𝝅 𝝅
Example: 𝟐 (𝒄𝒐𝒔 + 𝒊 𝒔𝒊𝒏 )
𝟔 𝟔
3. Exponential Form {𝑹𝒆±𝒊𝜽 }: Based on exponential function {𝒆𝒇(𝒙) }.
𝒊𝝅
Examples: 𝟒𝒆 𝟑
𝒊𝝅
−
√𝟐𝒆 𝟒
De Moivre’s Theorem: If complex number has a power 𝒏 other than 𝟏, then all three forms can
be rewritten as follows:
Cartesian Form with (𝒏 ≠ 𝟏): It is easy as we simply raise the power of (𝒉 ± 𝝊𝒊) to 𝒏:
(𝒉 ± 𝝊𝒊)𝒏
Exponential Form with (𝒏 ≠ 𝟏): It is also easy as we simply raise the power of (𝑹 𝒆±𝒊𝜽 ) to 𝒏:
𝒏
= (𝑹 𝒆±𝒊𝜽 )
𝒏
= (𝑹)𝒏 (𝒆±𝒊𝜽 )
𝑹𝒏 𝒆±(𝒊𝜽)𝒏 = 𝑹𝒏 𝒆±𝒊𝒏𝜽
Observation: In Exponential Form, via De Moivre’s theorem, 𝑹 is raised to 𝑹𝒏 and 𝒆±(𝒊𝜽) is
raised to 𝒆±(𝒊𝜽)𝒏.
Polar Form with (𝒏 ≠ 𝟏): Using observations from Exponential form, raise 𝑹 to 𝑹𝒏 and 𝜽
becomes 𝒏𝜽:
______________________________________________________________________________
𝑅𝑛 𝑛𝜃 𝑛𝜃
⏞ (𝒄𝒐𝒔 ⏞
𝑹 𝜽 ± 𝒊 𝒔𝒊𝒏 ⏞
𝜽)
𝑹𝒏 {𝒄𝒐𝒔 (𝒏𝜽) ± 𝒊 𝒔𝒊𝒏 (𝒏𝜽)}

Summarizing: Gathering all forms of Complex Number(Roots) under De Moivre’s theorem:
(𝒉 ± 𝝊𝒊)𝒏 =
⏟ 𝑹𝒏 𝒆±𝒊𝒏𝜽
⏟ =⏟𝑹𝒏 {𝒄𝒐𝒔 (𝒏𝜽) ± 𝒊 𝒔𝒊𝒏 (𝒏𝜽)}
𝐶𝑎𝑟𝑡𝑒𝑠𝑖𝑎𝑛 𝑓𝑜𝑟𝑚 𝐸𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙 𝑓𝑜𝑟𝑚 𝑃𝑜𝑙𝑎𝑟 𝑓𝑜𝑟𝑚
TOPIC 135: ALTERNATIVE FORMS OF COMPLEX ROOTS: INTERCONVERSION
Inter-conversion should be easy while remembering the basic trigonometric ratios:
𝒑𝒆𝒓𝒑 𝒃𝒂𝒔𝒆
= 𝐬𝐢𝐧 𝜽 = 𝐜𝐨𝐬 𝜽
𝒉𝒚𝒑𝒕 𝒉𝒚𝒑𝒕
𝝊 𝒉
= 𝐬𝐢𝐧 𝜽 = 𝐜𝐨𝐬 𝜽
𝑹 𝑹
Rearranging;
𝝊 𝒉
= 𝐬𝐢𝐧 𝜽 , = 𝐜𝐨𝐬 𝜽:
𝑹 𝑹
𝝊 = 𝑹 𝒔𝒊𝒏𝜽 , 𝒉 = 𝑹 𝒄𝒐𝒔𝜽
Extracting value of 𝜽 from both
equations
𝝊 𝒉 Moreover, Pythagorean Theorem also guides:
𝜽 = 𝒔𝒊𝒏−𝟏 ( ) , 𝜽 = 𝒄𝒐𝒔−𝟏 ( )
𝑹 𝑹 (𝑯𝒚𝒑𝒐𝒕𝒆𝒏𝒖𝒔𝒆)𝟐 = (𝑩𝒂𝒔𝒆)𝟐 + (𝑷𝒆𝒓𝒑𝒆𝒏𝒅𝒊𝒄𝒖𝒍𝒂𝒓)𝟐
𝝊 𝒉 𝑹𝟐 = 𝒉𝟐 + 𝝊𝟐
𝜽 = 𝒔𝒊𝒏−𝟏 ( ) = 𝒄𝒐𝒔−𝟏 ( )
𝑹 𝑹 𝑹 = √𝒉𝟐 + 𝝊𝟐
𝟑𝒊𝝅
Example 6 (p-520): Find the Cartesian form of the following Exponential form 𝟓𝒆 𝟐 .
Solution: Rewriting the given Exponential form to compare with the standard form:
𝟑𝝅
𝒊( )
𝟓𝒆 𝟐
𝑹𝒆±𝒊𝜽
𝟑𝝅
Here 𝑹 = 𝟓, 𝜽 = 𝟐 .
Cartesian Form: (𝒉 ± 𝝊𝒊); however, we don’t have values of 𝒉 and 𝝊.
As already known: 𝝊 = 𝑹 𝒔𝒊𝒏𝜽 & 𝒉 = 𝑹 𝒄𝒐𝒔𝜽
𝝊 = 𝑹 𝒔𝒊𝒏 𝜽 & 𝒉 = 𝑹 𝒄𝒐𝒔 𝜽 ; putting values:
𝟑𝝅 𝟑𝝅
𝝊 = 𝟓 𝒔𝒊𝒏 & 𝒉 = 𝟓 𝒄𝒐𝒔
𝟐 𝟐
𝝊 = 𝟓(−𝟏) & 𝒉 = 𝟓 (𝟎)
𝝊 = −𝟓 & 𝒉 = 𝟎
Returning values of 𝝊 & 𝒉 in (𝒉 ± 𝝊𝒊):
= 𝟎 ± −𝟓𝒊
𝑪𝒂𝒓𝒕𝒆𝒔𝒊𝒂𝒏 𝑭𝒐𝒓𝒎 = −𝟓𝒊
In addition to Cartesian form, Polar form can also be found:
______________________________________________________________________________
={ 𝑹
⏟ (𝒄𝒐𝒔 𝜽
⏟ ± 𝒊 𝒔𝒊𝒏 𝜽
⏟ )}
𝑔𝑖𝑣𝑒𝑛 𝑔𝑖𝑣𝑒𝑛 𝑔𝑖𝑣𝑒𝑛
= {𝑹
⏟ )}
5 3𝜋 3𝜋
2 2
𝟑𝝅 𝟑𝝅
𝑷𝒐𝒍𝒂𝒓 𝑭𝒐𝒓𝒎 = [𝟓 {𝒄𝒐𝒔 ( ) ± 𝒊 𝒔𝒊𝒏 ( )}]
𝟐 𝟐
Note: This equation can be further simplified to two numerical values, due to (±).
Example 7 (p-520): Find Polar and Exponential forms of the following Cartesian form (𝟏 + √3𝒊).
Solution: Labelling the given Cartesian form and comparing with the standard form:
( 𝟏 + √𝟑 𝒊)
(𝒉 ± 𝝊𝒊)
Here 𝒉 = 𝟏, 𝝊 = √𝟑.
Polar Form: { 𝑹
⏟ (𝒄𝒐𝒔 ⏟
𝜽 ± 𝒊 𝒔𝒊𝒏 ⏟
𝜽 )}
𝑢𝑛𝑘𝑛𝑜𝑤𝑛 𝑢𝑛𝑘𝑛𝑜𝑤𝑛 𝑢𝑛𝑘𝑛𝑜𝑤𝑛
we have values of 𝒉 and 𝝊 to find value of 𝑹 = √𝒉𝟐 + 𝝊𝟐 :
𝟐
𝑹 = √(𝟏)𝟐 + (√𝟑)
𝑹=𝟐
{ 𝑹
⏟ (𝒄𝒐𝒔 ⏟
𝜽 )}
𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑒𝑑 𝑢𝑛𝑘𝑛𝑜𝑤𝑛 𝑢𝑛𝑘𝑛𝑜𝑤𝑛
𝝊 𝒉
It’s possible to find value of 𝜽 as well using 𝜽 = 𝒔𝒊𝒏−𝟏 ( ) = 𝒄𝒐𝒔−𝟏 ( )
𝑹 𝑹
√𝟑 𝟏
𝜽 = 𝒔𝒊𝒏−𝟏 ( ) = 𝒄𝒐𝒔−𝟏 ( )
𝟐 𝟐
√𝟑 𝟏
𝜽 = 𝒔𝒊𝒏−𝟏 ( ) = 𝒄𝒐𝒔−𝟏 ( )
𝟐 𝟐
𝜽 = 𝟔𝟎° = 𝟔𝟎°
{ 𝑹
⏟ (𝒄𝒐𝒔 ⏟
𝜽 )}
𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑒𝑑 𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑒𝑑 𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑒𝑑
= {𝑹
⏟ )}
2 60° 60°
𝑷𝒐𝒍𝒂𝒓 𝑭𝒐𝒓𝒎 = {𝟐 × (𝒄𝒐𝒔 𝟔𝟎° ± 𝒊 𝒔𝒊𝒏 𝟔𝟎°)}
Note: This equation can be further simplified to two numerical values, due to (±).
In addition to Polar form, Exponential form can also be found:
Exponential Form (𝑹𝒆±𝒊𝜽 ):
𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑒𝑑 𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑒𝑑
⏞ ±𝒊 ⏞
𝜽
𝑹 𝒆
2 60°
⏞ 𝒆±𝒊 𝜽⏞
𝑹
______________________________________________________________________________
= 𝟐𝒆±𝒊( 60° )
𝟏𝟖𝟎° 𝝅
𝜽 can also be written in terms of 𝝅, as (𝟔𝟎° = 𝟑
= 𝟑
):
𝝅
±𝒊( )
= 𝟐𝒆 𝟑
𝝅
±𝒊( )
𝑬𝒙𝒑𝒐𝒏𝒆𝒏𝒕𝒊𝒂𝒍 𝑭𝒐𝒓𝒎 = 𝟐𝒆 𝟑
Hence, interconversion among the three forms (Cartesian, Polar & Exponential) is possible using
the formula given above.
______________________________________________________________________________
Lesson 31
A MARKET MODEL WITH PRICE EXPECTATIONS
TOPIC 136: A MARKET MODEL WITH PRICE EXPECTATIONS: FRAMEWORK
In a simple market model:

𝑸
⏟𝒅 = 𝒇 (𝑷)
⏟ 𝑸
⏟𝒔 = 𝒇 (𝑷)
⏟
𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑐𝑢𝑟𝑟𝑒𝑛𝑡 𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑐𝑢𝑟𝑟𝑒𝑛𝑡
𝑞𝑢𝑎𝑛𝑡𝑖𝑡𝑦 𝑝𝑟𝑖𝑐𝑒 𝑞𝑢𝑎𝑛𝑡𝑖𝑡𝑦 𝑝𝑟𝑖𝑐𝑒
𝑑𝑒𝑚𝑎𝑛𝑑 𝑠𝑢𝑝𝑝𝑙𝑦
Current quantity demanded depends only on Current quantity supplied depends only on
Current Price. Current Price.
However, in real-life, trend of price-change can affect the 𝑸𝒅 . Trends help in forming (price)
expectations which affect the current 𝑸𝒅 .
Trends show how the changes in variable takes place. i.e.
𝒅𝑷 𝒅𝑷
1. ( )>𝟎 1. ( )>𝟎
𝒅𝒕 𝒅𝒕
𝟐
𝒅 𝑷 𝟐
𝒅 𝑷
2. ( 𝟐) > 𝟎 2. ( 𝟐) < 𝟎
𝒅𝒕 𝒅𝒕
𝒅𝑷 𝒅𝑷
1. ( )<𝟎 1. ( )<𝟎
𝒅𝒕 𝒅𝒕
𝒅𝟐 𝑷 𝒅𝟐 𝑷
2. ( 𝒅𝒕𝟐 ) >𝟎 2. ( 𝒅𝒕𝟐 ) <𝟎
So, the Trend can be understood from 2 points.

𝒅𝑷
1. Is price increasing or decreasing over time – Rate of change of price ( ).
𝒅𝒕
2. Is Rate of change of price increasing or decreasing over time – Slope of Rate of
𝒅 𝒅𝑷
change of price {𝒅𝑷 ( 𝒅𝒕 )}.
______________________________________________________________________________
Mathematically, the trend is measured can be measured via 1st order and 2nd order
derivatives and the new demand and supply functions will be:
𝑷𝒓𝒊𝒄𝒆−𝑻𝒓𝒆𝒏𝒅
⏞
𝒅𝑷(𝒕) 𝒅𝟐 𝑷(𝒕)
𝑸
⏟𝒅 =𝒇 𝑷(𝒕)
⏟ , ⏟𝒅𝒕
, ⏟𝒅𝒕𝟐
𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑐𝑢𝑟𝑟𝑒𝑛𝑡
𝑝𝑟𝑖𝑐𝑒 𝑅𝑎𝑡𝑒 𝑜𝑓 𝑆𝑙𝑜𝑝𝑒 𝑜𝑓 𝑅𝑎𝑡𝑒
𝑞𝑢𝑎𝑛𝑡𝑖𝑡𝑦
𝑝𝑟𝑖𝑐𝑒−𝑐ℎ𝑎𝑛𝑔𝑒 𝑜𝑓 𝑝𝑟𝑖𝑐𝑒−𝑐ℎ𝑎𝑛𝑔𝑒
𝑑𝑒𝑚𝑎𝑛𝑑 { }
𝑜𝑣𝑒𝑟 𝑡𝑖𝑚𝑒 𝑜𝑣𝑒𝑟 𝑡𝑖𝑚𝑒
𝑷𝒓𝒊𝒄𝒆−𝑻𝒓𝒆𝒏𝒅
⏞
𝒅𝑷(𝒕) 𝒅𝟐 𝑷(𝒕)
𝑸
⏟𝒔 =𝒇 𝑷(𝒕)
⏟ , ⏟𝒅𝒕
, ⏟𝒅𝒕𝟐
𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑐𝑢𝑟𝑟𝑒𝑛𝑡
𝑝𝑟𝑖𝑐𝑒 𝑅𝑎𝑡𝑒 𝑜𝑓 𝑆𝑙𝑜𝑝𝑒 𝑜𝑓 𝑅𝑎𝑡𝑒
𝑞𝑢𝑎𝑛𝑡𝑖𝑡𝑦
𝑝𝑟𝑖𝑐𝑒−𝑐ℎ𝑎𝑛𝑔𝑒 𝑜𝑓 𝑝𝑟𝑖𝑐𝑒−𝑐ℎ𝑎𝑛𝑔𝑒
𝑠𝑢𝑝𝑝𝑙𝑖𝑒𝑑 { }
𝑜𝑣𝑒𝑟 𝑡𝑖𝑚𝑒 𝑜𝑣𝑒𝑟 𝑡𝑖𝑚𝑒
Current quantity demanded and quantity demanded depend not only on Current Price, but
also the Price-Trend.
Writing the general equations of the price expectations augmented 𝑸𝒅 and 𝑸𝒔 functions:
𝑒𝑓𝑓𝑒𝑐𝑡 𝑜𝑓
𝑐𝑢𝑟𝑟𝑒𝑛𝑡 𝑒𝑓𝑓𝑒𝑐𝑡 𝑜𝑓
𝑝𝑟𝑖𝑐𝑒 𝑃𝑟𝑖𝑐𝑒−𝑇𝑟𝑒𝑛𝑑
⏞
𝑸𝒅 = 𝜶 −𝜷𝑷
⏟ ⏞⏟
+𝒎𝑷′ + 𝒏𝑷′′ (𝜶, 𝜷 > 𝟎)
𝑆𝑖𝑚𝑝𝑙𝑒 𝐷𝑒𝑚𝑎𝑛𝑑 𝒑𝒓𝒊𝒄𝒆 𝒆𝒙𝒑𝒆𝒄𝒕𝒂𝒕𝒊𝒐𝒏𝒔
𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝒂𝒖𝒈𝒆𝒎𝒆𝒏𝒕𝒆𝒅
𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑒𝑓𝑓𝑒𝑐𝑡 𝑜𝑓
𝑃𝑟𝑖𝑐𝑒 𝑃𝑟𝑖𝑐𝑒−𝑇𝑟𝑒𝑛𝑑
⏞
𝑸𝒔 = −𝜸 +𝜹𝑷
⏟ ⏞
+𝒖𝑷′ + 𝒘𝑷′′
⏟ (𝜸, 𝜹 > 𝟎)
𝑆𝑖𝑚𝑝𝑙𝑒 𝑆𝑢𝑝𝑝𝑙𝑦 𝒑𝒓𝒊𝒄𝒆 𝒆𝒙𝒑𝒆𝒄𝒕𝒂𝒕𝒊𝒐𝒏𝒔
𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝒂𝒖𝒈𝒆𝒎𝒆𝒏𝒕𝒆𝒅
𝜶, 𝜷, 𝜸, 𝜹 are the conventional parameters, however, 𝒎, 𝒏, 𝒖 and 𝒘 are parameters generated
by terms of price expectations.
Assuming price expectations in demand function only and not in supply function:
𝑠𝑖𝑚𝑝𝑙𝑒 𝑏𝑢𝑦𝑒𝑟𝑠 ℎ𝑎𝑣𝑒
𝑑𝑒𝑚𝑎𝑛𝑑 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝑝𝑟𝑖𝑐𝑒 𝑒𝑥𝑝𝑒𝑐𝑡𝑎𝑡𝑖𝑜𝑛𝑠
⏞
𝑸𝒅 = 𝜶 − 𝜷𝑷 + ⏞
⏟ 𝑷′ + 𝒏
𝒎 ⏟ 𝑷′′
≠0 ≠0
Effect of 𝒎 and 𝒏: If 𝒎 > 𝟎, 𝑸𝒅 ↑, as buyer is expecting the price to increase over time and
consider it cheaper to buy now.
Similarly, if 𝒏 > 𝟎, 𝑸𝒅 will ↑ even faster as rate of price-increase is increasing over time.
𝑠𝑖𝑚𝑝𝑙𝑒 𝑠𝑢𝑝𝑝𝑙𝑖𝑒𝑟𝑠 ℎ𝑎𝑣𝑒 𝑛𝑜
𝑠𝑢𝑝𝑝𝑙𝑦 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝑝𝑟𝑖𝑐𝑒 𝑒𝑥𝑝𝑒𝑐𝑡𝑎𝑡𝑖𝑜𝑛𝑠
⏞
𝑸𝒔 = −𝜸 + 𝜹𝑷 + ⏞
⏟ 𝑷′ + 𝒘
𝒖 ⏟ 𝑷′′
=0 =0
Therefore, we have modified market model with price expectations:
𝑸𝒅 = 𝜶 − 𝜷𝑷 + 𝒎𝑷′ + 𝒏𝑷′′
𝑸𝒔 = −𝜸 + 𝜹𝑷
TOPIC 137: A MARKET MODEL WITH PRICE EXPECTATIONS: SOLUTION
Assuming Market-Clearing at all points in time (𝑸𝒅 = 𝑸𝒔 ).

𝜶 −𝜷𝑷 + 𝒎𝑷′ + 𝒏𝑷′′ = −𝜸 +𝜹𝑷
______________________________________________________________________________
𝒏 𝑷′′ + 𝒎𝑷′ −𝜷𝑷 − 𝜹𝑷 = −𝜶 − 𝜸

Normalizing the equation by dividing by 𝒏.
𝒎 (𝜷+𝜹) −(𝜶+𝜸)
𝑷′′ + ( 𝒏 ) 𝑷′ + {− 𝒏 } 𝑷 = 𝒏
Comparing with the standard form of First Order Differential Equation
𝒚′′ + 𝒂𝟏 𝒚′ + 𝒂𝟐 𝒚 = 𝒃
𝒎 (𝜷+𝜹) 𝜶+𝜸
𝑷⏟′′ + (⏟
𝒏
⏟′ + {−
)𝑷 ⏟ 𝒏 }𝑷 ⏟ = (−
⏟ 𝒏 ) comparable with
𝑦 ′′ 𝑦′ 𝑦
𝑎1 𝑎2 𝑏
′ ′ ′′ ′′ 𝒎 (𝜷+𝜹) 𝜶+𝜸
Here; 𝒚 = 𝑷, 𝒚 = 𝑷 , 𝒚 = 𝑷 , 𝒂𝟏 = , 𝒂𝟐
= − 𝒏 , 𝒃 = −( 𝒏 )
𝒏
Particular Solution: Particular solution of a Second Order Differential Equation 𝒚𝒑 has three
cases. Starting from 1st case; 𝒂𝟐 ≠ 𝟎.
>0 >0
⏞+⏞
(𝜷 𝜹)
Here; 𝒂𝟐 = − { 𝒏
⏟
}; 𝜷 and 𝜹 are already known to be positive. 𝒏 is considered to be non-zero
≠0
(𝜷+𝜹)
(𝒏 ≠ 𝟎), since buyer is assumed to have price-expectations. Therefore, the expression − ≠
𝒏
𝒃
𝟎. Therefore, we use the Case – I of the 𝒚𝒑 = 𝒂𝟐
.
𝜶+𝜸
−( )
𝒚𝒑 = 𝒏
(𝜷 + 𝜹)
−
𝒏
𝜶+𝜸
𝒚𝒑 =
𝜷+𝜹
Interpretation: Since all parameters; 𝜶, 𝜷, 𝜸 and 𝜹 are positive constants, so 𝒚𝒑 will also be a
constant value. Therefore, a Stationary Equilibrium will be there.
TOPIC 138: A MARKET MODEL WITH PRICE EXPECTATIONS: POSSIBLE CASES
Depending on the value of discriminant, the nature of roots of characteristic equation gets defined.
This gives rise to three cases of complementary functions:
𝒂𝟐𝟏 − 𝟒𝒂𝟐
−𝒂𝟏 ± √ ⏟
𝒅𝒊𝒔𝒄𝒓𝒊𝒎𝒊𝒏𝒂𝒏𝒕
𝒓 = 𝒓𝟏 , 𝒓𝟐 =
𝟐
𝒂𝟐𝟏 − 𝟒𝒂𝟐 ⋛ 𝟎 OR 𝒂𝟐𝟏 ⋛ 𝟒𝒂𝟐
1. Distinct Real Roots (𝒂𝟐𝟏 > 𝟒𝒂𝟐 ): Here; 𝒂𝟏 = 𝒏 , 𝒂𝟐 = − 𝒏
, 𝒃 = −(𝒏
), so substituting
the values, we get:
𝒂𝟐𝟏 > 𝟒 𝒂𝟐
𝒎 𝟐 (𝜷 + 𝜹)
( ) >𝟒−
𝒏 𝒏
𝒎 𝟐 (𝜷 + 𝜹)
( ) > −𝟒
𝒏 𝒏
Forming the complementary function with Distinct Real Roots:
𝑷𝒄 = 𝑨𝟏 𝒆 𝒓𝟏 𝒕 + 𝑨𝟐 𝒆 𝒓𝟐 𝒕
Boxes avoid clutter:
______________________________________________________________________________
𝟏 𝒎 𝒎 𝟐 𝜷+𝜹
Where, 𝒓𝟏 , 𝒓𝟐 = 𝟐 {− 𝒏 ± √( 𝒏 ) + 𝟒 ( 𝒏
)}
Accordingly, General Solution is:
𝜶+𝜸
𝑨 𝟏 𝒆 𝒓𝟏 𝒕 + 𝑨 𝟐 𝒆 𝒓𝟐 𝒕 + (
𝑷(𝒕) = ⏟ )
𝑃𝑐
⏟𝜷 + 𝜹
𝑃𝑝
2. Repeated Real Roots (𝒂𝟐𝟏 = 𝟒𝒂𝟐 ): Here; 𝒂𝟏 = 𝒏 , 𝒂𝟐 = − 𝒏
, 𝒃 = −(𝒏
), so substituting
the values, we get:
𝒂𝟐𝟏 = 𝟒 𝒂𝟐
𝒎 𝟐 (𝜷 + 𝜹)
( ) =𝟒−
𝒏 𝒏
𝒎 𝟐 (𝜷 + 𝜹)
( ) = −𝟒
𝒏 𝒏
Forming the complementary function with Repeated Real Roots:
𝑷𝒄 = 𝑨𝟑 𝒆 𝒓 𝒕 + 𝑨𝟒 𝒕𝒆 𝒓 𝒕
𝟏 𝒎 𝒎 𝟐 𝜷+𝜹
Where, 𝒓 = {− ± √(⏟ ) + 𝟒 ( )}
𝟐 𝒏 𝒏 𝒏
0
𝟏 𝒎
(− ± √𝟎) 𝒓=
𝟐 𝒏
𝒎
𝒓 =−
𝟐𝒏
Substituting the value of 𝒓 in the complementary function (𝒚𝒄 ):
𝒎 𝒎
− 𝒕 − 𝒕
𝑷𝒄 = 𝑨𝟑 𝒆 𝟐𝒏 + 𝑨𝟒 𝒕𝒆 𝟐𝒏
𝒎𝒕 𝒎𝒕
(− ) (− )
𝑷𝒄 = 𝑨𝟑 𝒆 𝟐𝒏 + 𝑨𝟒 𝒕𝒆 𝟐𝒏
𝒎𝒕
(− )
𝒎𝒕
(− ) 𝜶+𝜸
𝑷(𝒕) = ⏟
𝑨𝟑 𝒆 𝟐𝒏 + 𝑨𝟒 𝒕𝒆 𝟐𝒏 +( )
𝑃𝑐
⏟𝜷 + 𝜹
𝑃𝑝
3. Complex Roots (𝒂𝟐𝟏 < 𝟒𝒂𝟐 ): Here; 𝒂𝟏 = =− 𝒏
, 𝒂𝟐 𝒏
, 𝒃 = −( 𝒏
), so substituting the
values, we get:
𝒂𝟐𝟏 < 𝟒 𝒂𝟐
𝒎 𝟐 (𝜷 + 𝜹)
( ) <𝟒−
𝒏 𝒏
𝒎 𝟐 (𝜷 + 𝜹)
( ) < −𝟒
𝒏 𝒏
Forming the complementary function with Complex Roots:
𝑷𝒄 = 𝒆 𝒉 𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝝊 𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝝊 𝒕)
𝒂𝟏 𝟏 𝟐
Where, 𝒉 = − , 𝝊 = √𝟒 𝒂𝟐 − ( 𝒂𝟏 )
𝟐 𝟐
𝒎
𝟐
𝒏 𝟏 𝜷+𝜹 𝒎
𝒉=− , 𝝊 = √−𝟒 −( )
𝟐 𝟐 𝒏 𝒏
______________________________________________________________________________
𝒎 𝟏 𝜷+𝜹 𝒎 𝟐
𝒉 = − 𝟐𝒏 , 𝝊 = 𝟐 √−𝟒 ( 𝒏
) − (𝒏)
𝜶+𝜸
𝒆 𝒉 𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝝊 𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝝊 𝒕) + (
𝑷(𝒕) = ⏟ )
𝑃𝑐
⏟𝜷 + 𝜹
𝑃𝑝
𝒎 𝜶+𝜸
− 𝒕
𝑷(𝒕) = ⏟𝒆 𝟐𝒏 (𝑨 𝟓 𝐜𝐨𝐬 𝝊 𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝝊 𝒕) + ( )
𝑃𝑐
⏟𝜷 + 𝜹
𝑃𝑝
Parametric Interpretation for Dynamic Stability:

Parameters involved: 𝜷, 𝜹, 𝒎 and 𝒏
Given stipulations: 𝜷 > 𝟎, 𝜹 > 𝟎
Therefore, the dynamic stability depends on values of 𝒎 and 𝒏.
𝒎 𝒎 𝟐 𝜷+𝜹
As 𝒏 is in all of the terms i.e. ( 𝒏 ), ( 𝒏 ) and {−𝟒 ( 𝒏
)}, its importance is pivotal in determining
the nature of roots as well as dynamic stability.
Dropping the possibility 𝒏 = 𝟎, as the result of terms given above will become undefined.
Pivotal Role of (𝒏) and secondary role of (𝒎) in Dynamic Stability

𝒏>𝟎 𝒏<𝟎
Only Distinct Real Roots can exist. All Cases can exist:
(𝒓𝟏 > 𝟎, 𝒓𝟐 < 𝟎) 1. Distinct Real Roots, dynamically
(Dynamically Unstable) stable if (𝒎 < 𝟎).
Can be Dynamically Stable only if (𝑨𝟏 = 𝟎). 2. Repeated Real Roots,
𝑷𝒄 = 𝑨𝟏 𝒆𝒓𝟏𝒕 + 𝑨𝟐 𝒆𝒓𝟐𝒕 dynamically stable if (𝒎 < 𝟎).
+𝒗𝒆 −𝒗𝒆 3. Complex Roots, dynamically
⏞𝟏 𝒕
𝒓 ⏞𝟐 𝒕
𝒓
𝑷
⏟𝒄 = 𝑨
⏟𝟏 × 𝒆⏟ +𝑨 ⏟𝟐 × 𝒆⏟ stable if (𝒎 < 𝟎).
→0 =0 𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙 ≠0 𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙
⏟ 𝑔𝑟𝑜𝑤𝑡ℎ ⏟ 𝑑𝑒𝑐𝑎𝑦
=0 →0
Conclusion: Dynamic Stability is ensured if {(𝒏 < 𝟎) ∧ (𝒎 < 𝟎)} with the exception of where
(𝒏 > 𝟎) ⋀(𝑨𝟏 = 𝟎).
TOPIC 139: A MARKET MODEL WITH PRICE EXPECTATIONS: NUMERICAL
Example 1 (p. # 530): Let the Demand and Supply Functions be:
𝑸𝒅 = 𝟒𝟐 − 𝟒𝑷 − 𝟒𝑷′ + 𝑷′′
𝑸𝒔 = −𝟔 + 𝟖𝑷
With initial conditions 𝑷(𝟎) = 𝟔 and 𝑷′ (𝟎) = 𝟒. Assuming market clearance at every point of time,
find the time path 𝑷(𝒕).
Recalling the standard form of expectations-augmented demand and supply functions:
𝑸𝒅 = 𝜶 − 𝜷𝑷 + 𝒎𝑷′ + 𝒏𝑷′′
𝑸𝒔 = −𝜸 + 𝜹𝑷 + 𝒖𝑷′ + 𝒘𝑷′′
Comparison gives:
𝛼 𝛽 𝑚 𝑛
⏞ −⏞
𝑸𝒅 = 𝟒𝟐 𝟒𝑷 + ⏞ ⏞ 𝑷′′
(−𝟒) 𝑷′ + (𝟏)
′
𝑸𝒔 = − ⏟
𝟔+⏟𝟖𝑷+⏟ 𝟎𝑷 +⏟ 𝟎 𝑷′′
𝛾 𝛿 𝑢 𝑤
______________________________________________________________________________
Therefore, here 𝜶 = 𝟒𝟐, 𝜷 = 𝟒, 𝜸 = 𝟔, 𝜹 = 𝟖, 𝒎 = −𝟒 < 𝟎 , 𝒏 = 𝟏 > 𝟎 , 𝒖 = 𝟎, 𝒘 = 𝟎.

Double Boxed value of 𝒏 = 𝟏 > 𝟎 shows that only Distinct Real Roots are possible. It also
confirms that 𝒓𝟏 > 𝟎 and 𝒓𝟐 < 𝟎.
So dynamic unstability is present unless 𝑨𝟏 = 𝟎.
Therefore, we solve the system to get definite solution which evaluates 𝑨𝟏 and 𝑨𝟐 . Consequently,
possibility of 𝑨𝟏 = 𝟎 can be checked as well.
𝟏 𝒎 𝒎 𝟐 𝜷+𝜹
Distinct Real Roots: 𝒓𝟏 , 𝒓𝟐 = {− ± √( ) + 𝟒 ( )}
𝟐 𝒏 𝒏 𝒏
𝟐
𝟏 (−𝟒) (−𝟒) 𝟒+𝟖
𝒓𝟏 , 𝒓𝟐 = {− ± √{ } + 𝟒( )}
𝟐 𝟏 𝟏 𝟏
𝟏
𝒓𝟏 , 𝒓𝟐 = (𝟒 ± √𝟏𝟔 + 𝟒𝟖)
𝟐
𝟏
𝒓𝟏 , 𝒓𝟐 = (𝟒 ± √𝟔𝟒)
𝟐
𝟏
𝒓𝟏 , 𝒓𝟐 = (𝟒 ± 𝟖)
𝟐
𝒓𝟏 , 𝒓𝟐 = 𝟔, −𝟐
𝜶+𝜸
𝑨 𝟏 𝒆 𝒓𝟏 𝒕 + 𝑨 𝟐 𝒆 𝒓𝟐 𝒕 + (
𝑷(𝒕) = ⏟ )
𝑃𝑐
⏟𝜷 + 𝜹
𝑃𝑝
𝟒𝟐 + 𝟔
𝑨𝟏 𝒆(𝟔)𝒕 + 𝑨𝟐 𝒆(−𝟐)𝒕 + {
𝑷(𝒕) = ⏟ }
⏟𝟒 + 𝟖
𝑃𝑐
𝑃𝑝
𝟔𝒕 −𝟐𝒕
𝑷(𝒕) = 𝑨𝟏 𝒆 + 𝑨𝟐 𝒆 +𝟒
Definitizing the General Solution by introducing the initial condition; (𝒕 = 𝟎):
⏟ = 𝑨𝟏 𝒆𝟔(𝟎) + 𝑨𝟐 𝒆−𝟐(𝟎) + 𝟒
𝑷(𝟎)
⏟ = 𝑨𝟏 𝒆𝟎 + 𝑨𝟐 𝒆𝟎 + 𝟒
𝑷(𝟎)
𝟔
𝟔 = 𝑨𝟏 (𝟏) + 𝑨𝟐 (𝟏) + 𝟒
𝟔 − 𝟒 = 𝑨𝟏 + 𝑨𝟐
𝟐 = 𝑨𝟏 + 𝑨 𝟐
Differentiating the General Solution for introducing 𝑷′(𝟎).
𝒅 𝒅
{𝑷(𝒕)} = (𝑨𝟏 𝒆𝟔𝒕 + 𝑨𝟐 𝒆−𝟐𝒕 + 𝟒)
𝒅𝒕 𝒅𝒕
′ (𝒕)
𝒅 𝒅 𝒅
𝑷 = (𝑨𝟏 𝒆𝟔𝒕 ) + (𝑨𝟐 𝒆−𝟐𝒕 ) + (𝟒)
𝒅𝒕 𝒅𝒕 𝒅𝒕
𝑷′ (𝒕) = 𝑨𝟏 (𝟔)𝒆𝟔𝒕 + 𝑨𝟐 (−𝟐)𝒆𝟔𝒕 + 𝟎
𝑷′ (𝒕) = 𝟔𝑨𝟏 𝒆𝟔𝒕 − 𝟐𝑨𝟐 𝒆𝟔𝒕
Substituting the initial condition for First Derivative of General Solution; 𝑷′ (𝟎):
𝑷′ (𝟎) = 𝟔𝑨𝟏 𝒆𝟔(𝟎) − 𝟐𝑨𝟐 𝒆𝟔(𝟎)
𝑷′ (𝟎) = 𝟔𝑨𝟏 𝒆𝟎 − 𝟐𝑨𝟐 𝒆𝟎
⏟
______________________________________________________________________________
𝑷′ (𝟎) = 𝟔𝑨𝟏 (𝟏) − 𝟐𝑨𝟐 (𝟏)

⏟
4
𝟒 = 𝟔𝑨𝟏 − 𝟐𝑨𝟐
Simultaneously solving the resultant equations:
𝟐
⏟ = 𝑨𝟏 + 𝑨𝟐 , 𝟒 = 𝟔 𝑨
⏟𝟏 − 𝟐𝑨𝟐
𝑨𝟏 =𝟐−𝑨𝟐 𝑠𝑢𝑏𝑠𝑡𝑖𝑡𝑢𝑡𝑖𝑛𝑔
𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑨𝟏
𝟒 = 𝟔(𝟐 − 𝑨𝟐 ) − 𝟐𝑨𝟐
𝟒 = 𝟏𝟐 − 𝟔𝑨𝟐 − 𝟐𝑨𝟐
𝟒 − 𝟏𝟐 = −𝟖𝑨𝟐
𝑨𝟐 = 𝟏
To find 𝑨𝟏 , plugin the value of 𝑨𝟐 in 𝑨𝟏 = 𝟐 − 𝑨𝟐 :
𝑨𝟏 = 𝟐 − 𝟏
𝑨𝟏 = 𝟏
Returning the values of 𝑨𝟏 and 𝑨𝟐 in General Solution:
𝑷(𝒕) = 𝑨𝟏 𝒆𝟔𝒕 + 𝑨𝟐 𝒆−𝟐𝒕 + 𝟒
𝑷(𝒕) = 𝟏 𝒆𝟔𝒕 + 𝟏 𝒆−𝟐𝒕 + 𝟒
𝑷(𝒕) = 𝒆𝟔𝒕 + 𝒆−𝟐𝒕 + 𝟒
Interpretation: As already observed, (𝒏 > 𝟎)⋀(𝑨𝟏 ≠ 𝟎), only dynamic unstable price time-path
with Distinct Real Roots is possible. Specifically, the dynamic unstability in Definite Solution is
due to the presence of a +𝒗𝒆 root (𝒓𝟏 = 𝟔).
______________________________________________________________________________
Example 1 (p. # 531): Let the Demand and Supply Functions be:
𝑸𝒅 = 𝟒𝟎 − 𝟐𝑷 − 𝟐𝑷′ − 𝑷′′
𝑸𝒔 = −𝟓 + 𝟑𝑷
With initial conditions 𝑷(𝟎) = 𝟏𝟐 and 𝑷′ (𝟎) = 𝟏. Assuming market clearance at every point of
time, find the time path 𝑷(𝒕).
Recalling the standard form of expectations-augmented demand and supply functions:
𝑸𝒅 = 𝜶 − 𝜷𝑷 + 𝒎𝑷′ + 𝒏𝑷′′
𝑸𝒔 = −𝜸 + 𝜹𝑷 + 𝒖𝑷′ + 𝒘𝑷′′
Comparison gives:
𝛼 𝛽 𝑚 𝑛
⏞ −⏞
𝑸𝒅 = 𝟒𝟎 𝟐𝑷 +⏞
(−𝟐) 𝑷′ + ⏞
(−𝟏) 𝑷′′
′
𝑸𝒔 = − ⏟
𝟓+⏟𝟑𝑷+⏟𝟎𝑷 +⏟ 𝟎 𝑷′′
𝛾 𝛿 𝑢 𝑤
Therefore, here 𝜶 = 𝟒𝟎, 𝜷 = 𝟐, 𝜸 = 𝟓, 𝜹 = 𝟑, 𝒎 = −𝟐 < 𝟎 , 𝒏 = −𝟏 < 𝟎 , 𝒖 = 𝟎, 𝒘 = 𝟎.
Double Boxed value of 𝒏 = −𝟏 < 𝟎 shows that any of three types of Roots are possible.
______________________________________________________________________________
Single Boxed value of 𝒎 = −𝟐 < 𝟎 in addition to (𝒏 < 𝟎) shows Dynamic Stability in any of the
cases. For deciding which case is present here, we apply the test for nature of roots.
Therefore, we apply the test: 𝒂𝟐𝟏 ⋛ −𝟒𝒂𝟏. However, here the given parameters are 𝒎, 𝒏, 𝜷 and
𝜹, so the test will appear as follows:
𝒎 𝟐 (𝜷 + 𝜹)
( ) ⋛ −𝟒
𝒏 𝒏
−𝟐 𝟐 (𝟐 + 𝟑)
( ) ⋛ −𝟒
−𝟏 −𝟏
𝟏𝟔 ⏟ < 𝟐𝟎
𝑐𝑜𝑚𝑝𝑙𝑒𝑥
𝑟𝑜𝑜𝑡𝑠
𝟏 𝒎 𝒎 𝟐 𝜷+𝜹
Complex Roots: 𝒓𝟏 , 𝒓𝟐 = 𝒉 ± 𝝊𝒊 = 𝟐 {− 𝒏 ± √( 𝒏 ) + 𝟒 ( 𝒏
)}
Substituting the values:
𝟐
𝟏 (−𝟐) (−𝟐) 𝟐+𝟑
𝒓𝟏 , 𝒓𝟐 = {− ± √{ } + 𝟒( )}
𝟐 (−𝟏) (−𝟏) −𝟏
𝟏
𝒓𝟏 , 𝒓𝟐 = (−𝟐 ± √𝟒 − 𝟐𝟎)
𝟐
𝟏
𝒓𝟏 , 𝒓𝟐 = (−𝟐 ± √−𝟏𝟔
⏟ )
𝟐
⏟
√ 16×√
⏟−1
4 𝑖
𝟏
⏟𝟏 , 𝒓𝟐 = (−𝟐 ± 𝟒𝒊)
𝒓
𝟐
𝒉±𝝊𝒊
𝒓𝟏 , 𝒓𝟐 = ⏟
−𝟏 ± 𝟐𝒊
𝒉±𝝊𝒊
Extracted values: 𝒉 = −𝟏 and 𝝊 = 𝟐 .
𝜶+𝜸
⏟𝒉 𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝝊 𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝝊 𝒕) + (
𝑷(𝒕) = 𝒆 )
𝑃𝑐
⏟𝜷 + 𝜹
𝑃𝑝
−𝟏 𝒕
𝑷(𝒕) = 𝒆
⏟ (𝑨𝟓 𝐜𝐨𝐬 𝟐 𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝟐 𝒕) + ⏟
𝟗
𝑃𝑐 𝑃𝑝
−𝒕 {𝑨
𝑷(𝒕) = 𝒆 𝟓 𝐜𝐨𝐬(𝟐𝒕) + 𝑨𝟔 𝐬𝐢𝐧(𝟐𝒕)} + 𝟗
Definitizing the General Solution by introducing the initial condition; (𝒕 = 𝟎):
⏟ = 𝒆−(𝟎) [𝑨𝟓 𝐜𝐨𝐬{𝟐(𝟎)} + 𝑨𝟔 𝐬𝐢𝐧{𝟐(𝟎)}] + 𝟗
𝑷(𝟎)
⏟ = 𝟏 × {𝑨𝟓 𝐜𝐨𝐬(𝟎) + 𝑨𝟔 𝐬𝐢𝐧(𝟎)} + 𝟗
𝑷(𝟎)
12
𝟏𝟐 = {𝑨𝟓 (𝟏) + 𝑨𝟔 (𝟎)} + 𝟗
𝟏𝟐 = 𝑨𝟓 + 𝟗
𝑨𝟓 = 𝟑
Differentiating the General Solution for introducing 𝑷′(𝟎).
𝒅 𝒅
{𝑷(𝒕)} = { 𝒆−𝒕 {𝑨𝟓 𝐜𝐨𝐬(𝟐𝒕) + 𝑨𝟔 𝐬𝐢𝐧(𝟐𝒕)} + 𝟗}
𝒅𝒕 𝒅𝒕
______________________________________________________________________________
ℎ(𝑥) 𝑔(𝑥) ℎ(𝑥) 𝑔(𝑥)

′ (𝒕) 𝒅 ⏞ ⏞ 𝒅 ⏞ ⏞ 𝒅
𝑷 = 𝒅𝒕
[ 𝒆−𝒕 {𝑨𝟓 𝐜𝐨𝐬(𝟐𝒕)}] +
𝒅𝒕
[ 𝒆−𝒕 {𝑨𝟔 𝐬𝐢𝐧(𝟐𝒕)}] +
𝒅𝒕
(𝟗)
⏟ ⏟
𝑝𝑟𝑜𝑑𝑢𝑐𝑡 𝑟𝑢𝑙𝑒 𝑝𝑟𝑜𝑑𝑢𝑐𝑡 𝑟𝑢𝑙𝑒
𝒅 −𝒕 𝒅
𝑷′ (𝒕) = [ (𝒆 ) × {𝑨𝟓 𝐜𝐨𝐬(𝟐𝒕)} + 𝑨𝟓 {𝐜𝐨𝐬(𝟐𝒕)} × (𝒆−𝒕 )]
⏟
𝒅𝒕 ⏟
𝒅𝒕
−𝑒 −𝑡 −2×sin(2𝑡)
𝒅 −𝒕 𝒅
+[ (𝒆 ) × {𝑨𝟔 𝐬𝐢𝐧(𝟐𝒕)} + 𝑨𝟔 {𝐬𝐢𝐧(𝟐𝒕)} × (𝒆−𝒕 )]
⏟
𝒅𝒕 ⏟
𝒅𝒕
−𝑒 −𝑡 2×cos(2𝑡)
𝑷′ (𝒕) = [−𝒆−𝒕 × [{𝑨𝟓 𝐜𝐨𝐬(𝟐𝒕)}] + 𝑨𝟓 {−𝟐 × 𝐬𝐢𝐧(𝟐𝒕)} × (𝒆−𝒕 )]
+ [−𝒆−𝒕 × {𝑨𝟔 𝐬𝐢𝐧(𝟐𝒕)} + 𝑨𝟔 {𝟐 × 𝐜𝐨𝐬(𝟐𝒕)} × (𝒆−𝒕 )]
Substituting the initial condition for First Derivative of General Solution; 𝑷′ (𝟎):
𝑷′ (𝟎) = [{−𝒆−(𝟎) } × [𝑨𝟓 𝐜𝐨𝐬{𝟐(𝟎)}] + 𝑨𝟓 [−𝟐 × 𝐬𝐢𝐧{𝟐(𝟎)}] × {𝒆−(𝟎) }]
+ [{−𝒆−(𝟎) } × [𝑨𝟔 𝐬𝐢𝐧{𝟐(𝟎)}] + 𝑨𝟔 [𝟐 × 𝐜𝐨𝐬{𝟐(𝟎)}] × {𝒆−(𝟎) }]
𝑷′ (𝟎) = [( − 𝟏) × [𝑨𝟓 𝐜𝐨𝐬{𝟐(𝟎)}] + 𝑨𝟓 [−𝟐 × 𝐬𝐢𝐧{𝟐(𝟎)}] × (𝟏)]
⏟
+ [( − 𝟏) × [𝑨𝟔 𝐬𝐢𝐧{𝟐(𝟎)}] + 𝑨𝟔 [𝟐 × 𝐜𝐨𝐬{𝟐(𝟎)}] × (𝟏)]
𝑷′ (𝟎) = [ − 𝑨𝟓 𝐜𝐨𝐬(𝟎)
⏟ ⏟ + 𝑨𝟓 {−𝟐 × 𝐬𝐢𝐧(𝟎)
⏟ }] + [{ − 𝑨𝟔 𝐬𝐢𝐧(𝟎)
⏟ } + 𝑨𝟔 {𝟐 × 𝐜𝐨𝐬(𝟎)
⏟ }]
1 𝟏 0 0 1
𝟏 = [(−𝑨𝟓 × 𝟏) + 𝑨𝟓 (−𝟐 × 𝟎)] + [(−𝑨𝟔 × 𝟎) + 𝑨𝟔 (−𝟐 × 𝟏)]
𝟏 = (−𝑨𝟓 + 𝟎) + (𝟎 − 𝟐𝑨𝟔 )
𝟏 = −𝑨𝟓 − 𝟐𝑨𝟔
Simultaneously solving the resultant equations:
𝑨𝟓 = 𝟑 , 𝟏 = − 𝑨
⏟𝟓 − 𝟐𝑨𝟔
3
To find 𝑨𝟔 , plugin the value of 𝑨𝟓 in 𝟏 = −𝑨𝟓 − 𝟐𝑨𝟔 :
𝑨𝟓 = 𝟑 , 𝟏 = −𝟑 − 𝟐𝑨𝟔
−𝟐𝑨𝟔 = −𝟑 − 𝟏
𝑨𝟔 = 𝟐
Returning the values of 𝑨𝟓 and 𝑨𝟔 in General Solution:
𝑷(𝒕) = 𝒆−𝒕 { 𝑨𝟓 𝐜𝐨𝐬(𝟐𝒕) + 𝑨𝟔 𝐬𝐢𝐧(𝟐𝒕)} + 𝟗
𝑷(𝒕) = 𝒆−𝒕 { 𝟑 𝐜𝐨𝐬(𝟐𝒕) + 𝟐 𝐬𝐢𝐧(𝟐𝒕)} + 𝟗
−𝒕
𝑷(𝒕) = 𝒆⏟ ⏟𝐜𝐨𝐬(𝟐𝒕) + 𝟐 𝐬𝐢𝐧(𝟐𝒕)} +
{𝟑 ⏟
𝟗
𝑑𝑎𝑚𝑝𝑒𝑛𝑖𝑛𝑔 𝑤𝑎𝑣𝑒 𝑙𝑖𝑘𝑒 𝑓𝑙𝑢𝑐𝑡𝑎𝑢𝑡𝑖𝑜𝑛 𝑒𝑞𝑢𝑖𝑙𝑖𝑏𝑟𝑖𝑢𝑚
Interpretation: As already observed, (𝒏 < 𝟎) ∧ (𝒎 < 𝟎), dynamic stability is present in the
price time-path. Specifically, the time-path will be fluctuating as well as dampening.
______________________________________________________________________________
______________________________________________________________________________
Lesson 32
PHILLIPS CURVE: THE PHILLIPS RELATION
TOPIC 140: PHILLIPS CURVE: THE PHILLIPS RELATION
The Interaction of Inflation and Unemployment

An application of the second-order differential equation with a macro model dealing with the
problem of inflation and unemployment.
Phillips relation is the most widely used concepts in the modern analysis of the problem of inflation
and unemployment.
Original Phillips Relation:
Its original formulation depicts an empirically based negative relation between the rate of growth
of money wage and the rate of unemployment:
𝒘 = 𝒇(𝑼) [𝒇′ (𝑼) < 𝟎]

Here, 𝒘: Rate of Growth of Money Wage (𝑾).
𝑾̇
𝒘=
𝑾
𝑼: Rate of Unemployment.
Adapted Phillips Relation:
Unemployment pertains to labor market. However, later version of Phillips Relation uses inflation
rate instead of wage rate.
Its intuitive as, increasing 𝒘 increases the cost of production, which causes (cost push) inflation
(𝑷̇):
𝒘 ↑ ⇒ ↑ (𝒄𝒐𝒔𝒕 𝒐𝒇 𝒑𝒓𝒐𝒅𝒖𝒄𝒕𝒊𝒐𝒏) ⇒ ↑ 𝒑
𝒅𝑷
Here; 𝒑 is the rate of growth of inflation(𝑷̇ = 𝒅𝒕 ).
𝑷̇
𝒑=
𝑷
Dual Effect of Increased Money-Wages:
𝒘↑
______________________________________________________________________________
⇓
𝑸
Productivity 𝑻 = ( ) ↑
𝑳
⇓
𝑪 ⇓
Per unit Cost (𝑳) ↓
⇓
Cost of Production (𝑪) ↓
⇓
𝒑↓ 𝒑↑
Effect of Productivity on Inflation Effect of money-wage on Inflation
𝒑𝑻 ↓ 𝒑𝒘 ↑
𝒑𝑻 ↓ ⋛ 𝒑𝒘 ↑
Outcome of Off-setting between inflation increase & inflation decrease
Therefore, introducing the role of productivity (𝑻) in the link between money-wage (𝒘) and
inflation (𝑷̇):
𝟏
(𝒑 ∝ 𝒘) (𝒑 ∝ )
𝑻
𝒑= ⏞
𝒘
⏟ − ⏞
⏟
𝑻
𝑒𝑥𝑜𝑔𝑒𝑛𝑜𝑢𝑠 𝑒𝑥𝑜𝑔𝑒𝑛𝑜𝑢𝑠
Original Phillips Relation: 𝒘 = 𝒇(𝑼). Assuming a linear relationship, we get:
𝒘 = 𝜶 − 𝜷𝑼
Substituting value of 𝒘 in 𝒑 = 𝒘 − 𝑻:
𝒑 = 𝜶 − 𝜷𝑼 − 𝑻
𝒑 = 𝜶 − 𝑻 − 𝜷𝑼 (𝜶, 𝜷 > 𝟎)
Interpretation: From the Original Phillips Relation, we have derived the Adapted Phillips
Relation. Where, Productivity (𝑻) and Unemployment rate (𝑼) can negatively affect the inflation-
growth rate (𝒑).
TOPIC 141: PHILLIPS CURVE: EXPECTATIONS AUGMENTED PHILLIPS RELATION
More recently, economists have preferred to use the expectations-augmented version of the
Phillips relation:
𝒘 = 𝒇(𝑼) + 𝒈𝝅 (𝟎 < 𝒈 ≤ 𝟏)
𝑒𝑛𝑐ℎ𝑎𝑛𝑐𝑒𝑑 𝑒𝑓𝑓𝑒𝑐𝑡
𝑜𝑓 𝑖𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛
𝑒𝑥𝑝𝑒𝑐𝑡𝑎𝑡𝑖𝑜𝑛𝑠
𝑜𝑛 𝑤𝑎𝑔𝑒𝑠
𝒘 = 𝒇(𝑼) + ⏞
𝒈 𝝅
⏟
𝑒𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑟𝑎𝑡𝑒
𝑜𝑓 𝑖𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛
Nobel laureate Professor Friedman, suggests that if an inflationary trend has been in effect long
enough, people are apt to form certain inflation expectations which they then attempt to
incorporate into their money-wage demands. Thus 𝒘 should be an increasing function of 𝝅.
Inflation Expectations can also be incorporated in the Adapted Phillips Relation (𝒑 = 𝜶 − 𝑻 − 𝜷𝑼):
𝒑 = 𝜶 − 𝑻 − 𝜷𝑼
______________________________________________________________________________
𝑖𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛
𝒑 = 𝜶 − 𝑻 − 𝜷𝑼 + ⏞
𝒈𝝅 (𝟎 < 𝒈 ≤ 𝟏)
Formation of Expectations – Adaptive Expectations:

After introducing 𝝅, it’s meaningful to assess how 𝝅 are formed. Here, we assume Adaptive
Expectations hypothesis instead of rational expectations hypothesis.
Note: Adaptive Expectations are when one makes forecasts of future values of a variable using
only past values of the variable. Rational expectations are when forecasts of future values are
made using all available information.
Adaptive Expectations can be assessed by rate of change of expectations. Therefore, we
calculate the derivative of Expectations variable (𝝅):
𝑎𝑐𝑡𝑢𝑎𝑙 𝑟𝑎𝑡𝑒 𝑒𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑟𝑎𝑡𝑒
𝑜𝑓 𝑖𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑖𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛
𝒅𝝅
∝ ( ⏞
𝒑 − ⏞
𝝅 )
⏟
𝒅𝒕
𝑅𝑎𝑡𝑒 𝑜𝑓 ⏟
𝑐ℎ𝑎𝑛𝑔𝑒 𝑜𝑓
𝑑𝑖𝑠𝑐𝑟𝑒𝑝𝑎𝑛𝑐𝑦
𝒅𝝅
= ⏟𝒋 (𝒑 − 𝝅) (𝟎 < 𝒋 ≤ 𝟏)
𝒅𝒕 𝑎𝑑𝑗𝑢𝑠𝑡𝑚𝑒𝑛𝑡
𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡
The discrepancy term (𝒑 − 𝝅) can have following possibilities:
Actual Inflation exceeds Actual Inflation equals Expected Inflation exceeds
Expected Inflation Expected Inflation Actual Inflation
(𝒑 > 𝝅) (𝒑 = 𝝅) (𝒑 < 𝝅)
Upward revision of Expected No adjustment needed. Downward revision of
Inflation needed. Expected Inflation needed.
Interpretation: Inflation Expectations are incorporated in the form of Expected rate of inflations
(𝝅). While Adaptive Inflation Expectations is introduced using the derivative of 𝝅.
TOPIC 142: PHILLIPS CURVE: FEEDBACK FROM INFLATION TO UNEMPLOYMENT
Expectations-augmented Phillips Relation & Adaptive Expectations Hypothesis, can serve as a

two-equation system with two endogenous variables ( 𝒑 & 𝝅 )and one exogenous (𝑼):
𝒑 = 𝜶 − 𝑻 − 𝜷𝑼 + 𝒈 𝝅
⏟ (𝟎 < 𝒈 ≤ 𝟏)
𝐸𝑥𝑝𝑒𝑐𝑡𝑎𝑡𝑖𝑜𝑛𝑠−𝑎𝑢𝑔𝑚𝑒𝑛𝑡𝑒𝑑 𝑃ℎ𝑖𝑙𝑙𝑖𝑝𝑠
𝑅𝑒𝑙𝑎𝑡𝑖𝑜𝑛
𝒅𝝅
= 𝒋( 𝒑 − 𝝅 ) (𝟎 < 𝒋 ≤ 𝟏)
⏟𝒅𝒕
𝐴𝑑𝑎𝑝𝑡𝑖𝑣𝑒 𝐸𝑥𝑝𝑒𝑐𝑡𝑎𝑡𝑖𝑜𝑛𝑠
𝐻𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠
𝑼 can also be explained like 𝒑 and 𝝅, so the model becomes richer in behavioral characteristics.
Feedback Effect of 𝑼: Explaining 𝑼 can help us understand the feedback effect of inflation on
unemployment (𝑼 ⇄ 𝒑).
1. 𝑼 → 𝒑: Expectations-augmented Phillips relation is based mostly on supply side and deals
with 𝑼 → 𝒑:
______________________________________________________________________________
𝑤 (𝑙𝑎𝑏𝑜𝑟 𝑖𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛
𝑐𝑜𝑠𝑡) 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑣𝑖𝑡𝑦 𝑒𝑥𝑝𝑒𝑐𝑡𝑎𝑡𝑖𝑜𝑛𝑠
𝒑=⏞
𝜶 − 𝜷𝑼 − ⏟
⏟ ⏞
𝑻 + ⏞
𝒈𝝅
⏟
𝑠𝑢𝑝𝑝𝑙𝑦 𝑠𝑢𝑝𝑝𝑙𝑦 𝑠𝑢𝑝𝑝𝑙𝑦
𝑠𝑖𝑑𝑒 𝑠𝑖𝑑𝑒 𝑠𝑖𝑑𝑒 𝑣𝑖𝑎
⏟ 𝑤−𝑖𝑛𝑐𝑟𝑒𝑎𝑠𝑒
𝑠𝑢𝑝𝑝𝑙𝑦 𝑠𝑖𝑑𝑒 𝑒𝑓𝑓𝑒𝑐𝑡 𝑜𝑛 𝑖𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛
2. 𝑼 ← 𝒑: However, the Feedback effect on unemployment on inflation can be through:
Feedback Effect: 𝑼 ← 𝒑
𝑼 ←←←←←←←←←←← 𝒑
Government Intervention
Demand Side
Fiscal Policy Monetary Policy
𝒑↑ ∆𝑮 ↑: (𝑔𝑖𝑣𝑒𝑛 𝒑) ∆𝑴 ↑: (𝑔𝑖𝑣𝑒𝑛 𝒑)
⇓ ⇓ ⇓
∆(Consumption-Saving ∆(Real-Expenditure) ∆(Real Money)
Pattern)
⇓ ⇓ ⇓
∆(𝑨𝑫) ∆(𝑨𝑫) ∆(𝑨𝑫)
⇓ ⇓ ⇓
∆(𝑼) ∆(𝑼) ∆(𝑼)
For simplicity sake, we consider Monetary Policy channel only.
Let 𝑴 be the nominal-money balance.
𝑴̇
𝒎 = be the rate of growth of nominal money balance.
𝑴
Recalling the Monetary Policy channel:
∆(𝑼) ← ∆(𝑨𝑫) ← ∆(𝑹𝒆𝒂𝒍 𝑴𝒐𝒏𝒆𝒚) ← ∆(𝑴)
∆(𝑼) ← ∆(𝑹𝒆𝒂𝒍 𝑴𝒐𝒏𝒆𝒚)

Mathematizing the same:
𝒅𝑼 𝟏
( )∝
𝒅𝒕 (𝒎 − 𝒑)
𝒓𝑴
( )
𝑘>0 𝑷
𝒅𝑼 ⏞
= ⏞
−𝒌
⏟ (𝒎 − 𝒑)
⏟
⏟
𝒅𝒕 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑜𝑓 𝑟𝑎𝑡𝑒 𝑜𝑓 𝑔𝑟𝑜𝑤𝑡ℎ
𝑟𝑎𝑡𝑒 𝑜𝑓 𝑐ℎ𝑎𝑛𝑔𝑒 𝑝𝑟𝑜𝑝𝑜𝑟𝑡𝑖𝑜𝑛𝑎𝑙𝑖𝑡𝑦 𝑜𝑓 𝑟𝑒𝑎𝑙 𝑚𝑜𝑛𝑒𝑦
𝑜𝑓 𝑢𝑛𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡
Expression (𝒎 − 𝒑) can be explained as follows:
𝑴̇ 𝑷̇
(𝒎 − 𝒑) = −
𝑴 𝑷
= 𝒓𝑴 − 𝒓𝑷
Rate of growth of real money = 𝒓(𝑴)
𝑷
𝒅𝑼
Interpretation: Feedback effect of inflation on unemployment is developed: 𝒅𝒕
= −𝒌(𝒎 − 𝒑).
Negative relation exists between the two variables.
TOPIC 143: PHILLIPS CURVE: TIME PATH OF EXPECTED RATE OF INFLATION
After developing feedback effect (monetary policy) of inflation on unemployment, we have a

system of equation, rich in behavioral characteristics:
______________________________________________________________________________
𝒅𝑼
= −𝒌(𝒎 − 𝒑)
𝒅𝒕
Including Expectations-augmented Phillips Relation:
𝒑 = 𝜶 − 𝑻 − 𝜷𝑼 + 𝒈𝝅
we get a system of equations with the three variables 𝝅, 𝒑 and 𝑼.
Time Path: Two of the three variables need to be removed to get a time-path of the remaining
(single) variable. We remove 𝒑 and 𝑼 in step – I and step – II, respectively.
Step – I (Remove 𝒑):
𝒅𝝅
Given the Adaptive Expectations Hypothesis: 𝒅𝒕
= 𝒋( 𝒑 − 𝝅), we substitute
𝒑 = 𝜶 − 𝑻 − 𝜷𝑼 + 𝒈𝝅
𝒅𝝅
= 𝒋( 𝜶 − 𝑻 − 𝜷𝑼 + 𝒈𝝅 − 𝝅)
𝒅𝒕
𝒅𝝅
= 𝒋(𝜶 − 𝑻 − 𝜷𝑼 + 𝒈𝝅 − 𝝅)
𝒅𝒕
𝒅𝝅
= 𝒋{𝜶 − 𝑻 − 𝜷𝑼 + (𝒈 − 𝟏)𝝅}
𝒅𝒕
𝒅𝝅
= 𝒋(𝜶 − 𝑻 − 𝜷 𝑼 ) + 𝒋(𝒈 − 𝟏)𝝅
⏟
𝒅𝒕
𝒑 𝑖𝑠 𝑟𝑒𝑚𝑜𝑣𝑒𝑑,𝑼 𝑟𝑒𝑚𝑎𝑖𝑛𝑠
Step – II (Remove 𝑼): Substitution of 𝑼 is not straightforward, as value of 𝑼 is not available.
𝒅𝑼 𝒅𝑼
Instead, we have ( 𝒅𝒕 ). So, we create ( 𝒅𝒕 ) by differentiating the given equation with 𝑼:
𝒅𝟐 𝝅 𝒅
= {𝒋(𝜶 − 𝑻 − 𝜷 𝑼 ) + 𝒋(𝒈 − 𝟏)𝝅}
𝒅𝒕𝟐 𝒅𝒕
𝑒𝑥𝑜𝑔𝑒𝑛𝑜𝑢𝑠 =𝑓(𝑡)
𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟
𝒅 ⏞ 𝒅 𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟 =𝑓(𝑡)
= 𝒋 × {( ⏞
𝜶 − ⏞
𝑻 − 𝜷 𝑼 )} + 𝒋 × {( ⏞
𝒈 ⏞ }
− 𝟏) 𝝅
𝒅𝒕 𝒅𝒕
𝒅𝟐 𝝅 𝒅 𝒅
𝟐
= 𝒋 × (−𝜷) (𝑼) + 𝒋 × (𝒈 − 𝟏) (𝝅)
𝒅𝟐 𝝅 𝒅𝑼 𝒅𝝅
= −𝜷𝒋 + 𝒋(𝒈 − 𝟏)
𝒅𝒕𝟐 ⏟ 𝒅𝒕 𝒅𝒕
𝑓𝑒𝑒𝑑𝑏𝑎𝑐𝑘
𝒑 𝑜𝑛 𝑼
𝒑 𝑟𝑒𝑖𝑛𝑡𝑟𝑜𝑑𝑢𝑐𝑒𝑑
𝒅𝟐 𝝅 ⏞ 𝒅𝝅
= −𝜷𝒋 {−𝒌 (𝒎 − 𝒑 )} + 𝒋(𝒈 − 𝟏)
⏟
𝑼 𝑟𝑒𝑚𝑜𝑣𝑒𝑑
𝒅𝟐 𝝅 𝒅𝝅
𝟐
= 𝒋𝜷𝒌𝒎 − 𝒋𝜷𝒌 𝒑 + 𝒋(𝒈 − 𝟏)
𝒅𝒕 𝒅𝒕
𝒅𝝅 𝒅𝝅
To remove 𝒑 again, we recall 𝒅𝒕
= 𝒋(𝒑 − 𝝅) and rearrange to get 𝒑 in terms of 𝝅 & 𝒅𝒕
.
𝒅𝝅 𝟏
= (𝒑 − 𝝅)
𝒅𝒕 𝒋
𝒅𝝅 𝟏
+𝝅=𝒑
𝒅𝒕 𝒋
𝒅𝝅 𝟏
𝒑= +𝝅
𝒅𝒕 𝒋
______________________________________________________________________________
𝒅𝝅
Substituting the reintroduced 𝒑 to replace it with 𝝅 & 𝒅𝒕
.
𝒅𝟐 𝝅 𝒅𝝅 𝟏 𝒅𝝅
𝟐
= 𝒋𝜷𝒌𝒎 − 𝒋𝜷𝒌 ( + 𝝅) + 𝒋(𝒈 − 𝟏)
𝒅𝒕 𝒅𝒕 𝒋 𝒅𝒕
𝒅𝟐 𝝅 𝒅𝝅 𝟏 𝒅𝝅
𝟐
= 𝒋𝜷𝒌𝒎 − 𝒋𝜷𝒌 ( ) − 𝒋𝜷𝒌 (𝝅) + 𝒋(𝒈 − 𝟏)
𝒅𝒕 𝒅𝒕 𝒋 𝒅𝒕
𝒇′′ (𝒙) 𝒇′ (𝒙) 𝒇′ (𝒙)
⏞
𝒅𝟐𝝅 ⏞
𝒅𝝅
𝒇(𝒙) ⏞
𝒅𝝅
= 𝒋𝜷𝒌𝒎 − 𝜷𝒌 ⏞ ) + 𝒋(𝒈 − 𝟏)
− 𝒋𝜷𝒌 ( 𝝅
𝒅𝒕𝟐 𝒅𝒕 𝒅𝒕
⏟ ( ) ( )
𝑝𝑜𝑡𝑒𝑛𝑡𝑖𝑎𝑙 𝑠𝑒𝑐𝑜𝑛𝑑 𝑜𝑟𝑑𝑒𝑟 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑙 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛
𝒅𝟐 𝝅 𝒅𝝅 𝒅𝝅
𝟐
+ 𝜷𝒌 ( ) − 𝒋(𝒈 − 𝟏) ( ) + 𝒋𝜷𝒌𝝅 = 𝒋𝜷𝒌𝒎
𝒅𝟐 𝝅 𝒅𝝅
+⏟{𝜷𝒌 − 𝒋(𝒈 − 𝟏)} ( ) + 𝒋𝜷𝒌
⏟ (𝝅) = 𝒋𝜷𝒌𝒎
⏟
𝑎1 𝑎2 𝑏
Here, 𝒂𝟏 = {𝜷𝒌 − 𝒋(𝒈 − 𝟏)}, 𝒂𝟐 = 𝒋𝜷𝒌, 𝑏 = 𝒋𝜷𝒌𝒎
𝒃
Particular Solution: 𝒚𝒑 = , if 𝑎2 ≠ 0.
𝒂𝟐
>0 >0 >0
⏞
Here 𝒂𝟐 = ⏟
𝒋 𝜷 ⏞ ≠ 𝟎. So, we can use 𝒚𝒑 = 𝒃 .
⏞ 𝒌
𝒂 𝟐
>0
𝒃 𝒋𝜷𝒌𝒎
Here 𝝅𝒑 = =
𝒂𝟐 𝒋𝜷𝒌
𝝅𝒑 = 𝒎
Interpretation: Thus, in this model, the intertemporal equilibrium value of the expected rate of
inflation (𝝅) hinges exclusively on the rate of growth of nominal money (𝒎).
Complementary Function: For choosing among the three cases of complementary function, we
test for the nature of roots.
Nature of Roots: Deploying 𝒂𝟐𝟏 ⋛ 𝟒𝒂𝟐
{𝜷𝒌 − 𝒋(𝒈 − 𝟏)} ⋛ 𝟒(𝒋𝜷𝒌)
Perhaps it is not possible to confirm one of the three possibilities ⋛ , given the parametric
restrictions of symbols 𝜷, 𝒌, 𝒈 and 𝒋. So, we assume the possibility of any of the nature of roots
i.e. Distinct Real, Repeated Real and Complex Roots.
𝟏
Formula for Roots: 𝒓𝟏 , 𝒓𝟐 = 𝟐 (−𝒂𝟏 ± √𝒂𝟐𝟏 − 𝟒𝒂𝟐 )
+𝑣𝑒
⏞
+𝑣𝑒
⏞ −𝒋
Here, 𝒂𝟏 = 𝜷𝒌 𝒈
⏟ −𝟏 > 𝟎 and 𝒂𝟐 = 𝟒(𝒋𝜷𝒌) > 𝟎
⏟
<1
( −𝑣𝑒
)
{ }
Dynamic Stability: It all boils down to question of dynamic stability of all three cases:
______________________________________________________________________________
<𝑎1
𝑟1 (−𝑣𝑒)
⏞
𝟏
i- Distinct Real Roots: 𝒓𝟏 , 𝒓𝟐 = 𝟐 −𝒂𝟏 ⏞
±
⏟ 𝟐
√𝒂
⏟𝟏 − 𝟒𝒂𝟐
𝑟2 (−𝑣𝑒) +𝑣𝑒
( )
Since, both 𝑟1 and 𝑟2 are −𝒗𝒆, the time-path will be dynamic stable.
𝑎
𝑟1 =− 1 <0
2
𝟏
ii- Repeated Real Roots: 𝒓𝟏 , 𝒓𝟐 = −𝒂𝟏 ⏞
±
⏟ 𝟐
𝟐 √𝒂
⏟𝟏 − 𝟒𝒂𝟐
𝑎 =0
2 𝑟 =− 1 <0
( 2 )
Since, 𝑟1 and 𝑟2 are equal and −𝒗𝒆, the time-path will be dynamic stable.
𝜐
ℎ
⏞
iii- Complex Roots: 𝒓𝟏 , 𝒓𝟐 = ⏞
−
𝒂𝟏 𝟏
± 𝟐 √𝒂𝟐
⏟ 𝟐 ⏟𝟏 − 𝟒𝒂𝟐
𝑟𝑒𝑎𝑙 −𝑣𝑒
𝑝𝑎𝑟𝑡
⏟
𝑖𝑚𝑎𝑔𝑖𝑛𝑎𝑟𝑦
( 𝑝𝑎𝑟𝑡 )
𝒂𝟏
Since, 𝒉 = − < 𝟎, the time-path will be dynamic stable.
𝟐
Caveat: In this derivation, time-path of expected rate of inflation was calculated. However, one
can do it for other variables (𝑼&𝒑) as well.
TOPIC 144: PHILLIPS CURVE: TIME PATH OF EXPECTED RATE OF INFLATION -

NUMERICAL
Example: Let the three equations of the model take the specific forms:
Expectations-augmented Phillips Relation: 𝒑 = 𝜶 − 𝑻 − 𝜷𝑼 + 𝒈𝝅
𝟏
𝒑 =( )−⏟ 𝟑 𝑼 + (𝟏)
⏟𝝅
⏟𝟔 𝛽 𝑔
𝛼
𝒅𝝅
Adaptive Expectations Hypothesis: = 𝒋(𝒑 − 𝝅)
𝒅𝒕
𝒅𝝅 𝟑
= ( ) (𝒑 − 𝝅)
𝒅𝒕 ⏟ 𝟒
𝑗
𝒅𝑼
Feedback effect of Inflation on Unemployment: = −𝒌(𝒎 − 𝒑)
𝒅𝒕
𝒅𝑼 𝟏
= − ( ) (𝒎 − 𝒑)
𝒅𝒕 ⏟𝟐
𝑘
𝟑 𝟏
The given data; 𝜷 = 𝟑, 𝑔 = 𝟏, 𝒋 = and 𝒌 = Values of 𝒂𝟏 , 𝒂𝟐 and 𝒃 can be found by using:
𝟒 𝟐
.
𝒂𝟏 = 𝜷𝒌 + 𝒋(𝟏 − 𝒈) 𝒂𝟐 = 𝒋𝜷𝒌 and 𝒃 = 𝒋𝜷𝒌𝒎
𝟏 𝟑 𝟑 𝟏 𝟑 𝟏
𝒂𝟏 = (𝟑) (𝟐) + 𝟒 (𝟏 − 𝟏) 𝒂𝟐 = (𝟒) (𝟑) (𝟐) and 𝒃 = (𝟒) (𝟑) (𝟐) 𝒎
𝟑 𝟗 𝟗
𝒂𝟏 = 𝟐 𝒂𝟐 = 𝟖 and 𝒃 = (𝟖) 𝒎
𝒃 𝟗
Particular Integral (𝝅𝒑 ): 𝝅𝒑 = 𝒂 as (𝒂𝟐 = 𝟖 ≠ 𝟎)
𝟐
______________________________________________________________________________
𝟗
(𝟖) 𝒎
𝝅𝒑 =
𝟗
(𝟖)
𝝅𝒑 = 𝒎
Interpretation: Thus, in this model, the intertemporal equilibrium value of the expected rate of
inflation (𝝅) hinges exclusively on the rate of growth of nominal money (𝒎).
Complementary Function: For choosing among the three cases of complementary function, we
test for the nature of roots.
Nature of Roots: Deploying 𝒂𝟐𝟏 ⋛ 𝟒𝒂𝟐
𝟑 𝟐 𝟗
( ) ⋛ 𝟒( )
𝟐 𝟖
𝟗 𝟗
<
𝟒 𝟐
Complex Roots are present in this case.
𝟏
Formula for Roots: 𝒓𝟏 , 𝒓𝟐 = 𝟐 (−𝒂𝟏 ± √𝒂𝟐𝟏 − 𝟒𝒂𝟐 )
𝟏 𝟑 𝟑 𝟐 𝟗
𝒓𝟏 , 𝒓𝟐 = {− ( ) ± √( ) − 𝟒 ( )}
𝟐 𝟐 𝟐 𝟖
𝟏 𝟑 𝟏 𝟗
𝒓𝟏 , 𝒓𝟐 = − ( ) ( ) ± ( ) √−
𝟐 𝟐 𝟐 ⏟ 𝟒
𝟑
×√𝒊
𝟐
𝟏 𝟑 𝟏 𝟑
𝒓𝟏 , 𝒓𝟐 = − ( ) ( ) ± ( ) √𝒊
𝟐 𝟐 𝟐 𝟐
𝟑 𝟑
𝒓𝟏 , 𝒓𝟐 = − ± √𝒊
⏟𝟒 ⏟𝟒
ℎ 𝜐
𝟑 𝟑
Substituting the extracted values 𝒉 = − and 𝝊 = in the complex root case of 𝒚𝒄 :
𝟒 𝟒
𝒉 𝒕 (𝑨
𝝅𝒄 = 𝒆 𝟓 𝐜𝐨𝐬 𝝊 𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝝊 𝒕 )
𝟑 𝟑 𝟑
− 𝒕
𝝅𝒄 = 𝒆 𝟒 (𝑨𝟓 𝐜𝐨𝐬 𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝒕)
𝟒 𝟒
𝟑
(− )𝒕 𝟑 𝟑
𝝅𝒄 = 𝒆 𝟒 {𝑨𝟓 𝐜𝐨𝐬 ( 𝒕) + 𝑨𝟔 𝐬𝐢𝐧 ( 𝒕)}
𝟒 𝟒
General Solution 𝝅(𝒕): We already know; 𝝅(𝒕) = 𝝅𝒄 + 𝝅𝒑 , so by substituting:
𝟑
(− )𝒕 𝟑 𝟑
𝝅(𝒕) = 𝒆 𝟒 {𝑨𝟓 𝐜𝐨𝐬 ( 𝒕) + 𝑨𝟔 𝐬𝐢𝐧 ( 𝒕)} + 𝒎
𝟒 𝟒
Interpretation: The time-path of Expected Rate of Inflation-growth (𝝅), is
𝟑
(− )𝒕 𝟑 𝟑 𝟑
𝝅(𝒕) = 𝒆 𝟒 {𝑨𝟓 𝐜𝐨𝐬 (𝟒 𝒕) + 𝑨𝟔 𝐬𝐢𝐧 (𝟒 𝒕)} + 𝒎 . It is dynamically stable as 𝒉 = − 𝟒 < 𝟎. It has
a fluctuating time-path converging towards equilibrium value of 𝒎.
______________________________________________________________________________
TOPIC 145: PHILLIPS CURVE: TIME PATH OF INFLATION - NUMERICAL
Time-Path of Inflation (𝒑): Easiest possible method of finding 𝒑(𝒕) should be based on simplest
𝒅𝝅
equation containing 𝒑 and already solved variable (𝝅). i.e. 𝒅𝒕 = 𝒋(𝒑 − 𝝅).
𝒅𝝅 𝟏
= (𝒑 − 𝝅)
𝒅𝒕 𝒋
𝒅𝝅 𝟏
+𝝅 =𝒑
𝒅𝒕 𝒋
______________________________________________________________________________
𝑁𝑜𝑡 𝑔𝑖𝑣𝑒𝑛 𝑎𝑙𝑟𝑒𝑎𝑑𝑦

⏞
𝒅𝝅 𝟏
𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑒𝑑
𝒑(𝒕)
⏟ = ⏞
+ ⏟ 𝝅(𝒕)
𝑡𝑖𝑚𝑒−𝑝𝑎𝑡ℎ
𝒅𝒕
⏟ 𝒋 𝑡𝑖𝑚𝑒−𝑝𝑎𝑡ℎ
𝑜𝑓 𝒑 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑜𝑓 𝝅
𝑜𝑓 𝝅(𝒕)
𝒅𝝅
Therefore, we need to calculate ( ), before finding 𝒑(𝒕), by differentiating 𝝅(𝒕) w.r.t 𝒕.
𝒅𝒕
𝒅 𝒅 (−𝟑)𝒕 𝟑 𝟑
{𝝅(𝒕)} = [𝒆 𝟒 {𝑨𝟓 𝐜𝐨𝐬 ( 𝒕) + 𝑨𝟔 𝐬𝐢𝐧 ( 𝒕)} + 𝒎]
𝒅𝒕 𝒅𝒕 𝟒 𝟒
𝑔(𝑥) ℎ(𝑥) 𝑔(𝑥) ℎ(𝑥) 𝑑𝑒𝑣𝑜𝑖𝑑
𝑜𝑓 𝒕
𝒅{𝝅(𝒕)} 𝒅 ⏞(−𝟑)𝒕 ⏞ 𝟑 𝒅 ⏞(−𝟑)𝒕 ⏞ 𝟑 𝒅
= 𝒆 𝟒 {𝑨𝟓 𝐜𝐨𝐬 ( 𝒕)} + 𝒆 𝟒 {𝑨𝟔 𝐬𝐢𝐧 ( 𝒕)} + ( 𝒎⏞ )
⏟ 𝒅𝒕 𝒅𝒕 ⏟ 𝟒 𝒅𝒕 ⏟ 𝟒 𝒅𝒕
𝑑𝜋 [ 𝑝𝑟𝑜𝑑𝑢𝑐𝑡 𝑟𝑢𝑙𝑒 ] [ 𝑝𝑟𝑜𝑑𝑢𝑐𝑡 𝑟𝑢𝑙𝑒 ]
𝑑𝑡
𝟑 𝟑
𝒅𝝅 𝒅 (− )𝒕 𝟑 𝒅 𝟑 (− )𝒕
𝒅𝒕
= 𝒅𝒕
{𝒆 𝟒 } × {𝑨𝟓 𝐜𝐨𝐬 (𝟒 𝒕)} + 𝒅𝒕
{𝑨𝟓 𝐜𝐨𝐬 (𝟒 𝒕)} × {𝒆 𝟒 }+
𝟑 𝟑
𝒅 (− )𝒕 𝟑 𝒅 𝟑 (− )𝒕
𝒅𝒕
{𝒆 𝟒 } × {𝑨𝟔 𝐬𝐢𝐧 (𝟒 𝒕)} + 𝒅𝒕
{𝑨𝟔 𝐬𝐢𝐧 (𝟒 𝒕)} × {𝒆 𝟒 }+𝟎
𝟑 𝟑
𝒅𝝅 𝟑 (− )𝒕 𝟑 𝟑 𝟑 (− )𝒕
𝒅𝒕
= (− 𝟒) 𝒆 𝟒 × {𝑨𝟓 𝐜𝐨𝐬 (𝟒 𝒕)} + 𝑨𝟓 {(− 𝟒) 𝐬𝐢𝐧 (𝟒 𝒕)} × {𝒆 𝟒 }+
𝟑 𝟑
𝟑 (− )𝒕 𝟑 𝟑 𝟑 (− )𝒕
(− 𝟒) 𝒆 𝟒 × {𝑨𝟔 𝐬𝐢𝐧 (𝟒 𝒕)} + 𝑨𝟔 {(𝟒) 𝐜𝐨𝐬 (𝟒 𝒕)} × {𝒆 𝟒 }
𝑒𝑥𝑖𝑠𝑡𝑠 𝑖𝑛 𝝅(𝒕)
𝒅𝝅 𝟑 𝟑
(− )𝒕 ⏞ 𝟑 𝟑 𝟑 𝟑
(− )𝒕
𝒅𝒕
= (− 𝟒) 𝒆 𝟒 × {𝑨𝟓 𝐜𝐨𝐬 (𝟒 𝒕)} + 𝑨𝟓 {(− 𝟒) 𝐬𝐢𝐧 (𝟒 𝒕) } × {𝒆 𝟒 }+
𝑒𝑥𝑖𝑠𝑡𝑠 𝑖𝑛 𝝅(𝒕)
𝟑 𝟑
(− )𝒕 ⏞ 𝟑 𝟑 𝟑 𝟑
(− )𝒕
(− 𝟒) 𝒆 𝟒 × {𝑨𝟔 𝐬𝐢𝐧 (𝟒 𝒕)} + 𝑨𝟔 {(𝟒) 𝐜𝐨𝐬 (𝟒 𝒕) } × {𝒆 𝟒 }
𝑒𝑥𝑖𝑠𝑡𝑠 𝑖𝑛 𝝅(𝒕) 𝑒𝑥𝑖𝑠𝑡𝑠 𝑖𝑛 𝝅(𝒕)
𝒅𝝅 𝟑 𝟑
(− )𝒕 ⏞ 𝟑 𝟑 𝟑
(− )𝒕 ⏞ 𝟑 𝟑 𝟑
= [(− 𝟒) 𝒆 𝟒 × {𝑨𝟓 𝐜𝐨𝐬 ( 𝒕)} + (− ) 𝒆 𝟒 × {𝑨𝟔 𝐬𝐢𝐧 ( 𝒕)}] + [𝑨𝟓 {(− ) 𝐬𝐢𝐧 ( 𝒕) } ×
𝒅𝒕 𝟒 𝟒 𝟒 𝟒 𝟒
𝟑 𝟑
(− )𝒕 𝟑 𝟑 (− )𝒕
{𝒆 𝟒 } + 𝑨𝟔 {(𝟒) 𝐜𝐨𝐬 (𝟒 𝒕) } × {𝒆 𝟒 }]
𝑐𝑜𝑚𝑚𝑜𝑛 𝑒𝑥𝑖𝑠𝑡𝑠 𝑖𝑛 𝝅(𝒕) 𝑒𝑥𝑖𝑠𝑡𝑠 𝑖𝑛 𝝅(𝒕)
𝒅𝝅 ⏞ 𝟑 (−𝟑)𝒕 ⏞ 𝟑 ⏞ 𝟑
= (− ) 𝒆 𝟒 {𝑨𝟓 𝐜𝐨𝐬 ( 𝒕)} + {𝑨𝟔 𝐬𝐢𝐧 ( 𝒕)}
𝒅𝒕 𝟒 𝟒 𝟒
[ ]
𝑐𝑜𝑚𝑚𝑜𝑛
⏞ (−𝟑)𝒕 𝟑 𝟑 𝟑 𝟑
+ {𝒆 𝟒 } [𝑨𝟓 {(− ) 𝐬𝐢𝐧 ( 𝒕) } + 𝑨𝟔 {( ) 𝐜𝐨𝐬 ( 𝒕) }]
𝟒 𝟒 𝟒 𝟒
𝒅𝝅 𝟑 (− )𝒕
𝟑 𝟑 𝟑 𝟑
(− )𝒕 𝟑 𝟑 𝟑 𝟑
= (− ) 𝒆 𝟒 [{𝑨𝟓 𝐜𝐨𝐬 ( 𝒕)} + {𝑨𝟔 𝐬𝐢𝐧 ( 𝒕)}] + {𝒆 𝟒 } [𝑨𝟓 {(− ) 𝐬𝐢𝐧 ( 𝒕)} + 𝑨𝟔 {( ) 𝐜𝐨𝐬 ( 𝒕)}]
𝒅𝒕 𝟒 𝟒 𝟒 𝟒 𝟒 𝟒 𝟒
Recalling the main purpose i.e. 𝒑(𝒕):
______________________________________________________________________________
𝑟𝑒𝑐𝑒𝑛𝑡𝑙𝑦
𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑒𝑑 𝑝𝑟𝑒𝑣𝑖𝑜𝑢𝑠𝑙𝑦
⏞
𝒅𝝅 𝟏
𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑒𝑑
𝒑(𝒕)
⏟ = × ⏞
+ ⏟ 𝝅(𝒕)
𝒅𝒕
⏟ 𝒋 𝑡𝑖𝑚𝑒−𝑝𝑎𝑡ℎ
𝑜𝑓 𝒑 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑜𝑓 𝝅
𝑜𝑓 𝝅(𝒕)
𝒑(𝒕) =
𝟑 𝟑
𝟑 (− )𝒕 𝟑 𝟑 (− )𝒕 𝟑 𝟑 𝟑 𝟑
[(− 𝟒) 𝒆 𝟒 [{𝑨𝟓 𝐜𝐨𝐬 (𝟒 𝒕)} + {𝑨𝟔 𝐬𝐢𝐧 (𝟒 𝒕)}] + {𝒆 𝟒 } [𝑨𝟓 {(− 𝟒) 𝐬𝐢𝐧 (𝟒 𝒕)} + 𝑨𝟔 {(𝟒) 𝐜𝐨𝐬 (𝟒 𝒕)}]] ×
𝟑
𝟏 (− )𝒕 𝟑 𝟑
𝟑 + [𝒆 𝟒 {𝑨𝟓 𝐜𝐨𝐬 (𝟒 𝒕) + 𝑨𝟔 𝐬𝐢𝐧 (𝟒 𝒕)} + 𝒎]
( )
𝟒
⏟
𝜋(𝑡)
𝒑(𝒕)
𝟑 𝟑
(− )𝒕 𝟑 𝟑 𝟑
(− )𝒕 𝟑 𝟑 𝟑 𝟑
= [(− ) 𝒆 𝟒 [{𝑨𝟓 𝐜𝐨𝐬 ( 𝒕)} + {𝑨𝟔 𝐬𝐢𝐧 ( 𝒕)}] + {𝒆 𝟒 } [𝑨𝟓 {(− ) 𝐬𝐢𝐧 ( 𝒕)} + 𝑨𝟔 {( ) 𝐜𝐨𝐬 ( 𝒕)}]]
𝟒 𝟒 𝟒 𝟒 𝟒 𝟒 𝟒
𝟒 𝟑
(− )𝒕 𝟑 𝟑
× + 𝒆 𝟒 {𝑨𝟓 𝐜𝐨𝐬 ( 𝒕) + 𝑨𝟔 𝐬𝐢𝐧 ( 𝒕)} + 𝒎
𝟑 𝟒 𝟒
𝟑
(− )𝒕 𝟑 𝟑 𝟑
(− )𝒕 𝟑 𝟑
𝒑(𝒕) = −𝒆 𝟒 [{𝑨𝟓 𝐜𝐨𝐬 ( 𝒕)} + {𝑨𝟔 𝐬𝐢𝐧 ( 𝒕)}] + {𝒆 𝟒 } [𝑨𝟓 {− 𝐬𝐢𝐧 ( 𝒕)} + 𝑨𝟔 {𝐜𝐨𝐬 ( 𝒕)}]
𝟒 𝟒 𝟒 𝟒
𝟑
(− )𝒕 𝟑 𝟑
+ 𝒆 𝟒 {𝑨𝟓 𝐜𝐨𝐬 ( 𝒕) + 𝑨𝟔 𝐬𝐢𝐧 ( 𝒕)} + 𝒎
𝟒 𝟒
𝟑
(− )𝒕 𝟑 𝟑
(− )𝒕 𝟑 𝟑
(− )𝒕 𝟑 𝟑
(− )𝒕 𝟑
𝒑(𝒕) = −𝒆 𝟒 {𝑨𝟓 𝐜𝐨𝐬 ( 𝒕)} − 𝒆 𝟒 {𝑨𝟔 𝐬𝐢𝐧 ( 𝒕)} − 𝒆 𝟒 {𝑨𝟓 𝐬𝐢𝐧 ( 𝒕)} + 𝒆 𝟒 {𝑨𝟔 𝐜𝐨𝐬 ( 𝒕)}
𝟒 𝟒 𝟒 𝟒
𝟑
(− )𝒕 𝟑 𝟑
(− )𝒕 𝟑
+ 𝒆 𝟒 {𝑨𝟓 𝐜𝐨𝐬 ( 𝒕)} + 𝒆 𝟒 {𝑨𝟔 𝐬𝐢𝐧 ( 𝒕)} + 𝒎
𝟒 𝟒
𝟑 𝟑 𝟑 𝟑
(− )𝒕 𝟑 (− )𝒕 𝟑 (− )𝒕 𝟑 (− )𝒕 𝟑
𝒑(𝒕) = −𝒆 𝟒 {𝑨𝟓 𝐜𝐨𝐬 (𝟒 𝒕)} − 𝒆 𝟒 {𝑨𝟔 𝐬𝐢𝐧 (𝟒 𝒕)} −𝒆 𝟒 {𝑨𝟓 𝐬𝐢𝐧 (𝟒 𝒕)} + 𝒆 𝟒 {𝑨𝟔 𝐜𝐨𝐬 (𝟒 𝒕)} +
𝟑 𝟑
(− )𝒕 𝟑 (− )𝒕 𝟑
𝒆 𝟒 {𝑨𝟓 𝐜𝐨𝐬 (𝟒 𝒕)} + 𝒆 𝟒 {𝑨𝟔 𝐬𝐢𝐧 (𝟒 𝒕)} +𝒎
𝟑
(− )𝒕 𝟑 𝟑
(− )𝒕 𝟑
𝒑(𝒕) = −𝒆 𝟒 {𝑨𝟓 𝐬𝐢𝐧 ( 𝒕)} + 𝒆 𝟒 {𝑨𝟔 𝐜𝐨𝐬 ( 𝒕)} +𝒎
𝟒 𝟒
⏞(−𝟑)𝒕 𝟑 𝟑
𝒑(𝒕) = 𝒆 𝟒 [{−𝑨𝟓 𝐬𝐢𝐧 ( 𝒕)} + {𝑨𝟔 𝐜𝐨𝐬 ( 𝒕)}] + 𝒎
𝟒 𝟒
𝟑
(− )𝒕 𝟑 𝟑
𝒑(𝒕) = 𝒆 𝟒 {𝑨𝟔 𝐜𝐨𝐬 ( 𝒕) − 𝑨𝟓 𝐬𝐢𝐧 ( 𝒕)} + 𝒎
𝟒 𝟒
Interpretation: The time-path of Actual rate of inflation-growth (𝒑), is
𝟑
(− )𝒕 𝟑 𝟑 𝟑
𝒑(𝒕) = 𝒆 𝟒 {𝑨𝟔 𝐜𝐨𝐬 (𝟒 𝒕) − 𝑨𝟓 𝐬𝐢𝐧 (𝟒 𝒕)} + 𝒎 . It is dynamically stable as 𝒉 = − 𝟒 < 𝟎. Like
time-path of Expected rate of inflation-growth, 𝒑(𝒕)has a fluctuating time-path converging
towards equilibrium value of 𝒎 as well.
______________________________________________________________________________
TOPIC 146: PHILLIPS CURVE: TIME PATH OF UNEMPLOYMENT – NUMERICAL
𝟏
Time-Path of Unemployment (𝑼): Feedback effect of Inflation on Unemployment is: 𝒑 = 𝟔 −
𝟑𝑼 + 𝝅.
𝟏
Rewriting it for 𝑼, we get: 𝟑𝑼 = 𝟔 + 𝝅 − 𝒑
𝟏 𝝅 𝒑
𝑼= + −
𝟏𝟖 𝟑 𝟑
𝟏 𝟏
𝑼 = (𝝅 − 𝒑) +
𝟑 𝟏𝟖
Given the time-paths of 𝝅 and 𝒑, we can get time-path of 𝑼.
______________________________________________________________________________
𝟏 𝟏
𝑼= (𝝅 − 𝒑)+
𝟑 𝟏𝟖
𝒄𝒐𝒏𝒗𝒆𝒓𝒕𝒆𝒅 𝒕𝒐
𝒕𝒊𝒎𝒆−𝒑𝒂𝒕𝒉
𝟑
⏞ 𝟏 (− )𝒕 𝟑 𝟑
𝑼(𝒕) = [ 𝒆 𝟒 {𝑨𝟓 𝐜𝐨𝐬 (𝟒 𝒕) + 𝑨𝟔 𝐬𝐢𝐧 (𝟒 𝒕)} + 𝒎 −
𝟑
𝟑
(− )𝒕 𝟑 𝟑 𝟏
𝒆 𝟒 {𝑨𝟔 𝐜𝐨𝐬 (𝟒 𝒕) − 𝑨𝟓 𝐬𝐢𝐧 (𝟒 𝒕)} + 𝒎 ] + 𝟏𝟖
𝟑
𝟏 (− )𝒕 𝟑 𝟑 𝟑 𝟑 𝟏
𝑼(𝒕) = 𝒆 𝟒 [𝑨𝟓 𝐜𝐨𝐬 (𝟒 𝒕) + 𝑨𝟔 𝐬𝐢𝐧 (𝟒 𝒕) + 𝒎 − 𝑨𝟔 𝐜𝐨𝐬 (𝟒 𝒕) − 𝑨𝟓 𝐬𝐢𝐧 (𝟒 𝒕) − 𝒎] + 𝟏𝟖
𝟑
𝟏 (−𝟑)𝒕 𝟑 𝟑 𝟑 𝟑 𝟏
𝑼(𝒕) = 𝟑 𝒆 𝟒 [𝑨𝟓 𝐜𝐨𝐬 (𝟒 𝒕) − 𝑨𝟔 𝐜𝐨𝐬 (𝟒 𝒕) + 𝑨𝟔 𝐬𝐢𝐧 (𝟒 𝒕) − 𝑨𝟓 𝐬𝐢𝐧 (𝟒 𝒕)] + 𝟏𝟖
𝟏 (−𝟑)𝒕 ⏞ 𝟑 𝟑 𝟏
𝑼(𝒕) = 𝒆 𝟒 [(𝑨𝟓 − 𝑨𝟔 ) 𝐜𝐨𝐬 ( 𝒕) + (𝑨𝟔 − 𝑨𝟓 ) 𝐬𝐢𝐧
⏟ ( 𝒕)] +
𝟑 𝟒 𝟒 𝟏𝟖
𝟏 (−𝟑)𝒕 𝟑 𝟑 𝟏
𝑼(𝒕) = 𝒆 𝟒 [(𝑨𝟓 − 𝑨𝟔 ) {𝐜𝐨𝐬 ( 𝒕)} + (𝑨𝟔 − 𝑨𝟓 ) {𝐬𝐢𝐧 ( 𝒕)}] +
⏟
𝟑 𝟒 𝟒 ⏟
𝟏𝟖
𝑈𝑐 𝑈𝑝
Interpretation: The time-path of Unemployment (𝑼), is

𝟑
𝟏 (− )𝒕 𝟑 𝟑 𝟏
𝑼(𝒕) = 𝟑 𝒆 𝟒 [(𝑨𝟓 − 𝑨𝟔 ) {𝐜𝐨𝐬 (𝟒 𝒕)} + (𝑨𝟔 − 𝑨𝟓 ) {𝐬𝐢𝐧 (𝟒 𝒕)}] + 𝟏𝟖 . It is dynamically stable as
𝟑
𝒉 = − 𝟒 < 𝟎. Like time-paths of Expected & Actual rates of inflation-growth {𝝅(𝒕), 𝒑(𝒕)}, 𝑼(𝒕)has
𝟏
a fluctuating time-path converging towards equilibrium value of ( ).
𝟏𝟖
______________________________________________________________________________
𝟏
𝑼 converges to the constant (𝟏𝟖) regardless of the rate of growth of nominal money (𝒎) and
hence regardless of the equilibrium rate of inflation. This constant equilibrium value of 𝑼 is referred
to as the natural rate of unemployment (𝑼𝑵𝑹 ).
Replotting, the Phillips Relation using 𝑼𝑵𝑹 . We get vertical straight line parallel to the 𝒑-axis.
That vertical line relating the equilibrium values of 𝑼 and 𝒑 to each other, is known as the long-
run Phillips curve.
______________________________________________________________________________
______________________________________________________________________________
Lesson 33
HIGHER ORDER DIFFERENTIAL EQUATIONS
TOPIC 147: DIFFERENTIAL EQUATION WITH VARIABLE TERM
Second Order Differential Equation with constant coefficient and constant term:
𝒚′′ (𝒕) + 𝒂
⏟𝟏 𝒚′ (𝒕) + 𝒂
⏟𝟐 𝒚= ⏟
𝒃
𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑡𝑒𝑟𝑚
What if: instead of 𝒃, there is a variable term {𝒇(𝒕)} in terms of 𝒕.
𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡−𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒−𝑡𝑒𝑟𝑚 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑙 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛
⏞
𝒚′′ (𝒕) + 𝒂
⏟𝟏 𝒚′ (𝒕) + 𝒂
⏟𝟐 𝒚 = 𝒇(𝒕)
⏟
𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 ⏟
𝒕𝒏 :𝑛≠0
⏟ 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 (𝒚𝒑 ) 𝑠ℎ𝑎𝑙𝑙
𝑛𝑜 𝑐ℎ𝑎𝑛𝑔𝑒 𝑜𝑛 𝑳.𝑯.𝑺 ⇒ (𝒚𝒄 ) 𝑟𝑒𝑚𝑎𝑖𝑛𝑠 𝑢𝑛𝑐ℎ𝑎𝑛𝑔𝑒𝑑, 𝑐ℎ𝑎𝑛𝑔𝑒
𝑎𝑠 𝑖𝑡 𝑎𝑙𝑤𝑎𝑦𝑠 𝑑𝑒𝑝𝑒𝑛𝑑𝑠 𝑜𝑛 𝒓𝟐 +𝒂𝟏 𝒓+𝒂𝟐 =𝟎 .
Solution: Such situation calls for a certain method, known as Method of Undetermined
Coefficients.
Pre-condition: Method of Undermined Coefficients requires that the variable term and its
successive derivatives together contain only a finite and distinct types of expressions:
Function & its Finite & Distinct Infinite & Repetitive Type of Expressions
Derivatives Type of
Expressions
𝒚(𝒕) 𝟔𝒕𝟐 − 𝒕 − 𝟏 𝟐𝒆−𝟒𝒕 𝒔𝒊𝒏(𝒕)
′ (𝒕) −𝟒𝒕
𝒚 𝟏𝟐𝒕 − 𝟏 −𝟖𝒆 𝒄𝒐𝒔(𝒕)
𝒚′′ (𝒕) 𝟏𝟐 𝟑𝟐𝒆−𝟒𝒕 −𝒔𝒊𝒏(𝒕)
𝒚′′′ (𝒕) 𝟎 −𝟏𝟐𝟖𝒆−𝟒𝒕 −𝒄𝒐𝒔(𝒕)
′′′′ (𝒕) … …
𝒚 −
Numerical Example: Find the Particular Integral of:

𝒚′′ (𝒕) + ⏟
𝟓 𝒚′ (𝒕) + ⏟ ⏟𝟐 −𝒕−𝟏
𝟑 𝒚(𝒕) = 𝟔𝒕
𝑎1 𝑎2 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒 𝑡𝑒𝑟𝑚
𝒇(𝒕)
Here, 𝒂𝟏 = 𝟓, 𝒂𝟐 = 𝟑 and 𝒇(𝒕) = 𝟔𝒕𝟐 − 𝒕 − 𝟏.
Here, particular integral (𝒚𝒑 ) is calculated differently due to change in 𝑳. 𝑯. 𝑺.
𝐿.𝐻.𝑆 𝑚𝑢𝑠𝑡
𝑏𝑒 𝑒𝑞𝑢𝑎𝑙 𝑅.𝐻.𝑆
⏞′′ (𝒕) ⏞
𝒚
⏟ + 𝟓 𝒚′ (𝒕) + 𝟑 𝒚(𝒕) =
⏞ 𝟔𝒕𝟐 − 𝒕 − 𝟏
⏟
𝑞𝑢𝑎𝑑𝑟𝑎𝑡𝑖𝑐
𝒚(𝒕),𝒚′(𝒕)& 𝒚′′(𝒕) 𝑖𝑛𝑡𝑒𝑟𝑚𝑠 𝑜𝑓 (𝒕)
𝑖𝑛𝑡𝑒𝑟𝑚𝑠 𝑜𝑓 (𝒕)
Both sides are in terms of (𝒕), however, the degree of polynomial on 𝑳. 𝑯. 𝑺 is not explicit.
Equality (=) requires that if 𝑹. 𝑯. 𝑺 is quadratic in terms of (𝒕), then 𝑳. 𝑯. 𝑺 should also be quadratic
in terms of (𝒕). So, we assume a (trial) general quadratic function that could satisfy
𝒚(𝒕), 𝒚′ (𝒕) & 𝒚′′ (𝒕) on 𝑳. 𝑯. 𝑺 to satisfy equality: 𝒚(𝒕) = 𝑩𝟏 𝒕𝟐 + 𝑩𝟐 𝒕 + 𝑩𝟑 .
Developing 𝒚′ (𝒕) and 𝒚′′ (𝒕) for the given ( 𝒚′′ (𝒕) + 𝟓 𝒚′ (𝒕) + 𝟑 𝒚(𝒕) = 𝟔𝒕𝟐 − 𝒕 − 𝟏).
𝒚′ (𝒕) = 𝟐𝑩𝟏 𝒕 + 𝑩𝟐
𝒚′′ (𝒕) = 𝟐𝑩𝟏
Since 𝒚(𝒕), 𝒚′ (𝒕) and 𝒚′′ (𝒕) are now symbolized, we substitute them on 𝑳. 𝑯. 𝑺:
______________________________________________________________________________
𝐿.𝐻.𝑆
𝑅.𝐻.𝑆
⏞′′ (𝒕)
𝒚 + 𝟓 𝒚′ (𝒕) + 𝟑 𝒚(𝒕) = ⏞
𝟔𝒕𝟐 − 𝒕 − 𝟏
𝐿.𝐻.𝑆 𝑅.𝐻.𝑆
⏞
𝟐𝑩𝟏 + 𝟓 𝟐𝑩𝟏 𝒕 + 𝑩𝟐 + 𝟑 𝑩𝟏 𝒕𝟐 + 𝑩𝟐 𝒕 + 𝑩𝟑 = ⏞
𝟔𝒕𝟐 − 𝒕 − 𝟏
⏞
𝟐𝑩𝟏 + 𝟏𝟎𝑩𝟏 𝒕 + 𝟓𝑩𝟐 + 𝟑𝑩𝟏 𝒕𝟐 + 𝟑𝑩𝟐 𝒕 + 𝟑𝑩𝟑 = ⏞
𝟔𝒕𝟐 − 𝒕 − 𝟏
⏞
𝟑𝑩𝟏 𝒕𝟐 + 𝟏𝟎𝑩𝟏 𝒕 + 𝟑𝑩𝟐 𝒕 + 𝟐𝑩𝟏 + 𝟓𝑩𝟐 + 𝟑𝑩𝟑 = ⏞
𝟔𝒕𝟐 − 𝒕 − 𝟏
⏞
(𝟑𝑩𝟏 )𝒕𝟐 + (𝟏𝟎𝑩𝟏 + 𝟑𝑩𝟐 )𝒕 + 𝟐𝑩𝟏 + 𝟓𝑩𝟐 + 𝟑𝑩𝟑 = ⏞
𝟔𝒕𝟐 − 𝒕 − 𝟏
⏞ ⏞
(𝟑𝑩𝟏 ) 𝒕𝟐 + (𝟏𝟎𝑩𝟏 + 𝟑𝑩𝟐 )𝒕 + (𝟐𝑩𝟏 + 𝟓𝑩𝟐 + 𝟑𝑩𝟑 ) = 𝟔 𝒕𝟐 + (−𝟏)𝒕 + (−𝟏)
Comparing the coefficients of 𝒕𝟐 & 𝒕 and the constants on both sides of equation:
(𝟑𝑩𝟏 ) = 𝟔 (𝟏𝟎𝑩𝟏 + 𝟑𝑩𝟐 ) = (−𝟏) (𝟐𝑩𝟏 + 𝟓𝑩𝟐 + 𝟑𝑩𝟑 ) = (−𝟏)
𝑩𝟏 = 𝟐 𝟏𝟎 𝟐 + 𝟑𝑩𝟐 = −𝟏 𝟐 𝟐 + 𝟓(−𝟕) + 𝟑𝑩𝟑 = −𝟏
𝑩𝟏 = 𝟐 𝟐𝟎 + 𝟑𝑩𝟐 = −𝟏 𝟒 − 𝟑𝟓 + 𝟑𝑩𝟑 = −𝟏
𝑩𝟏 = 𝟐 𝑩𝟐 = −𝟕 𝑩𝟑 = 𝟏𝟎
So, the Undetermined coefficients are now determined.
Substituting 𝑩𝟏 , 𝑩𝟐 and 𝑩𝟑 in 𝒚(𝒕) = 𝑩𝟏 𝒕𝟐 + 𝑩𝟐 𝒕 + 𝑩𝟑 assumed above as the (particular) solution.
⏟ = (𝟐)𝒕𝟐 + (−𝟕)𝒕 + (𝟏𝟎)
𝒚(𝒕)
𝑦𝑝
𝒚𝒑 = 𝟐𝒕𝟐 − 𝟕𝒕 + 𝟏𝟎
Interpretation: So, the particular integral of the ‘constant-coefficient and variable-term second
order differential equation’ 𝒚′′ (𝒕) + 𝟓𝒚′ (𝒕) + 𝟑𝒚(𝒕) = 𝟔𝒕𝟐 − 𝒕 − 𝟏 is 𝒚𝒑 = 𝟐𝒕𝟐 − 𝟕𝒕 + 𝟏𝟎 .
Method used to find (𝒚𝒑 ) is Method of Undetermined Coefficients. Whereas, (𝒚𝒄 ) can be found
using methods same as ‘constant-coefficient and constant-term second order differential
equation’.
TOPIC 148: DIFFERENTIAL EQUATION WITH VARIABLE TERM – A MODIFICATION
Sometimes, ‘constant-coefficient and variable-term second order differential equation’ {𝒚′′ (𝒕) +
𝒂𝟏 𝒚′ (𝒕) + 𝒂𝟐 𝒚 = 𝒇(𝒕)} lacks term(s) on 𝑳. 𝑯. 𝑺. Under this situation, the Method of Undetermined
Coefficient is applied with a Modification: We experiment with a numerical example:
𝒚′′ (𝒕) + 𝟓𝒚′ (𝒕) = 𝟔𝒕𝟐 − 𝒕 − 𝟏
Labelling the given differential equation:
𝒚′′ (𝒕) + ⏟ 𝟓 𝒚′ (𝒕) + ⏟ 𝟔𝒕𝟐 − 𝒕 − 𝟏
𝟎 𝒚(𝒕) = ⏟
𝒇(𝒕)
Here, 𝒂𝟏 = 𝟓, 𝒂𝟐 = 𝟎 and 𝒇(𝒕) = 𝟔𝒕𝟐 − 𝒕 − 𝟏.
However, particular integral (𝒚𝒑 ) is differently calculated due to change in 𝑳. 𝑯. 𝑺.
𝐿.𝐻.𝑆 𝑚𝑢𝑠𝑡
𝑏𝑒 𝑒𝑞𝑢𝑎𝑙 𝑅.𝐻.𝑆
⏞′′ (𝒕) ⏞
𝒚
⏟ + 𝟓 𝒚′ (𝒕) + 𝟎 𝒚(𝒕) =
⏞ 𝟔𝒕𝟐 − 𝒕 − 𝟏
⏟
𝒚(𝒕),𝒚′(𝒕)& 𝒚′′(𝒕) 𝑖𝑛𝑡𝑒𝑟𝑚𝑠 𝑜𝑓 (𝒕)
𝑖𝑛𝑡𝑒𝑟𝑚𝑠 𝑜𝑓 (𝒕)
______________________________________________________________________________
in terms of (𝒕). So, we assume a general quadratic function that could satisfy 𝒚(𝒕) on 𝑳. 𝑯. 𝑺 to
satisfy equality:
𝒚(𝒕) = 𝑩𝟏 𝒕𝟐 + 𝑩𝟐 𝒕 + 𝑩𝟑
Developing 𝒚′ (𝒕) and 𝒚′′ (𝒕) for the given ( 𝒚′′ (𝒕) + 𝟓 𝒚′ (𝒕) = 𝟔𝒕𝟐 − 𝒕 − 𝟏).
𝒚′ (𝒕) = 𝟐𝑩𝟏 𝒕 + 𝑩𝟐
𝒚′′ (𝒕) = 𝟐𝑩𝟏
𝐿.𝐻.𝑆
𝑅.𝐻.𝑆
⏞′′ (𝒕)
𝒚 + 𝟓 𝒚′ (𝒕) + 𝟎 𝒚(𝒕) = ⏞
𝟔𝒕𝟐 − 𝒕 − 𝟏
⏞
𝟐𝑩𝟏 + 𝟓 𝟐𝑩𝟏 𝒕 + 𝑩𝟐 + 𝟎 𝑩𝟏 𝒕𝟐 + 𝑩𝟐 𝒕 + 𝑩𝟑 = ⏞
𝟔𝒕𝟐 − 𝒕 − 𝟏
⏞ 𝟏 + 𝟏𝟎𝑩𝟏 𝒕 + 𝟓𝑩𝟐 + 𝟎 + 𝟎 + 𝟎 = ⏞
𝟐𝑩 𝟔𝒕𝟐 − 𝒕 − 𝟏
𝟏𝟎𝑩𝟏 𝒕 + 𝟐𝑩𝟏 + 𝟓𝑩𝟐 = ⏞
⏞ 𝟔𝒕𝟐 − 𝒕 − 𝟏
⏞
(𝟎)𝒕𝟐 + (𝟏𝟎𝑩𝟏 )𝒕 + 𝟐𝑩𝟏 + 𝟓𝑩𝟐 = ⏞
𝟔𝒕𝟐 − 𝒕 − 𝟏
⏞ 𝟐 ⏞
(𝟎) 𝒕 + (𝟏𝟎𝑩𝟏 )𝒕 + (𝟐𝑩𝟏 + 𝟓𝑩𝟐 ) = 𝟔 𝒕𝟐 + (−𝟏)𝒕 + (−𝟏)
Extracting the coefficients of 𝒕𝟐 & 𝒕 and the constants on both sides of equation:
(𝟎) = 𝟔
⏟ (𝟏𝟎𝑩𝟏 ) = (−𝟏) (𝟐𝑩𝟏 + 𝟓𝑩𝟐 ) = (−𝟏)
𝑰𝑳𝑳𝑶𝑮𝑰𝑪𝑨𝑳
Comparison gives an illogical coefficient, which rejects 𝒚(𝒕) = 𝑩𝟏 𝒕𝟐 + 𝑩𝟐 𝒕 + 𝑩𝟑 as the
(Particular) Solution of differential equation – need for Modification.
Modification: The trick is to experiment 𝒚(𝒕) = 𝒕 × (𝑩𝟏 𝒕𝟐 + 𝑩𝟐 𝒕 + 𝑩𝟑 ) as the trial (Particular)
Solution. The multiplicative term 𝒕 × increase the degree of 𝒚(𝑡), 𝒚′ (𝒕) and 𝒚′′ (𝒕) by 𝟏. In the last
comparison, 𝒕𝟐 was missing a coefficient. This might be fixed due to this modification.
So, we assume a (trial) general quadratic function (after modification) that could satisfy 𝒚(𝒕) on
𝑳. 𝑯. 𝑺 to satisfy equality: 𝒚(𝒕) = 𝒕 × (𝑩𝟏 𝒕𝟐 + 𝑩𝟐 𝒕 + 𝑩𝟑 )
𝒚(𝒕) = (𝑩𝟏 𝒕𝟑 + 𝑩𝟐 𝒕𝟐 + 𝑩𝟑 𝒕)
Developing 𝒚′ (𝒕) and 𝒚′′ (𝒕) for the given ( 𝒚′′ (𝒕) + 𝟓 𝒚′ (𝒕) = 𝟔𝒕𝟐 − 𝒕 − 𝟏).
𝒚′ (𝒕) = (𝟑𝑩𝟏 𝒕𝟐 + 𝟐𝑩𝟐 𝒕 + 𝑩𝟑 )
𝒚′′ (𝒕) = 𝟔𝑩𝟏 𝒕 + 𝟐𝑩𝟐
𝐿.𝐻.𝑆
𝑅.𝐻.𝑆
⏞′′ (𝒕)
𝒚 + 𝟓 𝒚′ (𝒕) + 𝟎 𝒚(𝒕) = ⏞
𝟔𝒕𝟐 − 𝒕 − 𝟏
⏞
𝟔𝑩𝟏 𝒕 + 𝟐𝑩𝟐 + 𝟓 (𝟑𝑩𝟏 𝒕𝟐 + 𝟐𝑩𝟐 𝒕 + 𝑩𝟑 ) + 𝟎 (𝑩𝟏 𝒕𝟑 + 𝑩𝟐 𝒕𝟐 + 𝑩𝟑 𝒕) = ⏞
𝟔𝒕𝟐 − 𝒕 − 𝟏
______________________________________________________________________________
⏞
𝟔𝑩𝟏 𝒕 + 𝟐𝑩𝟐 + 𝟏𝟓𝑩𝟏 𝒕𝟐 + 𝟏𝟎𝑩𝟐 𝒕 + 𝟓𝑩𝟑 = ⏞
𝟔𝒕𝟐 − 𝒕 − 𝟏
⏞
(𝟏𝟓𝑩𝟏 )𝒕𝟐 + (𝟔𝑩𝟏 + 𝟏𝟎𝑩𝟐 )𝒕 + (𝟐𝑩𝟐 + 𝟓𝑩𝟑 ) = ⏞ 𝟔𝒕𝟐 − 𝒕 − 𝟏
The Modification has served its purpose. As 𝒕𝟐 has non-zero coefficient. Now the comparison is
foolproof.
Extracting the coefficients of 𝒕𝟐 & 𝒕 and the constants on both sides of equation:
(𝟏𝟓𝑩𝟏 ) = 𝟔 (𝟔𝑩𝟏 + 𝟏𝟎𝑩𝟐 ) = (−𝟏) (𝟐𝑩𝟐 + 𝟓𝑩𝟑 ) = (−𝟏)
𝟔 𝟐 𝟏𝟕
𝑩𝟏 = 𝟏𝟓 𝟔 𝟓
+ 𝟏𝟎𝑩𝟐 = −𝟏 𝟐 (− 𝟓𝟎) + 𝟓𝑩𝟑 = −𝟏
𝟐 𝟏𝟐 𝟑𝟒
𝑩𝟏 = 𝟓 𝟓
+ 𝟏𝟎𝑩𝟐 = −𝟏 − 𝟓𝟎 − 𝟓𝑩𝟑 = −𝟏
𝟐 𝟏𝟕 𝟖
𝑩𝟏 = 𝑩𝟐 = − 𝑩𝟑 = −
𝟓 𝟓𝟎 𝟏𝟐𝟓
So, all of the Undetermined coefficients are now determined.
𝟐 𝟏𝟕 𝟖
⏟ = ( ) 𝒕𝟐 + (− ) 𝒕 + (−
𝒚(𝒕) )
𝟓 𝟓𝟎 𝟏𝟐𝟓
𝑦𝑝
𝟐 𝟏𝟕 𝟖
𝒚𝒑 = ( ) 𝒕𝟐 + (− ) 𝒕 + (− )
𝟓 𝟓𝟎 𝟏𝟐𝟓
Interpretation: So, the particular integral of the ‘constant-coefficient and variable-term second-
𝟐 𝟏𝟕 𝟖
order differential equation’ 𝒚′′ (𝒕) + 𝟓𝒚′ (𝒕) = 𝟔𝒕𝟐 − 𝒕 − 𝟏 is 𝒚𝒑 = ( ) 𝒕𝟐 + (− )𝒕 + (− ) .
𝟓 𝟓𝟎 𝟏𝟐𝟓
Method of Undetermined Coefficients was applied with a Modification which was to multiply the
trial (quadratic function) solution with (𝒕). Complementary Function (𝒚𝒄 ) can be found using
methods same as ‘constant-coefficient and constant-term second order differential equation’.
TOPIC 149: HIGHER-ORDER DIFFERENTIAL EQUATION: PARTICULAR INTEGRAL
After comprehending Second-Order Differential Equation, we can extend the order of the
differential equation to 𝒏𝒕𝒉 (higher) order. It is synonymous to an 𝒏𝒕𝒉 degree polynomial with
degree of polynomial replaced with order of derivative:
←
𝑜𝑟𝑑𝑒𝑟 𝑜𝑓 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑖𝑛𝑐𝑟𝑒𝑎𝑠𝑒𝑠 {𝒚(𝒕),𝒚′ (𝒕) 𝑡𝑜 𝒚𝒏 (𝒕)}
⏞
𝒂𝟎 𝒚𝒏 (𝒕) + 𝒂𝟏 𝒚(𝒏−𝟏) (𝒕) + ⋯ + 𝒂𝒏−𝟏 𝒚′ (𝒕) + 𝒂𝒏 𝒚 = 𝒃
⏟
𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡−𝑠𝑢𝑏𝑠𝑐𝑟𝑖𝑝𝑡 𝑖𝑛𝑐𝑟𝑒𝑎𝑠𝑒𝑠 (𝒂𝟎 𝑡𝑜 𝒂𝒏 )
→
Solution: For higher order differential equation, we need Particular Integral (𝒚𝒑 ) and
Complementary Function (𝒚𝒄 ).
Particular Integral (𝒚𝒑 ):
Since second order differential equation is also a higher order differential equation, we adopt the
Particular Integrals from it:
2nd Order Differential Equation 𝒏𝒕𝒉 Order Differential Equation
Order 𝒏=𝟐 𝒏>𝟐
𝒚𝒑 𝒂𝟐 ≠ 𝟎, 𝒂𝟐 = 𝟎, 𝒂𝟐 = 𝟎, 𝒂𝒏 ≠ 𝟎, 𝒂𝒏 = 𝟎, 𝒂𝒏 = 𝟎, 𝒂𝒏−𝟏 = 𝟎
(𝒂𝟏 ≠ 𝟎) 𝒂𝟏 ≠ 𝟎 𝒂𝟏 = 𝟎 (𝒂𝒏−𝟏 ≠ 𝟎) 𝒂𝒏−𝟏 ≠ 𝟎
{ ⋁ } { ⋁ }
(𝒂𝟏 = 𝟎) (𝒂𝒏−𝟏 = 𝟎)
______________________________________________________________________________
𝒃 𝒚𝒑 𝒚𝒑 𝒃 𝒚𝒑 Use 𝒚 = 𝒌𝒕𝟐 as trial

𝒚𝒑 = 𝒚𝒑 =
𝒂𝟐 𝒃 𝒃 𝒂𝒏 𝒃 solution, as used in
= 𝒕 = 𝒕𝟐 = 𝒕 SODE.
𝒂𝟏 𝟐 𝒂𝒏−𝟏
Numerical Example: Find the Particular Solution of:
𝒚(𝟒) (𝒕) + ⏟
𝟔 𝒚′′′ (𝒕) + 𝟏𝟒
⏟ 𝒚′′ (𝒕) + 𝟏𝟔
⏟ 𝒚′ (𝒕) + ⏟
𝟖 𝒚 = 𝟐𝟒
⏟
𝑎1 𝑎2 𝑎3 𝑎4 𝑏
Here; 𝒂𝟏 = 𝟔, 𝒂𝟐 = 𝟏𝟒, 𝒂𝒏−𝟏 = 𝒂𝟑 = 𝟏𝟔, 𝒂
⏟𝒏 = 𝒂𝟒 = 𝟖 ≠ 𝟎 and ⏟
𝒃 = 𝟐𝟒 ≠ 𝟎 .
𝑏 𝑛𝑜𝑛−ℎ𝑜𝑚𝑜𝑔𝑒𝑛𝑜𝑢𝑠
𝑦𝑝 = 𝑐𝑎𝑠𝑒
𝑎𝑛
𝒃
𝒚𝒑 =
𝒂𝟒
𝟐𝟒
𝒚𝒑 =
𝟖
𝒚𝒑 = 𝟑
Interpretation: The 4th (higher) order differential equation,
(𝟒) ′′′ ′′ ′
𝒚 (𝒕) + 𝟔𝒚 (𝒕) + 𝟏𝟒𝒚 (𝒕) + 𝟏𝟔𝒚 (𝒕) + 𝟖𝒚 = 𝟐𝟒 has a particular integral 𝒚𝒑 = 𝟑 . It is a static
equilibrium as 𝒚𝒑 is a constant value.
𝒚′′′ (𝒕) + 𝒚′′ (𝒕) + 𝟑𝒚′ (𝒕) = 𝟏
′′′ (𝒕)
𝒚 +⏟ 𝟏 𝒚′′ (𝒕) + ⏟𝟑 𝒚′ (𝒕) + ⏟
𝟎𝒚=⏟
𝟏
𝑎1 𝑎2 𝑎3 𝑏
Here; 𝒂𝟏 = 𝟏, 𝒂𝒏−𝟏 = 𝒂𝟐 = 𝟑, 𝒂
⏟𝒏 = 𝒂𝟑 = 𝟎 and 𝒃=𝟏≠𝟎
⏟ .
𝑏 𝑛𝑜𝑛−ℎ𝑜𝑚𝑜𝑔𝑒𝑛𝑜𝑢𝑠
𝑏= 𝑡 𝑐𝑎𝑠𝑒
𝑎𝑛−1
𝒃
𝒚𝒑 =
𝒕
𝒂𝒏−𝟏
𝟏
𝒚𝒑 = 𝒕
𝟑
Interpretation: The 3 (higher) order differential equation, 𝒚′′′ (𝒕) + 𝒚′′ (𝒕) + 𝟑𝒚′ (𝒕) = 𝟏 has a
rd
𝟏
particular integral 𝒚𝒑 = 𝟑 𝒕 . It is a moving equilibrium as 𝒚𝒑 contains variable (𝒕).
𝟏
𝒚′′′ (𝒕) + 𝟑𝒚′′ (𝒕) =
𝟑
𝟏
𝒚′′′ (𝒕) + ⏟
𝟑 𝒚′′ (𝒕) + ⏟
𝟎 𝒚′ (𝒕) + ⏟
𝟎𝒚=
⏟
𝟑
𝑎1 𝑎2 𝑎3
𝑏
Here; 𝒂𝟏 = 𝟑, 𝒂𝒏−𝟏 = 𝒂𝟐 = 𝟎,𝒂
⏟𝒏 = 𝒂𝟑 = 𝟎 and 𝒃=𝟏≠𝟎
⏟ .
𝒃 𝑛𝑜𝑛−ℎ𝑜𝑚𝑜𝑔𝑒𝑛𝑜𝑢𝑠
𝒃= 𝒕 𝑐𝑎𝑠𝑒
𝒂𝒏−𝟏
Experiment with trial solution 𝒚 = 𝒌𝒕𝟐 which implies 𝒚 ′ (𝒕)

= 𝟐𝒌𝒕, 𝒚′′ (𝒕) = 𝟐𝒌 and 𝒚′′′ (𝒕) = 𝟎.
Substituting all of these values in given equation will give us 𝒌:
𝟏
𝟎 + 𝟑 (𝟐𝒌) + (𝟎)𝟐𝒌𝒕 + (𝟎)𝒌𝒕𝟐 =
𝟑
𝟏
𝟔𝒌=
𝟑
𝟏
𝒌=
𝟏𝟖
______________________________________________________________________________
Returning back the value of 𝒌 in, 𝒚

⏟ = 𝒌 𝒕𝟐 :
𝑡𝑟𝑖𝑎𝑙
𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛
Since trial solution has satisfied the differential equation, it can, now, be called particular solution
of the differential equation.
𝟏 𝟐
𝒚
⏟ = 𝒕
𝑏𝑒𝑐𝑜𝑚𝑒𝑠
𝟏𝟖
𝑝𝑎𝑟𝑡𝑖𝑐𝑢𝑙𝑎𝑟
𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛 (𝒚𝒑 )
𝟏 𝟐
𝒚𝒑 = 𝒕
𝟏𝟖
𝟏
Interpretation: The 3rd (higher) order differential equation, 𝒚′′′ (𝒕) + 𝟑𝒚′′ (𝒕) = has a particular
𝟑
𝟏 𝟐
integral 𝒚𝒑 = 𝒕 . It is a moving equilibrium as 𝒚𝒑 contains variable (𝒕).
𝟏𝟖
TOPIC 150: HIGHER-ORDER DIFFERENTIAL EQUATION: COMPLEMENTARY FUNCTION
For Higher Order Differential equations, the complementary functions can be built upon those of
Second (Higher) Order Differential equations. We consider the following numerical example:
Numerical Example: Find the Complementary Function of:
𝒚(𝟒) (𝒕) + 𝟔𝒚′′′ (𝒕) + 𝟏𝟒𝒚′′ (𝒕) + 𝟏𝟔𝒚′ (𝒕) + 𝟖𝒚 = 𝟐𝟒
𝒚(𝟒) (𝒕) + ⏟
𝟔 𝒚′′′ (𝒕) + 𝟏𝟒
⏟ 𝒚′′ (𝒕) + 𝟏𝟔
⏟ 𝒚′ (𝒕) + ⏟
𝟖 𝒚 = 𝟐𝟒
⏟
𝑎1 𝑎2 𝑎3 𝑎4 𝑏
Here; 𝒂𝟏 = 𝟔, 𝒂𝟐 = 𝟏𝟒, 𝒂𝒏−𝟏 = 𝒂𝟑 = 𝟏𝟔, 𝒂𝒏 = 𝒂𝟒 = 𝟖 ≠ 𝟎 and ⏟
𝒃 = 𝟐𝟒 ≠ 𝟎 .
𝑛𝑜𝑛−ℎ𝑜𝑚𝑜𝑔𝑒𝑛𝑜𝑢𝑠
𝑐𝑎𝑠𝑒
For Second Order Differential equation, the characteristic equation was:
𝒓𝟐 + 𝒂𝟏 𝒓 + 𝒂𝟐 = 𝟎
Accordingly, for Fourth Order Differential equation, the characteristic equation will be:
𝒓𝟒 + 𝒂𝟏 𝒓𝟑 + 𝒂𝟐 𝒓𝟐 + 𝒂𝟑 𝒓 + 𝒂𝟒 = 𝟎
To find the roots (𝒓), we need to solve the characteristic equation, therefore, factorizing
𝒓𝟒 + 𝒂𝟏 𝒓𝟑 + 𝒂𝟐 𝒓𝟐 + 𝒂𝟑 𝒓 + 𝒂𝟒 = 𝟎, we get:
(𝒓
⏟ + 𝟐) × (𝒓 (𝒓𝟐 + 𝟐𝒓 + 𝟐) = 𝟎
⏟ − 𝟐) × ⏟
𝑙𝑖𝑛𝑒𝑎𝑟 𝑙𝑖𝑛𝑒𝑎𝑟 𝑞𝑢𝑎𝑑𝑟𝑎𝑡𝑖𝑐
𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛
Either;
𝑎 𝑏 𝑐
(𝒓 + 𝟐) = 𝟎 (𝒓 − 𝟐) = 𝟎
⏞ 𝒓𝟐 + ⏞
((𝟏) 𝟐𝒓+⏞
𝟐) = 𝟎
⇓ ⇓ −𝒃 ± √𝒃𝟐 − 𝟒𝒂𝒄
𝒓𝟑 , 𝒓𝟒 =
𝟐
⇓ ⇓ −𝟐 ± √𝟐𝟐 − 𝟒(𝟏)(𝟐)
𝒓𝟑 , 𝒓𝟒 =
𝟐
𝒓𝟏 = −𝟐 𝒓𝟐 = 𝟐 𝒓𝟑 , 𝒓𝟒 = −𝟏
⏟ ±⏟ 𝟐𝒊
ℎ 𝜐
Summarizing the (hybrid) group of roots:
𝒓𝟏 = −𝟐, 𝒓𝟐 = 𝟐 , ⏟
⏟ 𝒓𝟑 = −𝟏 + 𝟐𝒊, 𝒓𝟒 = −𝟏 − 𝟐𝒊
𝑑𝑖𝑠𝑡𝑖𝑛𝑐𝑡 𝑟𝑒𝑎𝑙 𝑟𝑜𝑜𝑡𝑠 𝑐𝑜𝑛𝑗𝑢𝑔𝑎𝑡𝑒 𝑐𝑜𝑚𝑝𝑙𝑒𝑥 𝑟𝑜𝑜𝑡𝑠
Accordingly, the complementary function will also be hybrid:
______________________________________________________________________________
𝐻𝑦𝑏𝑟𝑖𝑑 𝑛𝑎𝑡𝑢𝑟𝑒 𝑜𝑓 𝑐𝑜𝑚𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑟𝑦 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
𝒚𝒄 = ⏞ 𝑨
⏟𝟏 𝒆−𝒓𝟏𝒕 + 𝑨𝟐 𝒆−𝒓𝟐𝒕 ⏟𝒉𝒕 (𝑨𝟑 𝐜𝐨𝐬 𝒕 + 𝑨𝟒 𝐬𝐢𝐧 𝒕)
+𝒆
𝐷𝑖𝑠𝑡𝑖𝑛𝑐𝑡 𝑅𝑒𝑎𝑙 𝑟𝑜𝑜𝑡𝑠 𝑠𝑒𝑔𝑚𝑒𝑛𝑡 𝐶𝑜𝑚𝑝𝑙𝑒𝑥 𝑟𝑜𝑜𝑡𝑠 𝑠𝑒𝑔𝑚𝑒𝑛𝑡
Interpretation: Complementary Function in case of higher order differential equations can be
hybrid. Here, it is a combination of Distinct Real roots and Complex roots.
TOPIC 151: CONVERGENCE AND THE ROUTH THEOREM
In Higher Order Differential Equations, the roots are more than 2 as well
as the nature of roots can be hybrid. i.e. Distinct Real Roots, Repeated
Real Roots and Complex Roots. As the number of roots increase, the
complexity in the calculations also increase. A simpler method for
ascertaining the dynamic stability without getting entangled in calculation
of roots can be instrumental. English Mathematician Edward Routh
(1831-1907) did just that by giving a qualitative method of judging the
dynamic stability of higher order differential equations.
Limitation: Routh Theorem doesn’t allow to plot time path, as it doesn’t calculate the roots.
Formation: It is based on the Characteristic Equation of Higher Order Differential Equation:
(=1)
⏞𝟎 𝒓𝒏 + 𝒂𝟏 𝒓𝒏−𝟏 + 𝒂𝟐 𝒓𝒏−𝟐 + ⋯ + 𝒂𝒏−𝟏 𝒓 + 𝒂𝒏 = 𝟎
𝒂
←
𝑝𝑜𝑤𝑒𝑟 𝑜𝑓 𝑟𝑜𝑜𝑡 (𝒓) 𝑖𝑛𝑐𝑟𝑒𝑎𝑠𝑒𝑠
⏞
𝒂𝟎 𝒓𝒏 + 𝒂𝟏 𝒓𝒏−𝟏 + 𝒂𝟐 𝒓𝒏−𝟐 + ⋯ + 𝒂𝒏−𝟏 𝒓 + 𝒂𝒏 = 𝟎
⏟
𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡−𝑠𝑢𝑏𝑠𝑐𝑟𝑖𝑝𝑡 𝑖𝑛𝑐𝑟𝑒𝑎𝑠𝑒𝑠
→
All roots will be negative (𝒓𝟏 , 𝒓𝟐 , … , 𝒓𝒏 ) < 𝟎, only if first 𝒏 of the following determinants are
positive:
𝒂𝟏 𝒂𝟑 𝒂𝟓 𝒂𝟕
𝒂𝟏 𝒂𝟑 𝒂𝟓
𝒂 𝒂 𝒂𝟎 𝒂𝟐 𝒂𝟒 𝒂𝟔
|𝒂𝟏 |, |𝒂𝟏 𝒂𝟑 | , |𝒂𝟎 𝒂𝟐 𝒂𝟒 | , | 𝟎 𝒂 𝒂 𝒂 | , …
𝟎 𝟐 𝟏 𝟑 𝟓
𝟎 𝒂𝟏 𝒂𝟑
⏟ 𝟎 𝒂𝟎 𝒂𝟐 𝒂𝟒
𝑎𝑙𝑙 𝑑𝑒𝑡𝑒𝑟𝑚𝑖𝑛𝑎𝑛𝑡𝑠 𝑠ℎ𝑜𝑢𝑙𝑑 𝑏𝑒 𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒
Pattern of Development:
1. Number of Determinants = Order of Differential Equation. e.g. For 𝒏 = 𝟒, four

determinants shall be evaluated.
2. Red Arrows: A (𝑽) like shape occurs from principal diagonal to the last column. It contains
coefficients from 𝒂𝟏 , … .
3. Yellow Arrows: remaining elements are filled in such a way that vertically there exists an
ascending order from bottom to top.
4. Zero(s) (𝟎) are used where elements are below (𝒂𝟎 ).
______________________________________________________________________________
5. Subscript of coefficient can be greater than order of given differential equation. e.g. if order
of given differential equation is 𝟒. There will be 𝟒 determinants. However, in these 4
determinants, we can see elements with subscripts greater than 𝟒. i.e. 𝒂𝟓 , 𝒂𝟔 , 𝒂𝟕 . See final
column of (𝟒 × 𝟒) determinant.
Numerical Example 1: Apply Routh Theorem and find the dynamic stability of the time-path
(without calculating the characteristic roots):
𝒚(𝟒) (𝒕) + 𝟔𝒚′′′ (𝒕) + 𝟏𝟒𝒚′′ (𝒕) + 𝟏𝟔𝒚′ (𝒕) + 𝟖𝒚 = 𝟐𝟒
⏟ 𝒚(𝟒) (𝒕) + ⏟
(𝟏) 𝟔 𝒚′′′ (𝒕) + 𝟏𝟒
⏟ 𝒚′′ (𝒕) + 𝟏𝟔
⏟ 𝒚′ (𝒕) + ⏟
𝟖 𝒚 = 𝟐𝟒
⏟
𝑎0 𝑎1 𝑎2 𝑎3 𝑎4 𝑏
Here; 𝒂
⏟𝟎 = 𝟏 , 𝒂𝟏 = 𝟔, 𝒂𝟐 = 𝟏𝟒, 𝒂𝟑 = 𝟏𝟔, 𝒂𝟒 = 𝟖 and 𝒃 = 𝟐𝟒
𝑜𝑓 𝒚(𝟒) (𝒕)
Routh Theorem States:
𝒂𝟏 𝒂𝟑 𝒂𝟓 𝒂𝟕
𝒂𝟏 𝒂𝟑 𝒂𝟓
𝒂 𝒂𝟑 𝒂 𝒂𝟐 𝒂𝟒 𝒂𝟔
|𝒂𝟏 |, |𝒂𝟏 𝒂 𝒂𝟐 𝒂𝟒 | , | 𝟎
𝟎 𝒂𝟐 | , | 𝟎 𝟎 𝒂𝟏 𝒂𝟑 𝒂𝟓 | , …
𝟎 𝒂𝟏 𝒂𝟑
⏟ 𝟎 𝒂𝟎 𝒂𝟐 𝒂𝟒
𝒂𝟏 𝒂𝟑 𝒂𝟓 𝒂𝟕
𝒂𝟏 𝒂𝟑 𝒂𝟓 𝒂𝟎 𝒂𝟐 𝒂𝟒 𝒂𝟔 |
𝒂 𝒂𝟑
|𝒂𝟏 |, |𝒂𝟏 𝒂𝟐 | , |𝒂𝟎 𝒂𝟐 𝒂𝟒 | , ||
𝟎
𝟎 𝒂𝟏 𝒂𝟑 𝟎 𝒂𝟏 𝒂𝟑 𝒂𝟓 |
⏟ 𝟎 𝒂𝟎 𝒂𝟐 𝒂𝟒
𝐹𝑜𝑟 𝑜𝑟𝑑𝑒𝑟 (𝟒),𝑜𝑛𝑙𝑦 𝒇𝒐𝒖𝒓 𝑑𝑒𝑡𝑒𝑟𝑚𝑖𝑛𝑎𝑛𝑡𝑠 𝑠ℎ𝑎𝑙𝑙 𝑏𝑒 𝑒𝑣𝑎𝑙𝑢𝑎𝑡𝑒𝑑
Substituting the coefficient values and aligning vertically:
|𝒂𝟏 | = |𝟔| = 𝟔 > 𝟎
𝒂𝟏 𝒂𝟑 𝟔 𝟏𝟔
|𝒂 𝒂 | = | | = 𝟔𝟖 > 𝟎
𝟎 𝟐 𝟏 𝟏𝟒
𝒂𝟏 𝒂𝟑 𝒂𝟓 𝟔 𝟏𝟔 𝟎
|𝒂𝟎 𝒂𝟐 𝒂𝟒 | = |𝟏 𝟏𝟒 𝟖 | = 𝟖𝟎𝟎 > 𝟎
𝟎 𝒂𝟏 𝒂𝟑 𝟎 𝟔 𝟏𝟔
𝟎 is substituted 𝒂𝟓 as this coefficient doesn’t exist in the given differential equation.
𝒂𝟏 𝒂𝟑 𝒂𝟓 𝒂𝟕 𝟔 𝟏𝟔 𝟎 𝟎
| 𝒂𝟎 𝒂 𝟐 𝒂𝟒 𝒂 𝟔 | = |𝟏 𝟏𝟒 𝟖 𝟎 | = 𝟔𝟒𝟎𝟎 > 𝟎
|𝟎 𝒂 𝒂 𝒂 | | |
𝟏 𝟑 𝟓 𝟎 𝟔 𝟏𝟔 𝟎
𝟎 𝒂𝟎 𝒂𝟐 𝒂𝟒 𝟎 𝟏 𝟏𝟒 𝟖
𝟎 is substituted 𝒂𝟓 , 𝒂𝟔 , 𝒂𝟕 as these coefficients don’t exist in the given differential equation.
Interpretation: For the given Fourth order differential equation, four determinants were
evaluated under Routh Theorem. All turned out to be positive (𝟔, 𝟔𝟖, 𝟖𝟎𝟎, 𝟔𝟒𝟎𝟎). Hence, without
finding the roots, we have found that the time-path of given differential equation
𝒚(𝟒) (𝒕) + 𝟔𝒚′′′ (𝒕) + 𝟏𝟒𝒚′′ (𝒕) + 𝟏𝟔𝒚′ (𝒕) + 𝟖𝒚 = 𝟐𝟒 , is dynamically stable.
Numerical Example 2: Apply Routh Theorem and find the dynamic stability of the time-path
(without calculating the characteristic roots):
𝒚′′′ (𝒕) − 𝟐𝒚′′ (𝒕) − 𝒚′ (𝒕) + 𝟐𝒚 = 𝟒
′′′
(𝟏)
⏟ 𝒚 (𝒕) + (−𝟐)⏟ 𝒚′′ (𝒕) + (−𝟏)
⏟ 𝒚′ (𝒕) + ⏟
𝟐𝒚=⏟
𝟒
𝑎0 𝑎1 𝑎2 𝑎3 𝑏
______________________________________________________________________________
𝐻𝑎𝑐𝑘
Here; 𝒂 ⏞
⏟𝟎 = 𝟏 , 𝒂𝟏 = −𝟐
⏟ , 𝒂𝟐 = −𝟏, 𝒂𝟑 = 𝟐 and 𝒃 = 𝟒
𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝒂𝟏 =1𝑠𝑡 𝑑𝑒𝑡𝑒𝑟𝑚𝑖𝑛𝑎𝑛𝑡
𝑜𝑓 𝒚′′′ (𝒕) |𝑎 |<0⇒𝐷𝑦𝑛𝑎𝑚𝑖𝑐𝑎𝑙𝑙𝑦 𝑈𝑛𝑠𝑡𝑎𝑏𝑙𝑒
1
𝒂𝟏 𝒂𝟑 𝒂𝟓 𝒂𝟕
𝒂𝟏 𝒂𝟑 𝒂𝟓
𝒂𝟏 𝒂𝟑 𝒂 𝒂𝟐 𝒂𝟒 𝒂𝟔
Routh Theorem States: |𝒂𝟏 |, |𝒂 𝒂 𝒂𝟐 𝒂𝟒 | , | 𝟎
𝟎 𝒂𝟐 | , | 𝟎 𝟎 𝒂𝟏 𝒂𝟑 𝒂𝟓 | , …
𝟎 𝒂𝟏 𝒂𝟑
⏟ 𝟎 𝒂𝟎 𝒂𝟐 𝒂𝟒
(−𝒗𝒆) 𝒂𝟏 𝒂𝟑 𝒂𝟓
𝒂 𝒂𝟑
⏞𝟏 | , |𝒂𝟏
| 𝒂 𝒂𝟐 | , |𝒂𝟎 𝒂𝟐 𝒂𝟒 |
𝟎
⏟ 𝟎 𝒂𝟏 𝒂𝟑
𝐹𝑜𝑟 𝑜𝑟𝑑𝑒𝑟 (𝟑),𝑜𝑛𝑙𝑦 𝒕𝒉𝒓𝒆𝒆 𝑑𝑒𝑡𝑒𝑟𝑚𝑖𝑛𝑎𝑛𝑡𝑠 𝑠ℎ𝑎𝑙𝑙 𝑏𝑒 𝑒𝑣𝑎𝑙𝑢𝑎𝑡𝑒𝑑
Since 𝒂𝟏 gives the value of 1st determinant and it is −𝒗𝒆. One can already guess that time-path
is dynamically unstable due to negativity of one (i.e. 1st) determinant.
Nevertheless, we evaluate all of the determinants to put a stamp of validity on this hack.
Substituting the coefficient values and aligning vertically:
|𝒂𝟏 | = |−𝟐| = −𝟐 < 𝟎
𝒂𝟏 𝒂𝟑 −𝟐 𝟐
|𝒂 𝒂 | = | |= 𝟎≯𝟎
𝟎 𝟐 𝟏 −𝟏
𝒂𝟏 𝒂𝟑 𝒂𝟓 −𝟐 𝟐 𝟎
|𝒂𝟎 𝒂𝟐 𝒂𝟒 | = | 𝟏 −𝟏 𝟎 | = 𝟎 ≯ 𝟎
𝟎 𝒂𝟏 𝒂𝟑 𝟎 −𝟐 𝟐
𝟎 is substituted 𝒂𝟒 and 𝒂𝟓 as these coefficients don’t exist in the given differential equation.
Interpretation: For the given Third order differential equation, three determinants were
evaluated under Routh Theorem. 1st is negative, while 2nd and 3rd are non-positive(= 𝟎).
Precisely speaking, the value of determinants are (−𝟐, 𝟎, 𝟎). Hence, without finding the roots, we
have found that the time-path of given differential equation 𝒚′′′ (𝒕) − 𝟐𝒚′′ (𝒕) − 𝒚′ (𝒕) + 𝟐𝒚 = 𝟒 , is
______________________________________________________________________________
Lesson 34
FIRST ORDER DIFFERENCE EQUATIONS
TOPIC 152: DISCRETE TIME: FIRST ORDER DIFFERENCE EQUATION
Till now, dynamic analysis has been conducted assuming time as a continuous variable.
However, there can situations in which time might be taken as non-continuous (discrete) variable.
Now, variable 𝒕 will be allowed to take integer values only.
Comparison of Continuous Time and Discrete Time Analysis:
Criterion Continuous Time Analysis Discrete Time Analysis
Etymology (Latin Origin) Uninterrupted Separate
Nature of Time (𝒕) (𝒕) is a continuous variable (𝒕) is a discrete variable
Nature of Values Decimals as well as Integers Integers only
Successive Gaps Not Necessarily Equal Equal
Possible Values of (𝒕) 𝟎, 𝟎. 𝟓, 𝟏, 𝟏. 𝟔, 𝟐, 𝟐. 𝟗, … 𝟎, 𝟏, 𝟐, 𝟑, 𝟒, 𝟓, …
AKA Point Analysis Period Analysis
Tool of Analysis Derivatives & Differentials Difference, Lags & Leads
Symbolization 𝒅(𝒚) 𝜹(𝒚) ∆(𝒚𝒕 ), (𝒚𝒕−𝟏 ), (𝒚𝒕+𝟏 )
𝒅(𝒚), {𝒅𝒕
}, 𝜹(𝒚), { }
𝜹𝒕
Type of Equation Differential Equations Difference Equations
Examples Stock Market Activity Price of Seasonal crops.
Exclusive Features of Period Analysis:

Period doesn’t necessarily relate to calendar rather than analytical sense.
A period is merely a length of time that elapses before the variable 𝒚 undergoes a change.
∆𝒚 𝒅𝒚
(∆𝒙) is the discrete-time counterpart of (𝒅𝒙).
Consecutive time periods ⇒ ∆𝒕 = 𝟏 or First Difference of 𝒚.
∆𝒚𝒕 ≡ 𝒚𝒕+𝟏 − 𝒚𝒕
Symbology:
Use of subscript 𝒕 for discrete-time.
𝒚𝒕 is the value of 𝒚 in time period 𝒕.
𝒚𝒕+𝟏 is the value of 𝒚 in time period immediately following 𝒕.
𝒚𝒕+𝟏 is known as ‘lead’ variable of 𝒚𝒕 .
𝒚𝒕−𝟏 is known as ‘lag’ variable of 𝒚𝒕 .
Common Terminology: Unlike book of A. C. Chiang, most widely used terminology is not
Difference Equations, rather its Recurrence Relations.
Examples:
∆𝒚𝒕 = 𝟐 ∆𝒚𝒕 = −𝟎. 𝟏 𝒚𝒕
(instead of 𝒅𝒚 = 𝟐) (instead of 𝒅𝒚 = −𝟎. 𝟏 𝒚)
Equivalent Forms
∆𝒚𝒕 = 𝟐 ∆𝒚𝒕 = −𝟎. 𝟏 𝒚𝒕
𝒚𝒕+𝟏 − 𝒚𝒕 = 𝟐 𝒚𝒕+𝟏 − 𝒚𝒕 = −𝟎. 𝟏 𝒚𝒕
𝒚𝒕+𝟏 = 𝒚𝒕 + 𝟐 𝒚𝒕+𝟏 − 𝒚𝒕 + 𝟎. 𝟏 𝒚𝒕 = 𝟎
𝒚𝒕+𝟏 − 𝟎. 𝟗 𝒚𝒕 = 𝟎
Standard Form:
𝒚𝒕+𝟏 + 𝒂𝒚𝒕 = 𝒄
Here, 𝒂 and 𝒄 are two constants.
______________________________________________________________________________
𝒚𝒕 is the value of the 𝒚 in current time period 𝒕.

𝒚𝒕+𝟏 is the value of the 𝒚 in next time period (𝒕 + 𝟏).
Solution Methods: Difference equations can be solved either by Iterative (Crude) Method or by
General (Formula) Method.
TOPIC 153: SOLVING DIFFERENCE EQUATION USING ITERATIVE METHOD
Like differential equation, difference equations can also be solved to get time-path. A crude
method of solving the first order difference equation involves repetition. It can be understood with
the help of an example.
Example: Solve the following difference equations by iteration, while 𝒚𝟎 = 𝟏𝟎:
𝒚𝒕+𝟏 = 𝒚𝒕 + 𝟏
Solution: Experimenting different values of 𝒕 in the given differential equation:
𝒚 𝒕 +𝟏 = 𝒚 𝒕 + 𝟏
Value of 𝒕 Substitution Simplification Final Form
𝒕=𝟎 𝒚 𝟎 +𝟏 = 𝒚𝟎 + 𝟏 𝒚𝟏 = 𝒚𝟎 + 𝟏 𝒚 𝟏 = 𝒚𝟎 + 𝟏
𝒕=𝟏 𝒚 𝟏 +𝟏 = 𝒚𝟏
⏟ + 𝟏 𝒚𝟐 = 𝒚𝟎 + 𝟏 + 𝟏 𝒚 𝟐 = 𝒚𝟎 + 𝟐
𝑦1 =𝑦0 +1
𝒕=𝟐 𝒚 𝟐 +𝟏 = 𝒚𝟐
⏟ + 𝟏 𝒚𝟑 = 𝒚𝟎 + 𝟐 + 𝟏 𝒚 𝟑 = 𝒚𝟎 + 𝟑
𝑦2 =𝑦0 +2
𝒕=𝟑 𝒚 𝟑 +𝟏 = 𝒚𝟑
⏟ + 𝟏 𝒚𝟒 = 𝒚𝟎 + 𝟑 + 𝟏 𝒚 𝟒 = 𝒚𝟎 + 𝟒
𝑦3 =𝑦0 +3
… … … …
𝒕=𝒕 Generalizing, we get: 𝒚 𝒕 = 𝒚𝟎 + 𝒕
Interpretation: So, we have found the Definite Solution 𝒚⏟𝒕 = 𝒚𝟎 + 𝒕 of the given
𝑐𝑢𝑟𝑟𝑒𝑛𝑡
𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 (𝒚)
difference equation 𝒚
⏟𝒕+𝟏 = 𝒚𝒕 + 𝟏 . We can also find its numerical value by substituting the
𝑓𝑢𝑡𝑢𝑟𝑒
initial condition (𝒚𝟎 = 𝟏𝟎).
𝒚𝒕 = 𝒚𝟎 + 𝒕
𝒚𝒕 = 𝟏𝟎 + 𝒕
𝒚𝒕 = 𝟏𝟎 + 𝒕
Example: Solve the following difference equations by iteration, while 𝒚𝟎 = 𝜷:
𝒚𝒕+𝟏 = 𝜶𝒚𝒕
Solution: Experimenting different values of 𝒕 in the given differential equation:
𝒚 𝒕 +𝟏 = 𝜶𝒚 𝒕
𝒕=𝟎 𝒚 𝟎 +𝟏 = 𝜶𝒚𝟎 𝒚𝟏 = 𝜶𝒚𝟎 𝒚 𝟏 = 𝜶 𝟏 𝒚𝟎
𝒕=𝟏 𝒚 𝟏 +𝟏 = 𝜶 ⏟
𝒚𝟏 𝒚𝟐 = 𝜶 𝜶𝒚𝟎 𝒚 = 𝜶 𝟐 𝒚𝟎
𝟐
𝑦1 =𝛼𝑦0
𝒕=𝟐 𝒚 𝟐 +𝟏 = 𝜶 ⏟
𝒚𝟐 𝒚𝟑 = 𝜶 𝜶𝟐 𝒚𝟎 𝒚 𝟑 = 𝜶 𝟑 𝒚𝟎
𝑦2 =𝛼 2 𝑦0
______________________________________________________________________________
𝒕=𝟑 𝒚 𝟑 +𝟏 = 𝜶 ⏟
𝒚𝟑 𝒚𝟒 = 𝜶 𝜶𝟑 𝒚𝟎 𝒚 𝟒 = 𝜶 𝟒 𝒚𝟎
𝑦3 =𝛼 3 𝑦0
… … … …
𝒕=𝒕 Generalizing, we get: 𝒚 𝒕 = 𝜶 𝒕 𝒚𝟎
Interpretation: So, we have found the Definite Solution 𝒚⏟𝒕 = 𝜶 𝒕 𝒚𝟎 of the given
⏟𝒕+𝟏 = 𝜶𝒚𝒕 . We can also find its numerical value by substituting the
initial condition (𝒚𝟎 = 𝜷).
𝒚𝒕 = 𝜶𝒕 𝒚𝟎
𝒚𝒕 = 𝜶𝒕 𝜷
𝒚𝒕 = 𝜶𝒕 𝜷
TOPIC 154: SOLVING DIFFERENCE EQUATION USING GENERAL METHOD: A≠-1
In addition to Iterative Method, we can use a General (Formula) Method to solve difference
equations to get their time-paths. General Solution contains two components:
𝒚⏟𝒕 = 𝒚𝒄
⏟ + 𝒚
⏟𝒑
𝑒𝑞𝑢𝑖𝑙𝑖𝑏𝑟𝑖𝑢𝑚 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝒚𝒕
𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝒚𝒕
Homogenous Case: Consider the following homogeneous difference equation:

𝒎𝒚𝒕+𝟏 − 𝒏𝒚𝒕 = 𝟎
Normalizing it, we get:
𝒏
𝒚𝒕+𝟏 − ( ) 𝒚𝒕 = 𝟎
𝒎
Rearranging to apply iteration method:
𝒏
𝒚𝒕+𝟏 = ( ) 𝒚𝒕
𝒎
Experimenting different values of 𝒕 in the given differential equation:
𝒏
𝒚 𝒕 +𝟏 = ( ) 𝒚 𝒕
𝒎
𝒕=𝟎 𝒏 𝒏 𝒏 𝟏
𝒚 𝟎 +𝟏 = ( ) 𝒚𝟎 𝒚𝟏 = ( ) 𝒚𝟎 𝒚 𝟏 = ( ) 𝒚𝟎
𝒎 𝒎 𝒎
𝒕=𝟏 𝒏 𝒏 𝒏 𝒏 𝟐
𝒚 𝟏 +𝟏 = ( ) ⏟ 𝒚𝟏 𝒚𝟐 = ( ) ( )𝒚 𝒚 𝟐 = ( ) 𝒚𝟎
𝒎
𝒏 𝒎 𝒎 𝟎 𝒎
𝑦1 =( )𝑦0
𝒎
𝒕=𝟐 𝒏 𝒏 𝒏 𝟐 𝒏 𝟑
𝒚 𝟐 +𝟏 =( ) 𝒚𝟐
⏟ 𝒚𝟑 = ( ) ( ) 𝒚𝟎 𝒚 𝟑 = ( ) 𝒚𝟎
𝒎 𝒎 𝒎 𝒎
𝒏 2
𝑦2 =( ) 𝑦0
𝒎
______________________________________________________________________________
𝒕=𝟑 𝒏 𝒏 𝟒
𝒚 𝟑 +𝟏 = ( ) 𝒚𝟑 𝒏 𝒏 𝟑
𝒎 ⏟ 𝒚𝟒 = ( ) ( ) 𝒚𝟎 𝒚 𝟒 = ( ) 𝒚𝟎
𝒏 3 𝒎 𝒎 𝒎
𝑦3 =( ) 𝑦0
𝒎
… … … …
𝒕=𝒕 Generalizing, we get: 𝒏 𝒕
𝒚 𝒕 = ( ) 𝒚𝟎
𝒎
𝒏 𝒕
Interpretation: So, we have found the Definite Solution 𝒚⏟𝒕 = (𝒎) 𝒚𝟎 of the given
𝒏
⏟𝒕+𝟏 = (𝒎) 𝒚𝒕 .
𝒏 𝒕
Consolidating the 𝒚𝒕 = ( ) 𝒚𝟎 into 𝒚𝒕 = (𝒃)𝒕 𝑨 or 𝒚𝒕 = 𝑨𝒃𝒕 .
𝒎
𝒚𝒕 = 𝑨𝒃𝒕 serves the same role as 𝒚𝒕 = 𝑨𝒆−𝒓𝒕 did in the differential equations.
(𝒓) played a pivotal role in the dynamic stability of the time-path of a differential equation. In case
of difference equations, (𝒃) plays the same role.
Recalling the standard form of first order difference equation:
Complementary Function: Reducing the standard form of first order difference equation to
(homogeneous) Reduced form to get the complementary function:
𝒚𝒕+𝟏 + 𝒂 𝒚𝒕 = 𝟎
As we found 𝒚𝒕 = 𝑨𝒃𝒕 , we can also guess its lead form, 𝒚𝒕 +𝟏 = 𝑨𝒃𝒕 +𝟏
𝒚𝒕+𝟏 + 𝒂 𝒚𝒕 = 𝟎
Substituting values of 𝒚𝒕 and 𝒚𝒕+𝟏 :
𝑨𝒃𝒕+𝟏 + 𝒂 𝑨𝒃𝒕 = 𝟎
𝑨𝒃𝒕+𝟏 + 𝒂𝑨𝒃𝒕 = 𝟎
𝑨𝒃𝒕 (𝒃 + 𝒂) = 𝟎
𝒃+𝒂=𝟎
𝒃 = −𝒂
As the General Solution of reduced form of complete equation is the complementary function, i.e.
(𝒚𝒕 )𝐶.𝐸
𝑅.𝐹 = 𝒚𝒄 .
Returning the value of 𝒃 in 𝒚⏟𝒕 = 𝑨𝒃𝒕 , we get:
𝑦𝑐
𝒚𝒄 = 𝑨(−𝒂)𝒕
Particular Integral: It is a solution that satisfies the difference equation, so we experiment with
the simplest (constant) trial solution, i.e. 𝒚𝒕 = 𝒌.
𝒚𝒕 = 𝒌
Extending the time period by 𝟏, leaves constant unchanged:
𝒚𝒕+𝟏 = 𝒌
Substituting 𝒚𝒕 and 𝒚𝒕+𝟏 in standard form of first order difference equation:
𝒌 + 𝒂𝒌 = 𝒄
𝒌(𝟏 + 𝒂) = 𝒄
______________________________________________________________________________
𝒄
⏟
𝒌 =
𝑤𝑎𝑠 𝑜𝑝𝑡𝑒𝑑 𝑎𝑠 𝟏+𝒂
𝑎 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛 𝑜𝑓 𝒚𝒕
𝒄
𝒚𝒑 =
𝟏+𝒂
General Solution: Summing the expressions of 𝒚𝒄 and 𝒚𝒑 :
𝒚𝒕 = 𝒚𝒄 + 𝒚𝒑
𝒄
𝒚𝒕 = 𝑨(−𝒂)𝒕 +
𝟏+𝒂
Interpretation: We have derived the General Solution of First Order Difference equation, with a
parametric restriction of 𝒂 ≠ −𝟏. As if 𝒂 ≠ −𝟏 is allowed then the general solution will become
𝒄
undefined: 𝒚𝒕 = 𝑨(−𝒂)𝒕 + .
𝟏+ 𝒂⏟
≠−𝟏
Definite Solution: To find the definite solution, we introduce the initial condition, (𝒕 = 𝟎) in the
𝒄
general solution, 𝒚 𝒕 = 𝑨(−𝒂) 𝒕 + 𝟏+𝒂.
𝒄
𝒚 𝟎 = 𝑨(−𝒂) 𝟎 +
𝟏+𝒂
𝒄
𝒚𝟎 = 𝑨 × (𝟏) +
𝟏+𝒂
𝒄
𝒚𝟎 − =𝑨
𝟏+𝒂
𝒄
𝑨 = 𝒚𝟎 −
𝟏+𝒂
𝒄
Returning the value of arbitrary constant (𝑨) in the general solution 𝒚𝒕 = 𝑨 (−𝒂)𝒕 + 𝟏+𝒂 to get
the definite solution:
𝒄 𝒄
(−𝒂)𝒕 +
𝒚𝒕 = 𝒚𝟎 −
𝟏+𝒂 𝟏+𝒂
𝒄 𝒄
𝒚𝒕 = (𝒚𝟎 − ) (−𝒂)𝒕 +
𝟏+𝒂 𝟏+𝒂
Interpretation: We have derived the Definite Solution of First Order Difference equation, with a
parametric restriction of 𝒂 ≠ −𝟏. As if 𝒂 ≠ −𝟏 is allowed then the definite solution will become
𝒄 𝒄
undefined: 𝒚𝒕 = (𝒚𝟎 − ) (−𝒂)𝒕 + .
𝟏+ 𝒂⏟ 𝟏+ 𝒂⏟
≠−𝟏 ≠−𝟏
TOPIC 155: DIFFERENCE EQUATION USING GENERAL METHOD: A≠-1 NUMERICAL
Solve the first-order difference equation:

𝟕
𝒚𝒕+𝟏 − 𝟓𝒚𝒕 = 𝟏 (𝒚𝟎 = 𝟒)
Recalling standard form of First-Order Difference Equation:
𝒚𝒕+𝟏 − 𝒂𝒚𝒕 = 𝒄
Comparing the given specific form, we extract values:
𝒚𝒕+𝟏 + (−𝟓)
⏟ 𝒚𝒕 = ⏟
𝟏
𝑎 𝑐
Here, ⏟
𝒂 = −𝟓 ≠ −𝟏 , 𝒄 = 𝟏.
𝒄
𝒚𝒑 = 𝑖𝑠 𝑠𝑢𝑖𝑡𝑎𝑏𝑙𝑒
𝟏+𝒂
______________________________________________________________________________
Definite Solution: The equation can be solved for the definite solution of time-path by using the
formula:
𝑑𝑒𝑓𝑖𝑛𝑖𝑡𝑖𝑧𝑒𝑑 𝑓𝑜𝑟𝑚 𝑜𝑓
𝑎𝑟𝑏𝑖𝑡𝑟𝑎𝑟𝑦 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 (𝑨)
⏞
𝒄 𝒄
𝒚𝒕 = ( 𝒚
⏟𝟎 − ) (−𝒂)𝒕 +
𝟏+𝒂 ⏟
𝟏+𝒂
⏟ 𝑔𝑖𝑣𝑒𝑛 𝑦 𝑝
𝑦𝑐
𝟕 𝟏 𝟏
𝒚𝒕 = { − } {−(−𝟓)}𝒕 +
𝟒 𝟏 + (−𝟓) 𝟏 + (−𝟓)
𝟕 𝟏 𝟏
𝒚𝒕 = ( − ) (𝟓)𝒕 +
𝟒 𝟏−𝟓 𝟏−𝟓
𝟕 𝟏 𝒕
𝟏
𝒚𝒕 = ( − ) (𝟓) +
𝟒 −𝟒 −𝟒
𝟕 𝟏 𝒕
𝟏
𝒚𝒕 = ( + ) (𝟓) −
𝟒 𝟒 𝟒
𝟏
𝒚𝒕 = (𝟐)(𝟓)𝒕 −
𝟒
𝒕
𝟏
𝒚𝒕 = ⏟
𝟐(𝟓) + (− )
⏟ 𝟒
𝑦 𝑐
𝑦𝑝
Interpretation: The definite solution of the first order difference equation 𝒚𝒕+𝟏 − 𝟓𝒚𝒕 = 𝟏 is
𝟏 𝟏
𝒚𝒕 = 𝟐(𝟓)𝒕 − . 𝒚𝒑 = , represents a static equilibrium.
𝟒 𝟒
______________________________________________________________________________
TOPIC 156: SOLVING DIFFERENCE EQUATION USING GENERAL METHOD: A=-1
General Solution of First Order Difference equation, with a parametric restriction of 𝒂 ≠ −𝟏 is

𝒄
𝒚𝒕 = 𝑨(−𝒂)𝒕 + 𝟏+ ⏟
𝒂
. However, if 𝒂 = −𝟏 is allowed then the general solution will need revisit.
≠−𝟏
Complementary Function: As found before, 𝒚𝒄 = 𝑨(−𝒂)𝒕 . If we allow 𝒂 = −𝟏 in it, 𝒚𝒄 will not

breakdown. So, we can use the same 𝒚𝒄 even if 𝒂 = −𝟏 Reducing
______________________________________________________________________________
Particular Integral: The simplest (constant) trial solution, should not be useful as it was with 𝒂 ≠
−𝟏. Therefore, we try next possibility (a variable function). i.e. 𝒚𝒕 = 𝒌𝒕.
𝒚𝒕 = 𝒌𝒕
Extending the time period by 𝟏:
𝒚𝒕+𝟏 = 𝒌(𝒕 + 𝟏)
Substituting 𝒚𝒕 and 𝒚𝒕+𝟏 in standard form of first order difference equation:
𝒌(𝒕 + 𝟏) + 𝒂𝒌𝒕 = 𝒄
𝒌𝒕 + 𝒌 + 𝒂𝒌𝒕 = 𝒄
𝒌(𝒕 + 𝟏 + 𝒂𝒕) = 𝒄
𝒄
𝒌=
𝒕+𝟏+ 𝒂 ⏟ 𝒕
=−1
𝒄
𝒌=
𝒕 + 𝟏 + (−𝟏)𝒕
𝒄
𝒌=
𝒕+𝟏−𝒕
𝒌=𝒄
𝒚𝒕 = 𝒌 𝒕
𝒚⏟𝒕 = 𝒄𝒕
𝑤𝑎𝑠 𝑜𝑝𝑡𝑒𝑑 𝑎𝑠
𝑎 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛 𝑜𝑓 𝒚𝒕
𝒚𝒑 = 𝒄𝒕
General Solution: Summing the expressions of 𝒚𝒄 and 𝒚𝒑 :
𝒚𝒕 = 𝑨(−𝒂)𝒕 + 𝒄𝒕
Interpretation: We have derived the General Solution of First Order Difference equation, without
the parametric restriction of 𝒂 ≠ −𝟏.
Definite Solution: To find the definite solution, we introduce the initial condition, (𝒕 = 𝟎) in the
general solution, 𝒚 𝒕 = 𝑨(−𝒂) 𝒕 + 𝒄 𝒕 .
𝒚 𝟎 = 𝑨(−𝒂) 𝟎 + 𝒄 𝟎
𝒚𝟎 = 𝑨 × (𝟏) + 𝟎
𝒚𝟎 = 𝑨
𝑨 = 𝒚𝟎
Returning the value of arbitrary constant (𝑨) in the general solution 𝒚𝒕 = 𝑨 (−𝒂)𝒕 + 𝒄𝒕 to get the
definite solution:
𝒚𝒕 = 𝒚𝟎 (−𝒂)𝒕 + 𝒄𝒕
𝒚𝒕 = (𝒚𝟎 )(−𝒂)𝒕 + 𝒄𝒕
Interpretation: We have derived the Definite Solution of First Order Difference equation, without
the parametric restriction of 𝒂 ≠ −𝟏.
TOPIC 157: DIFFERENCE EQUATION USING GENERAL METHOD: A=-1 NUMERICAL
Solve the first-order difference equation:

𝒚𝒕+𝟏 − 𝒚𝒕 = 𝟐 (𝒚𝟎 = 𝟎)
Recalling standard form of First-Order Difference Equation:
______________________________________________________________________________
𝒚𝒕+𝟏 − 𝒂𝒚𝒕 = 𝒄
Comparing the given specific form, we extract values:
𝒚𝒕+𝟏 + (−𝟏)
⏟ 𝒚𝒕 = ⏟
𝟐
𝑎 𝑐
Here, 𝒂 = −𝟏
⏟ , 𝒄 = 𝟐.
𝒚𝒑 =𝒄𝒕 𝑖𝑠 𝑠𝑢𝑖𝑡𝑎𝑏𝑙𝑒
Definite Solution: The equation can be solved for the definite solution of time-path by using the
formula:
𝑑𝑒𝑓𝑖𝑛𝑖𝑡𝑖𝑧𝑒𝑑 𝑓𝑜𝑟𝑚 𝑜𝑓
𝑎𝑟𝑏𝑖𝑡𝑟𝑎𝑟𝑦 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 (𝑨)
𝒚𝒕 = ⏞
(𝒚
⏟ 𝟎) (−𝒂)𝒕 + 𝒄𝒕
⏟
⏟ 𝑔𝑖𝑣𝑒𝑛 𝑦𝑝
𝑦𝑐
𝒚𝒕 = (𝟎){−(−𝟏)}𝒕 + 𝟐𝒕
𝒚𝒕 = (𝟎){𝟏}𝒕 + 𝟐𝒕
𝑛𝑜 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑎𝑙𝑟𝑒𝑎𝑑𝑦 𝑎𝑡
𝑓𝑟𝑜𝑚 𝑒𝑞𝑢𝑖𝑙𝑖𝑏𝑟𝑖𝑢𝑚 𝑒𝑞𝑢𝑖𝑙𝑖𝑏𝑟𝑖𝑢𝑚
𝒚𝒕 = ⏟ ⏞
𝟎 ⏞
+ ⏟ 𝟐𝒕
𝑦𝑐 𝑦𝑝
𝒚𝒕 = 𝟐𝒕
Interpretation: The definite solution of the first order difference equation 𝒚𝒕+𝟏 − 𝒚𝒕 = 𝟐 is 𝒚𝒕 = 𝟐𝒕 .
𝒚𝒑 = 𝟐𝒕, represents a moving (jumping/stepped) equilibrium.
______________________________________________________________________________
Lesson 35
DYNAMIC STABILITY OF FIRST ORDER DIFFERENCE EQUATIONS
TOPIC 158: DYNAMIC STABILITY OF FIRST ORDER DIFFERENCE EQUATION: NON-

OSCILLATORY CASES
Like differential equations, the dynamic stability of difference equations also depends on
complementary function (𝒚𝒄 ). Here, 𝒚𝒄 = 𝑨(𝒃)𝒕 , we consider (𝑨 = 𝟏) ⇒ 𝒚𝒄 = (𝒃)𝒕 . This means
that expression (𝒃)𝒕 shall determine the dynamics of 𝒚𝒄 .
Understanding (𝒃)𝒕 Expression: (𝒃)𝒕 contains 𝒃 which further is composed of 𝒂 (𝒃 = −𝒂).
𝒃 𝒂
Base in Solution of First Order Differential Coefficient in First Order Differential
Equation: Equation:
𝒕 𝒚𝒕 + 𝒂 𝒚𝒕−𝟏 = 𝒄
𝒚 = 𝑨( 𝒃 ) + 𝒚
𝒕 𝒑
𝒚𝒕 = 𝑨( −𝒂 )𝒕 + 𝒚𝒑
Determines Dynamic Stability Determines formula for 𝒚𝒄 .
(𝒃)𝒕 is an exponential expression, behavior of which is tabulated below by broadly using +𝒗𝒆,
𝒛𝒆𝒓𝒐 and −𝒗𝒆 values:
(𝒃)𝒕
𝒕 𝒃=𝟐>𝟎 𝒃=𝟎 𝒃 = −𝟐 < 𝟎
Non-negative Negative
0 1 0 1
1 0.5 0 -2
2 4 0 4
3 8 0 -8
4 16 0 16
5 32 0 -32
6 64 0 64
7 128 0 -128
8 256 0 256
9 512 0 -512
10 1024 0 1024
… … … …
Exponentially growth Neither Growth Exponential change with
Remarks
with +𝒗𝒆 signs. nor Decay. Alternative (+𝒗𝒆/−𝒗𝒆) signs.
st
1 Quadrant, Smooth Horizontal 1st & 4th Quadrants, Ups &
Outcomes
graph Straight line Downs in the graph
Terminology Non-oscillatory when (𝒃 ≮ 𝟎) Oscillatory when (𝒃 < 𝟎)
Interpretation: We observe that First Order Difference Equation, have an additional feature of
pattern (i.e. Non-oscillatory or Oscillatory):
Time-path Pattern Nature
First Order Differential Equation Non-oscillatory or Oscillatory Convergent or Divergent
Narrowing down, we focus on the role of 𝒃 only. Using number line to visualize various regions of
𝒃:
______________________________________________________________________________
The above-mentioned seven values of 𝒃: (−𝟏 > 𝒃 > 𝟏) include all possible patterns of graph of
(𝒃𝑡 ). Therefore, for First Order Difference Equations, these values are used to portray
cases(regions) of time-paths. Out of seven cases (VII) cases, first Four (I-IV) are non-oscillatory
(𝒃 ≮ 𝟎), which are depicted below:
Region – I 𝒚𝒄 = (𝒃)𝒕 𝒕=𝟎 𝒕=𝟏 𝒕=𝟐 𝒕=𝟑 𝒕=𝟒
(+𝒗𝒆)(𝟎 𝒕𝒐 𝟒… ) ⇒ +𝒗𝒆 values only: non-oscillatory.
𝒃>𝟏 |𝒃| > 𝟏 e.g.: 𝒚𝒄 = (𝟐)𝒕 1 2 4 8 16
Dynamic Stability: Time-path is divergent & hence dynamically unstable.
______________________________________________________________________________
Region – II 𝒚𝒄 = (𝒃)𝒕 𝒕=𝟎 𝒕=𝟏 𝒕=𝟐 𝒕=𝟑 𝒕=𝟒

(+𝒗𝒆) (𝟎 𝒕𝒐 𝟒… )
⇒ +𝒗𝒆 values only: non-oscillatory.
𝒃=𝟏 |𝒃| = 𝟏 e.g.: 𝒚𝒄 = (𝟏)𝒕 1 1 1 1 1
Dynamic Stability: Time-path is neither divergent nor convergent.
Region – III 𝒚𝒄 = (𝒃)𝒕 𝒕=𝟎 𝒕=𝟏 𝒕=𝟐 𝒕=𝟑 𝒕=𝟒

(+𝒗𝒆) (𝟎 𝒕𝒐 𝟒… )
⇒ +𝒗𝒆 values only: non-oscillatory.
𝟎<𝒃<𝟏 |𝒃| < 𝟏 𝟏 𝒕 1 𝟏 𝟏 𝟏 𝟏
e.g.: 𝒚𝒄 = ( )
𝟐 𝟐 𝟒 𝟖 𝟏𝟔
Dynamic Stability: Time-path is convergent & hence dynamically stable.
Region – IV 𝒚𝒄 = (𝒃)𝒕 𝒕=𝟎 𝒕=𝟏 𝒕=𝟐 𝒕=𝟑 𝒕=𝟒

(+𝒗𝒆)(𝟎 𝒕𝒐 𝟒… ) ⇒ +𝒗𝒆 values only: non-oscillatory.
𝒃=𝟎 |𝒃| = 𝟎 e.g.: 𝒚𝒄 = (𝟎)𝒕 0 0 0 0 0
______________________________________________________________________________
Dynamic Stability: Time-path is neither divergent nor convergent. Infact it is pre-converged

at equilibrium. However, since 𝒚𝒄 is 𝟎, it means that time-path is already on equilibrium.
Summary: Cases (Regions) I-IV have non-negative 𝒃. A non-negative number when raised to
any power generates a positive number, which gives rise to a graph (time-path, here) in 1st
Quadrant only – hence non-oscillatory.
𝒃 ≮ 𝟎 ⇒ 𝑵𝒐𝒏 − 𝒐𝒔𝒄𝒊𝒍𝒍𝒂𝒕𝒐𝒓𝒚
Only when |𝒃| < 𝟏, the time-path is convergent.
|𝒃| < 𝟏 ⇒ 𝑪𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒕
Tabulating the observations:
Region – I Region – II Region – III Region – IV
(𝒃 (𝒃 ≮ 𝟎)
⏟ > 𝟏) ⋁ ⏟ (𝒃 ≮ 𝟎)
⏟ = 𝟏) ⋁ ⏟
(𝒃 (𝒃 ≮ 𝟎)
⏟ = 𝟎) ⋁ ⏟
(𝒃
𝑛𝑜𝑛− 𝑛𝑜𝑛−
(𝟎 < 𝒃
⏟ < 𝑛𝑜𝑛−
𝟐 𝟏 𝟎
𝑜𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦 𝑜𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦 𝟏 𝑜𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦
( )
𝟐
(𝒃 ≮ 𝟎)
𝟏) ⋁ ⏟
𝑛𝑜𝑛−
𝑜𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦
(|𝒃|
⏟ (|𝒃|
⏟ (|𝒃| < 𝟏)
⏟ = 𝟎) ⋁ ⏟
(|𝒃|
𝟐 𝟏 (𝟎 < |𝒃|
⏟ ⏟𝟎 𝑝𝑟𝑒−
𝟏 𝑦𝑐 =0 𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑑
( )
> 𝟏) ⋁ (|𝒃|
⏟ ≮ 𝟏) = 𝟏) ⋁ (|𝒃|
⏟ ≮ 𝟏) 𝟐 𝑛𝑜 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛
𝑛𝑜𝑛− 𝑛𝑜𝑛−
𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡 𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡
(𝑑𝑖𝑣𝑒𝑟𝑔𝑒𝑛𝑡) (𝑛𝑒𝑖𝑡ℎ𝑒𝑟) < 𝟏) ⋁ (|𝒃|
⏟ < 𝟏)
𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡
TOPIC 159: DYNAMIC STABILITY OF FIRST ORDER DIFFERENCE EQUATION:

OSCILLATORY CASES
After explaining first four (I-IV) cases(regions), we explain the remaining three (V-VII)
cases(regions). Out of seven cases (VII) cases, last three (V-VII) are oscillatory (𝒃 < 𝟎), which
are depicted below:
Region – V 𝒚𝒄 = (𝒃)𝒕 𝒕=𝟎 𝒕=𝟏 𝒕=𝟐 𝒕=𝟑 𝒕=𝟒
(−𝒗𝒆)(𝟎 𝒕𝒐 𝟒… ) ⇒ +𝒗𝒆/−𝒗𝒆 (alternative signs): oscillatory.
______________________________________________________________________________
−𝟏 < 𝒃 < 𝟎 |𝒃| < 𝟎 𝟏 𝒕 1 𝟏 𝟏 𝟏 𝟏

e.g. 𝒚𝒄 = (− 𝟐) − −
𝟐 𝟒 𝟖 𝟏𝟔
Dynamic Stability: Time-path is convergent & hence dynamically stable. It moves above
and below (oscillates) the equilibrium with decreasing vertical distance.
Region – VI 𝒚𝒄 = (𝒃)𝒕 𝒕=𝟎 𝒕=𝟏 𝒕=𝟐 𝒕=𝟑 𝒕=𝟒

(−𝒗𝒆) (𝟎 𝒕𝒐 𝟒… )
⇒ +𝒗𝒆/−𝒗𝒆 (alternative signs): oscillatory.
𝒃 = −𝟏 |𝒃| = 𝟏 e.g.: 𝒚𝒄 = (−𝟏)𝒕 1 -1 1 -1 1
Dynamic Stability: Time-path is neither divergent nor convergent. It moves above and
below (oscillates) the equilibrium with uniform vertical distance.
Region – VII 𝒚𝒄 = (𝒃)𝒕 𝒕=𝟎 𝒕=𝟏 𝒕=𝟐 𝒕=𝟑 𝒕=𝟒

(−𝒗𝒆)(𝟎 𝒕𝒐 𝟒… ) ⇒ +𝒗𝒆/−𝒗𝒆 (alternative signs): oscillatory.
𝒃 < −𝟏 |𝒃| < 𝟏 e.g.: 𝒚𝒄 = (−𝟐)𝒕 1 -2 4 -8 16
______________________________________________________________________________
Dynamic Stability: Time-path is convergent & hence dynamically unstable. It moves above
and below (oscillates) the equilibrium with increasing vertical distance.
Summary: Cases (Regions) V-VII have negative 𝒃. A negative number when raised to any power
generates numbers with alternative signs(+𝒗𝒆/−𝒗𝒆), which give rise to a graph (time-path,
here) in 1st and 4th Quadrants – hence oscillatory.
𝒃 < 𝟎 ⇒ 𝑶𝒔𝒄𝒊𝒍𝒍𝒂𝒕𝒐𝒓𝒚
Only when |𝒃| < 𝟏, the time-path is convergent.
|𝒃| < 𝟏 ⇒ 𝑪𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒕
Tabulating the observations:
Region – V Region – VI Region – VII
(𝒃 (𝒃 < 𝟎)
⏟ = −𝟏) ⋁ ⏟ ( 𝒃 (𝒃 < 𝟎)
⏟ < −𝟏) ⋁ ⏟
(−𝟏 < 𝒃 (𝒃 < 𝟎)
⏟ < 𝟎) ⋁ ⏟ −𝟏 𝑜𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦 (−𝟐) 𝑜𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦
𝟏 𝑜𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦
(− )
𝟐
(|𝒃| (|𝒃| ≮ 𝟏)
⏟ = 𝟏) ⋁ ⏟ ⏟ > 𝟏) ⋁
(|𝒃| (|𝒃| ≮ 𝟏)
⏟
(|𝒃| < 𝟏)
⏟ < 𝟏) ⋁ ⏟
(|𝒃| 𝟏 𝑛𝑜𝑛− 𝟐 𝑛𝑜𝑛−𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡
𝟏 𝑐𝑜𝑛𝑣𝑒𝑔𝑒𝑛𝑡 (𝑑𝑖𝑣𝑒𝑟𝑔𝑒𝑛𝑡)
𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡
𝟐 (𝑛𝑒𝑖𝑡ℎ𝑒𝑟)
______________________________________________________________________________
TOPIC 160: DYNAMIC STABILITY OF FIRST ORDER DIFFERENCE EQUATION: ROLE OF 𝑨
Recalling 𝒚𝒄 = 𝑨
⏟ (𝒃)𝒕 . Out of seven cases, we treat the convergent ones, as they are
𝑵𝒐𝒏−𝒐𝒔𝒄𝒊𝒍𝒍𝒂𝒕𝒐𝒓𝒚 𝑶𝒔𝒄𝒊𝒍𝒍𝒂𝒕𝒐𝒓𝒚
the desired cases – ⏟⏞
𝑪𝒂𝒔𝒆 – 𝑰𝑰𝑰 & ⏞
𝑪𝒂𝒔𝒆 – 𝑽. In this selection, we get both of the patterns of
𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒕
time-path (Non-oscillatory and Oscillatory). However, remaining (undesired) 5 cases can also
be treated in the same fashion.
(𝑨) being a coefficient can affect 𝒚𝒄 in either or both of the two ways:
Dual Effect Value of 𝑨 on 𝒚𝒄
Magnitude Sign
|𝑨| = 𝟎 |𝑨| < 𝟏 |𝑨| = 𝟏 |𝑨| > 𝟏 𝑨>𝟎 𝑨<𝟎
Nullify Minify Neutral Amplify Retain Reverse
𝒚𝒄 = 𝟎 (𝒃)𝒕 𝒚𝒄 = 𝟎. 𝟓 (𝒃)𝒕 𝒚𝒄 = 𝟏 (𝒃)𝒕 𝒚𝒄 = 𝟐 (𝒃)𝒕 𝒚𝒄 = 𝑨 ⏟ (𝒃)𝒕 𝒚𝒄 = 𝑨⏟ (𝒃)𝒕
>0 <0
𝒚𝒄 = 𝟎 𝒚𝒄 = 𝟎. 𝟓𝒃𝒕 𝒚𝒄 = 𝒃𝒕 𝒚𝒄 = 𝟐𝒃𝒕 𝒚𝒄 > 𝟎 𝒚𝒄 < 𝟎
- Noteworthy - Noteworthy - Noteworthy

Scale Effect Mirror Effect
1. Scale Effect:
Case – III (Convergent & Non-oscillatory):
Three time-paths of complementary functions are plotted:
1. Original 𝒚𝒄
2. Doubled 𝒚𝒄
3. Halved 𝒚𝒄
𝟏 𝒕
Original Complementary function: 𝒚𝒄 = 𝑨 ( ) : [Case – III]
𝟐
______________________________________________________________________________
𝟏 𝒕
Doubled Complementary function: 𝒚𝒄 = 𝟐 (𝟐) : [Case – III]
𝟏 𝟏 𝒕
Halved Complementary function: 𝒚𝒄 = ( ) : [Case – III]
𝟐 𝟐
Interpretation: Original 𝒚𝒄 starts from 𝟏, however, Doubled 𝒚𝒄 from 𝟐 and Doubled 𝒚𝒄 from 𝟎. 𝟓
respectively. It is evident that magnitude of 𝑨(≷ 𝟏) affects the scale of 𝒚𝒄 by upscaling or
downscaling the time-path. Its, therefore, called the Scale Effect of 𝑨.
Case – V (Convergent & Oscillatory):

Three time-paths of complementary functions are plotted:
______________________________________________________________________________
2. Doubled 𝒚𝒄
3. Halved 𝒚𝒄
𝟏 𝒕
Original Complementary function: 𝒚𝒄 = 𝑨 (− 𝟐) : [Case – V]
𝟏 𝒕
Doubled Complementary function: 𝒚𝒄 = 𝟐 (− 𝟐) : [Case – V]
______________________________________________________________________________
𝟏 𝟏 𝒕
Halved Complementary function: 𝒚𝒄 = 𝟐
(− 𝟐) : [Case – V]
______________________________________________________________________________
Interpretation: Original 𝒚𝒄 starts from 𝟏, however, Doubled 𝒚𝒄 from 𝟐 and Doubled 𝒚𝒄 from 𝟎. 𝟓
respectively. It is evident that magnitude of 𝑨(≷ 𝟏) affects the scale of 𝒚𝒄 by upscaling or
downscaling the time-path. It’s, therefore, called the Scale Effect of 𝑨.
2. Mirror Effect:
Case – III (Convergent & Non-oscillatory):
Two time-paths of complementary functions are plotted:
2. Reversed 𝒚𝒄
𝟏 𝒕
Original Complementary function: 𝒚𝒄 = 𝑨 (𝟐) : [Case – III]
______________________________________________________________________________
𝟏 𝒕
Reversed Complementary function: 𝒚𝒄 = −𝟏 (𝟐) : [Case – III]
Interpretation: Both 𝒚𝒄 start from 𝟏, however, +𝟏 and −𝟏 respectively. It is evident that −𝒗𝒆 sign
of 𝑨(< 𝟎) reverses the shape of 𝒚𝒄 by making a mirror-image of the original time-path. It’s,
therefore, called the Mirror Effect of 𝑨.
Case – V (Convergent & Oscillatory):

Two time-paths of complementary functions are plotted:
2. Reversed 𝒚𝒄
𝟏 𝒕
Original Complementary function: 𝒚𝒄 = 𝑨 (− 𝟐) : [Case – V]
______________________________________________________________________________
𝟏 𝒕
Reversed Complementary function: 𝒚𝒄 = −𝟏 (− 𝟐) : [Case – V]
Interpretation: Both 𝒚𝒄 start from 𝟏, however, +𝟏 and −𝟏 respectively. It is evident that −𝒗𝒆 sign
of 𝑨(< 𝟎) reverses the shape of 𝒚𝒄 by making a mirror-image of the original time-path. It’s,
therefore, called the Mirror Effect of 𝑨.
TOPIC 161: DYNAMIC STABILITY OF FIRST ORDER DIFFERENCE EQUATION:

NUMERICALS
𝟒 𝒕
Example # 1: Check the dynamic stability the time-path 𝒚𝒕 = 𝟐 (− 𝟓) + 𝟗?
𝟒 𝒕 𝒄
𝟗 , the formula in the background is: 𝒚𝒕 = 𝑨(𝒃)𝒕 + 𝟏+𝒂.
Since 𝒚𝒕 = 𝟐 (− 𝟓) + ⏟
≠𝑓(𝑡)
______________________________________________________________________________
Breaking-up the definite solution into complementary and particular solutions:

𝑦𝑐
⏞ 𝟒 𝒕 𝑦𝑝
𝒚𝒕 = 𝟐 (− ) + ⏞
𝟗
⏟ 𝟓
𝑏
Dynamic stability of time-path depends on original and absolute values of 𝒃.
𝟒 𝟒
Interpretation: Here, 𝒃⏟ = − 𝟓 < 𝟏 & |𝒃|
⏟ = 𝟓 < 𝟎 . Therefore, time path is oscillatory
𝑜𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦 𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡
in pattern and convergent in nature. The time path is dynamically stable.
Note:
𝟒 𝒕 𝟒 𝒕
Time-path of 𝒚𝒕 = 𝟐 (− 𝟓) + 𝟗 vs. Time-path of 𝒚𝒕 = −𝟐 (𝟓) + 𝟗
Example # 2: Check the dynamic stability the time-path 𝒚𝒕 = 𝟑(𝟐)𝒕 + 𝟒?
𝒄
Since 𝒚𝒕 = 𝟑(𝟐)𝒕 + ⏟
𝟒 , the formula in the background is: 𝒚𝒕 = 𝑨(𝒃)𝒕 + 𝟏+𝒂.
≠𝑓(𝑡)
Breaking-up the definite solution into complementary and particular solutions:
𝑦𝑐 𝑦𝑝
⏞(𝟐)𝒕 + ⏞
𝒚𝒕 = 𝟑 𝟒
⏟
𝑏
Dynamic stability of time-path depends on original and absolute values of 𝒃.
Interpretation: Here, 𝒃 ⏟ = 𝟐 ≠ 𝟏 & |𝒃|
⏟ = 𝟐 ≮ 𝟎 . Therefore, time path is non
𝑛𝑜𝑛−𝑜𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦 𝑛𝑜𝑛−𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡
oscillatory in pattern and non convergent in nature. The time path is
______________________________________________________________________________
______________________________________________________________________________
Lesson 36
INCOME DYNAMICS OF DIFFERENCE EQUATIONS
TOPIC 162: LAGGED INCOME DETERMINATION MODEL: FRAMEWORK
Unlike simple income determination model, we can introduce lags in consumption function:
𝑪𝒕 = 𝑪𝟎 + 𝒄𝒀𝒕−𝟏
⏟
𝐼𝑛𝑐𝑜𝑚𝑒 𝑜𝑓 𝑙𝑎𝑠𝑡 𝑡𝑖𝑚𝑒−𝑝𝑒𝑟𝑖𝑜𝑑
𝑑𝑒𝑡𝑒𝑟𝑚𝑖𝑛𝑒𝑠 𝑐𝑢𝑟𝑟𝑒𝑛𝑡 𝑐𝑜𝑛𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛
𝒀𝒕 = 𝑪𝒕 + 𝑰𝒕
Here, 𝒄 = 𝑴𝑷𝑪
⏟ , ⏟
𝑰𝒕 = 𝑰𝟎
(0<𝑐<1) 𝑎𝑢𝑡𝑜𝑛𝑜𝑚𝑜𝑢𝑠
𝑖𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡
Substituting the value of 𝑪𝒕 in 𝒀𝒕 = 𝑪𝒕 + 𝑰𝒕
𝒀𝒕 = 𝑪𝟎 + 𝒄𝒀𝒕−𝟏 + 𝑰𝒕
𝒀𝒕 − 𝒄𝒀𝒕−𝟏 = 𝑪𝟎 + 𝑰𝒕
𝒀𝒕 + (−𝒄)
⏟ 𝒀𝒕−𝟏 = ⏟ 𝑪𝟎 + 𝑰𝒕
𝑎≠−1 𝑐
𝑎𝑠 𝑀𝑃𝐶<1
Here, ⏟
𝒂 = −𝒄, ⏟
𝒄 = 𝑪𝟎 + 𝑰𝒕
𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
𝑜𝑓 𝐹𝑂𝐷𝐸 𝑜𝑓 𝐹𝑂𝐷𝐸
As coefficient of First Order Difference equation is ≠ −𝟏, we use the following formula for the
definite time-path of National Income.
(𝐶0 +𝐼𝑡 ) (𝐶0 +𝐼𝑡 )
𝒀𝒕 𝒀𝟎
⏞
𝒄 𝒕 ⏞
𝒄
𝒚⏞𝒕 = (𝒚
⏞𝟎 − ) (𝒃)
⏟ +
𝟏+ 𝒂⏟ (=−𝑎) 𝟏+ 𝒂⏟
(−𝑐) (−𝑐)
(𝐶0 +𝐼𝑡 ) (𝐶0 +𝐼𝑡 )
𝒀𝒕 𝒀𝟎
⏞
𝒄 𝒕 ⏞
𝒄
𝒚⏞𝒕 = (𝒚
⏞𝟎 − ) (−𝒂)
⏟ +
𝟏+ 𝒂⏟ 𝟏+ 𝒂⏟
𝒂=−𝒄
(−𝑐) (−𝑐)
𝑪𝟎 + 𝑰𝒕 𝑪𝟎 + 𝑰𝒕
𝒀𝒕 = {𝒀𝟎 − } {−(−𝒄)}𝒕 +
𝟏 + (−𝒄) 𝟏 + (−𝒄)
𝑪𝟎 + 𝑰𝒕 𝑪 𝟎 + 𝑰𝒕
𝒀𝒕 = {𝒀𝟎 − } (𝒄)𝒕 +
𝟏−𝒄 𝟏−𝒄
𝑪𝟎 +𝑰𝒕 𝑪𝟎 +𝑰𝒕
Interpretation: 𝒀𝒕 = {𝒀𝟎 − 𝟏−𝒄
} (𝒄)𝒕 + 𝟏−𝒄
is the definite time-path of the first order difference
equation 𝒀𝒕 − 𝒄𝒀𝒕−𝟏 = 𝑪𝟎 + 𝑰𝒕 . Dynamic Stability of the time-path, mainly, depends on
exponential expression (𝒄𝒕 ).
𝒄≰𝟎
As per economic theory: ⏞
𝟎 < 𝑴𝑷𝑪 < 𝟏:
⏟
𝒄<𝟏
𝒕
|𝒄|≮𝟎
⏞ } ] : (|𝒄| ≮ 𝟎) ⇒ 𝒏𝒐𝒏 − 𝒐𝒔𝒄𝒊𝒍𝒍𝒂𝒕𝒐𝒓𝒚
(𝒄)
[{ ⏟
(𝒄 < 𝟏) ⇒ 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒕
𝒄<𝟏
Therefore, the time-path is dynamically stable with a non-oscillatory pattern and a convergent
nature.
______________________________________________________________________________
𝒀𝒄 (𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑓𝑟𝑜𝑚 𝒀∗)

𝒀𝒄 →𝟎,𝑎𝑠 𝒕→∞
⏞ 𝑪𝟎 + 𝑰𝒕 𝑪𝟎 + 𝑰𝒕
𝒀𝒕 = {𝒀𝟎 − ⏟𝒕 +
} (𝒄)
⏟ 𝟏−𝒄 ⏟𝟏 − 𝒄
→𝟎
𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑎𝑠 𝒕→∞ 𝒀𝒑 (=𝒀∗ )
⏟ 𝑜𝑓 (𝒄)𝒕
𝒀𝒕 →𝒀∗ ,𝑎𝑠 𝒕→∞
TOPIC 163: LAGGED INCOME DETERMINATION MODEL: NUMERICAL
In addition to symbolic determination of income determination model with time lags, we can solve
it numerically, as well:
𝑪𝒕 = 𝟐𝟎𝟎 + 𝟎. 𝟗𝒀𝒕−𝟏
⏟
𝑎𝑢𝑡𝑜𝑛𝑜𝑚𝑜𝑢𝑠
𝑖𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡
⏞
𝑰𝒕 = 𝟏𝟎𝟎
𝒀𝟎 = 𝟒𝟓𝟎𝟎
Here, ⏟
𝑴𝑷𝑪 = 𝟎. 𝟗.
(0<𝑐<1)
Substituting the value of 𝑪𝒕 and 𝑰𝒕 in 𝒀𝒕 = 𝑪𝒕 + 𝑰𝒕
𝐶0 𝑀𝑃𝐶
⏞ + 𝟎.
𝒀𝒕 = 𝟐𝟎𝟎 ⏞ 𝟗 𝒀𝒕−𝟏 + 𝟏𝟎𝟎
𝒀𝒕 − 𝟎. 𝟗𝒀𝒕−𝟏 = 𝟐𝟎𝟎 + 𝟏𝟎𝟎
(−𝟎. 𝟗) 𝒀𝒕−𝟏 = 𝟑𝟎𝟎
𝒀𝒕 + ⏟ ⏟
𝑎≠−1 𝑐
Comparing the standard form of First Order Difference equation:
𝒚𝒕 + 𝒂𝒚𝒕−𝟏 = 𝒄
Since 𝒂 = −𝟎. 𝟗 ≠ −𝟏, the definite time-path will be:
300 300
𝒀𝒕 𝒀𝟎 −𝒂 𝒕
⏞
𝒄 ⏞
𝒄
𝒚⏞𝒕 = (𝒚
⏞𝟎 − )(⏞
𝒃) +
𝟏+ 𝒂⏟ 𝟏+ 𝒂⏟
−0.9 −0.9
𝟑𝟎𝟎 𝟑𝟎𝟎
𝒀𝒕 = (𝟒𝟓𝟎𝟎 − ) {−(−𝟎. 𝟗)}𝒕 +
𝟏 − 𝟎. 𝟗 𝟏 − 𝟎. 𝟗
𝒀𝒕 = (𝟒𝟓𝟎𝟎 − 𝟑𝟎𝟎𝟎)(𝟎. 𝟗)𝒕 + 𝟑𝟎𝟎𝟎
𝒀𝒕 = (𝟏𝟓𝟎𝟎)(𝟎. 𝟗)𝒕 + 𝟑𝟎𝟎𝟎
Interpretation: 𝒀𝒕 = (𝟏𝟓𝟎𝟎)(𝟎. 𝟗)𝒕 + 𝟑𝟎𝟎𝟎 is the definite time-path of the first order difference
equation 𝒀𝒕 − 𝟎. 𝟗𝒀𝒕−𝟏 = 𝟑𝟎𝟎 . Dynamic Stability of the time-path, mainly, depends on
exponential expression (𝟎. 𝟗)𝒕 .
𝒕
|𝟎.𝟗|≮𝟎
⏞ (|𝟎. 𝟗| ≮ 𝟎) ⇒ 𝒏𝒐𝒏 − 𝒐𝒔𝒄𝒊𝒍𝒍𝒂𝒕𝒐𝒓𝒚
⏟ 𝟗) } ] :
[{(𝟎.
(𝟎. 𝟗 < 𝟏) ⇒ 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒕
𝟎.𝟗<𝟏
nature.
______________________________________________________________________________
𝒀𝒄 (𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑓𝑟𝑜𝑚 𝒀∗)

𝒀𝒕 = ⏞⏟ ⏟ 𝟗)𝒕 + 𝟑𝟎𝟎𝟎
(𝟏𝟓𝟎𝟎) (𝟎. ⏟
𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 →𝟎 𝒀𝒑 (=𝒀∗ )
⏟ 𝑜𝑓 (𝟎.𝟗)𝒕 𝑎𝑠 𝒕→∞
𝒀𝒕 →(𝒀∗ =𝟑𝟎𝟎𝟎),𝑎𝑠 𝒕→∞
TOPIC 164: INCOME DYNAMICS WITH INDUCED INVESTMENT USING DIFFERENCE

EQUATIONS
After introducing induced investment alongside autonomous investment, the income

determination model with time lags, can be done numerically, as follows:
𝑴𝑷𝑪
⏞
𝑪𝒕 = 𝟐𝟎𝟎 + 𝟎. 𝟒 𝒀𝒕−𝟏
⏟
𝑎𝑢𝑡𝑜𝑛𝑜𝑚𝑜𝑢𝑠 𝑖𝑛𝑑𝑢𝑐𝑒𝑑
𝑰𝒕 = ⏞
𝟏𝟎𝟎 +⏞ 𝟎. 𝟐𝒀𝒕−𝟏
𝒀𝟎 = 𝟗𝟓𝟎
Here, ⏟
𝑴𝑷𝑪 = 𝟎. 𝟒.
(0<𝑐<1)
Substituting the value of 𝑪𝒕 and 𝑰𝒕 in 𝒀𝒕 = 𝑪𝒕 + 𝑰𝒕
𝐶0 𝑀𝑃𝐶
⏞ + 𝟎.
𝒀𝒕 = 𝟐𝟎𝟎 ⏞ 𝟒 𝒀𝒕−𝟏 + 𝟏𝟎𝟎 + 𝟎. 𝟐𝒀𝒕−𝟏
𝒀𝒕 − 𝟎. 𝟒𝒀𝒕−𝟏 − 𝟎. 𝟐𝒀𝒕−𝟏 = 𝟐𝟎𝟎 + 𝟏𝟎𝟎
𝒀𝒕 − 𝟎. 𝟔𝒀𝒕−𝟏 = 𝟑𝟎𝟎
(−𝟎. 𝟔) 𝒀𝒕−𝟏 = 𝟑𝟎𝟎

𝒀𝒕 + ⏟ ⏟
𝑎≠−1 𝑐
______________________________________________________________________________
Since 𝒂 = −𝟎. 𝟔 ≠ −𝟏, the definite time-path will be:
300 300
𝒀𝒕 𝒀𝟎 −𝒂 𝒕
⏞
𝒄 ⏞
𝒄
𝒚⏞𝒕 = (𝒚
⏞𝟎 − )(⏞
𝒃) +
𝟏+ 𝒂⏟ 𝟏+ 𝒂⏟
−0.6 −0.6
𝟑𝟎𝟎 𝟑𝟎𝟎
𝒀𝒕 = (𝟗𝟓𝟎 − ) {−(−𝟎. 𝟔)}𝒕 +
𝟏 − 𝟎. 𝟔 𝟏 − 𝟎. 𝟔
𝒀𝒕 = (𝟗𝟓𝟎 − 𝟕𝟓𝟎)(𝟎. 𝟔)𝒕 + 𝟕𝟓𝟎
𝒀𝒕 = (𝟐𝟎𝟎)(𝟎. 𝟔)𝒕 + 𝟕𝟓𝟎
Interpretation: 𝒀𝒕 = (𝟐𝟎𝟎)(𝟎. 𝟔)𝒕 + 𝟕𝟓𝟎 is the definite time-path of the first order difference
equation 𝒀𝒕 − 𝟎. 𝟔𝒀𝒕−𝟏 = 𝟑𝟎𝟎 . Dynamic Stability of the time-path, mainly, depends on
exponential expression (𝟎. 𝟔)𝒕 .
𝒕
|𝟎.𝟔|≮𝟎
⏞ (|𝟎. 𝟔| ≮ 𝟎) ⇒ 𝒏𝒐𝒏 − 𝒐𝒔𝒄𝒊𝒍𝒍𝒂𝒕𝒐𝒓𝒚
⏟ 𝟔) } ] :
[{(𝟎.
(𝟎. 𝟔 < 𝟏) ⇒ 𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒕
𝟎.𝟔<𝟏
nature.
______________________________________________________________________________
𝒀𝒄 (𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑓𝑟𝑜𝑚 𝒀∗ )
𝒀𝒕 = ⏞(𝟐𝟎𝟎)
⏟ ⏟ 𝟔)𝒕 + 𝟕𝟓𝟎
(𝟎. ⏟
𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 →𝟎 𝒀𝒑 (=𝒀∗ )
⏟ 𝑜𝑓 (𝟎.𝟗)𝒕 𝑎𝑠 𝒕→∞
𝒀𝒕 →(𝒀∗ =𝟕𝟓𝟎),𝑎𝑠 𝒕→∞
TOPIC 165: INCOME DYNAMICS & CAPITAL MARKET USING DIFFERENCE EQUATIONS
Savings in an economy represent a share of current income given by the marginal propensity to
save (𝒔), whereas firms decide to invest a share of the income earned in the previous period.
Thus:
𝑺𝒕 = 𝒔𝒀𝒕 & 𝑰𝒕 = 𝜷𝒀𝒕−𝟏
Elaborating the given equations:
(0<𝑠<1)
⏞
𝑴𝑷𝑺
𝑺𝒕 = ⏞
𝒔 𝒀𝒕 𝒔>𝟎
𝑝𝑟𝑜𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑜𝑓 𝑙𝑎𝑠𝑡
𝑦𝑒𝑎𝑟 ′ 𝑠 𝑖𝑛𝑐𝑜𝑚𝑒
𝑡ℎ𝑎𝑡 𝑖𝑠 𝑖𝑛𝑣𝑒𝑠𝑡𝑒𝑑
𝑰𝒕 =
⏟ ⏞
𝜷 𝒀𝒕−𝟏 𝜷<𝟏
𝑑𝑒𝑡𝑒𝑟𝑚𝑖𝑛𝑒𝑠 𝑐𝑢𝑟𝑟𝑒𝑛𝑡 𝑖𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡
For equilibrium, savings should be equal to investment:𝑺𝒕 = 𝑰𝒕
𝑺𝒕 = 𝑰𝒕
𝒔𝒀𝒕 = 𝜷𝒀𝒕−𝟏
𝒔𝒀𝒕 − 𝜷𝒀𝒕−𝟏 = 𝟎
Normalizing the equation by dividing throughout by 𝒔:
𝜷
𝒀𝒕 − ( ) 𝒀𝒕−𝟏 = 𝟎
𝒔
𝜷
𝒀𝒕 + {− ( )} 𝒀𝒕−𝟏 = ⏟
𝟎
⏟ 𝒔 𝑐
𝑎≠−1
______________________________________________________________________________
+𝑣𝑒
⏞
𝜷
𝒂 = −( ) ≠
⏟ −𝟏
⏟
𝒔 (𝜷≠𝒔)
+𝑣𝑒
Recalling the formula for definite solution of First Order Difference equation with assumption that
𝜷
(𝒂)(𝜷≠𝒔) = − ≠ −𝟏.
𝒔
𝜷
Here; 𝒂 = − 𝒔, 𝒄 = 𝟎, 𝒚𝒕 = 𝒀𝒕 , 𝒚𝟎 = 𝒀𝟎 , 𝒃 = −𝒂.
𝒕
(−𝒂)
𝒄 𝒄
𝒚𝒕 = (𝒚𝟎 − ⏞ ) +
)( 𝒃
𝟏+𝒂 𝟏+𝒂
𝟎 𝟎
𝒀𝒕 = {𝒀𝟎 − } {(−𝒂)}𝒕 +
𝜷 𝜷
𝟏 + (− 𝒔 ) 𝟏 + (− 𝒔 )
𝜷 𝒕
𝒀𝒕 = {𝒀𝟎 − 𝟎} [{− (− )}] + 𝟎
𝒔
𝒕
𝜷
𝒀𝒕 = 𝒀𝟎 ( )
𝒔
𝜷 𝒕
Interpretation: 𝒀𝒕 = 𝒀𝟎 ( 𝒔 ) is the definite time-path of the first order difference equation
𝜷
𝒀𝒕 − 𝒔 𝒀𝒕−𝟏 = 𝟎 . Dynamic Stability of the time-path, mainly, depends on exponential expression
𝜷 𝒕
(𝒔) .
(>𝟏):𝑛𝑜𝑛−𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡 Pattern
Nature {⏟ ⏞
|𝒃| } (<𝟎):𝑜𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦
(<𝟏):𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡 { ⏟ ⏞
𝒃 }
(>𝟎):𝑛𝑜𝑛−𝑜𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦
𝜷 𝑖𝑓 (𝜷<𝒔) 𝜷 𝜷
|𝒃| = | | ⋚ 𝟏 ⇒ |𝒔| < 𝒃=( )≮𝟎⇒
𝟏
𝒔 ⏟ 𝒔
𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡 Non-oscillatory
Therefore, the time-path will be convergent in nature only if (𝜷 < 𝒔) while it will move in a non-
oscillatory pattern.
______________________________________________________________________________
TOPIC 166: INCOME DYNAMICS & MONEY MARKET USING DIFFERENCE EQUATIONS
On the money market, the government sets money supply as a percentage of national income
from the previous period. Furthermore, transaction demand for money depends on current income
such that:
𝑴𝒔 = 𝒔𝒀𝒕−𝟏
𝑴𝑫 = 𝜶 + 𝜷𝒀𝒕 (𝟎 < 𝜷 < 𝒔 < 𝟏 & 𝜶 > 𝟎)
Find the aggregate national income that ensures equilibrium on the money market, and check
whether the time path of national income is dynamically stable and oscillating.
Solution:
𝒔: proportion of previous year’s income that is converted into supply of money in current year.
Equilibrium in the money market requires 𝑴𝒔 = 𝑴𝑫:
𝒔𝒀𝒕−𝟏 = 𝜶 + 𝜷𝒀𝒕
−𝜷𝒀𝒕 + 𝒔𝒀𝒕−𝟏 = 𝜶
𝜷𝒀𝒕 − 𝒔𝒀𝒕−𝟏 = −𝜶
Normalizing the equation by dividing throughout by 𝜷:
𝒔 𝜶
𝒀𝒕 + (− ) 𝒀𝒕−𝟏 = −
𝜷 𝜷
𝒔 𝜶
𝒀𝒕 + {(− )} 𝒀𝒕−𝟏 = −
⏟ 𝜷 ⏟ 𝜷
𝑎≠−1 𝑐
+𝑣𝑒
⏞
𝒔
𝒂 = −( ) ≠
⏟ −𝟏
𝜷
⏟ (𝜷<𝒔)
+𝑣𝑒
𝜷
(𝒂)(𝜷<𝒔) = − ≠ −𝟏.
𝒔
𝒔 𝜶
Here; 𝒂 = − 𝜷, 𝒄 = − 𝜷, 𝒚𝒕 = 𝒀𝒕 , 𝒚𝟎 = 𝒀𝟎 , 𝒃 = −𝒂.
𝒕
(−𝒂)
𝒄 𝒄
𝒚𝒕 = (𝒚𝟎 − ⏞ ) +
)( 𝒃
𝟏+𝒂 𝟏+𝒂
𝜶 𝜶
(− ) (− )
𝜷 𝜷
𝒀𝒕 = {𝒀𝟎 − } (−𝒂)𝒕 +
𝒔 𝒔
𝟏 + (− ) 𝟏 + (− )
𝜷 𝜷
𝜶 𝜶
𝜷 𝒔 𝒕 𝜷
𝒀𝒕 = {𝒀𝟎 + 𝒔 } {− (− 𝜷)} − 𝒔
𝟏− 𝟏−
𝜷 𝜷
𝜶 𝜶
𝜷 𝒔 𝒕 𝜷
𝒀𝒕 = {𝒀𝟎 + }( ) −
𝜷−𝒔 𝜷−𝒔
( ) 𝜷 ( )
𝜷 𝜷
𝜶 𝜷 𝒔 𝒕 𝜶 𝜷
𝒀𝒕 = {𝒀𝟎 + × ( )} ( ) − × ( )
𝜷 𝜷−𝒔 𝜷 𝜷 𝜷−𝒔
𝜶 𝒔 𝒕 𝜶
𝒀𝒕 = (𝒀𝟎 + )( ) −
𝜷−𝒔 𝜷 𝜷−𝒔
______________________________________________________________________________
𝐸𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙
𝑒𝑥𝑝𝑟𝑒𝑠𝑠𝑖𝑜𝑛
𝜶 ⏞𝒔 𝒕 𝜶
𝒀𝒕 = (𝒀𝟎 + ) ( ) −
⏟ 𝜷−𝒔 𝜷 𝜷
⏟− 𝒔
𝑌𝑐 𝑌𝑝
𝜶 𝒔 𝒕 𝜶
Interpretation: 𝒀𝒕 = (𝒀𝟎 + 𝜷−𝒔
) (𝜷
) − 𝜷−𝒔 is the definite time-path of the first order difference
𝒔 𝜶
equation 𝒀𝒕 − 𝜷 𝒀𝒕−𝟏 = − 𝜷 . Dynamic Stability of the time-path, mainly, depends on exponential
𝒔 𝒕
expression ( ) .
𝜷
(>𝟏):𝑛𝑜𝑛−𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡 (<𝟎):𝑜𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦
Nature {⏟ ⏞
|𝒃| } Pattern { ⏟ ⏞
𝒃 }
(<𝟏):𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡 (>𝟎):𝑛𝑜𝑛−𝑜𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦
𝒔 𝑎𝑠 𝑔𝑖𝑣𝑒𝑛 (𝜷<𝒔) 𝒔 𝒔
|𝒃| = | | ⋚ 𝟏 ⇒ | |>𝟏 𝒃 = (𝜷) ≮ 𝟎 ⇒Non-oscillatory
𝜷 ⏟ 𝜷
𝑻𝒊𝒎𝒆−𝒑𝒂𝒕𝒉 𝒊𝒔 𝑪𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒕
Therefore, the time-path is convergent in nature as (𝜷 < 𝒔) while it will move in a non-oscillatory
pattern.
TOPIC 167: INCOME DYNAMICS & BALANCED TRADE USING DIFFERENCE EQUATIONS
The imports of a country are a share of national income from the current period. Furthermore,
exports depend both on current income and income from the previous period such that:
𝑴𝒕 = 𝒎𝒀𝒕 𝟎<𝒎<𝟏
𝑿𝒕 = 𝜶𝒀𝒕 + 𝜷𝒀𝒕−𝟏 (𝟎 < 𝜶, 𝜷 < 𝟏)
Find the time path of aggregate national income that ensures balanced foreign trade.
Solution: For balanced Foreign Trade; 𝑴𝒕 = 𝑿𝒕 :
𝒎𝒀𝒕 = 𝜶𝒀𝒕 + 𝜷𝒀𝒕−𝟏
𝒎𝒀𝒕 − 𝜶𝒀𝒕 − 𝜷𝒀𝒕−𝟏 = 𝟎
(𝒎 − 𝜶)𝒀𝒕 − 𝜷𝒀𝒕−𝟏 = 𝟎
Normalizing the equation by dividing throughout by (𝒎 − 𝜶):
𝜷 𝟎
𝒀𝒕 − 𝒀𝒕−𝟏 =
𝒎−𝜶 𝒎−𝜶
𝜷
𝒀𝒕 + {(− )} 𝒀𝒕−𝟏 = ⏟
𝟎
⏟ 𝒎−𝜶 𝑐
𝑎≠−1
𝜷
𝒂 = −( ) ≠ ⏟ −𝟏
𝒎 − 𝜶 (𝒎−𝜶≠𝜷)
𝜷
(𝒂)(𝒎−𝜶≠𝜷) = − ≠ −𝟏.
𝒔
𝜷
Here; 𝒂 = − 𝒎−𝜶, 𝒄 = 𝟎, 𝒚𝒕 = 𝒀𝒕 , 𝒚𝟎 = 𝒀𝟎 , 𝒃 = −𝒂.
𝒕
(−𝒂)
𝟎 𝒄
𝒚𝒕 = (𝒚𝟎 − )( ⏞
𝒃 ) +
𝟏+𝒂 𝟏+𝒂
______________________________________________________________________________
𝟎 𝟎
𝒀𝒕 = {𝒀𝟎 − } (−𝒂)𝒕 +
𝜷 𝜷
𝟏 + (− 𝒎 − 𝜶) 𝟏 + (− 𝒎 − 𝜶)
𝒕
𝜷
𝒀𝒕 = {𝒀𝟎 − 𝟎} {− (− )} − 𝟎
𝒎−𝜶
𝒕
𝜷
𝒀𝒕 = 𝒀𝟎 ( )
𝒎−𝜶
𝐸𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙
⏞ 𝜷 𝒕
𝒀𝒕 = 𝒀𝟎 ( ) −⏟
𝟎
⏟ 𝒎−𝜶
𝑌 𝑝
𝑌𝑐
𝜷 𝒕
Interpretation: 𝒀𝒕 = 𝒀𝟎 (𝒎−𝜶) is the definite time-path of the first order difference equation
𝜷
𝒀𝒕 − 𝒎−𝜶 𝒀𝒕−𝟏 = 𝟎 . Dynamic Stability of the time-path, mainly, depends on exponential
𝜷 𝒕
expression ( ).
𝒎−𝜶
Nature {⏟ ⏞
𝒃 }
𝜷 𝑖𝑓 (𝒎−𝜶>𝜷) 𝜷 >0
|𝒃| = | |⋚𝟏 ⇒ | |<𝟏 ⏞
𝜷 𝑖𝑓 𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡:(𝒎−𝜶>𝜷)
𝒎−𝜶 ⏟ 𝒎−𝜶 𝒃=( ⇒
)⏟ 𝑡ℎ𝑒𝑛:(𝒎−𝜶>𝟎)> 𝟎
𝑻𝒊𝒎𝒆−𝒑𝒂𝒕𝒉 𝒘𝒊𝒍𝒍 𝒃𝒆 𝑪𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒕 𝒎−𝜶
𝑵𝒐𝒏−𝑶𝒔𝒄𝒊𝒍𝒍𝒂𝒕𝒐𝒓𝒚
Therefore, the time-path will be convergent in nature if (𝒎 − 𝜶 > 𝜷) while it will move in a non-
oscillatory pattern as (𝒎 − 𝜶 > 𝟎).
______________________________________________________________________________
Lesson 37
VARIOUS ECONOMIC APPLICATIONS OF DIFFERENCE EQUATIONS
TOPIC 168: INVESTMENT DYNAMICS USING DIFFERENCE EQUATIONS

𝑰𝒕 is the amount invested in period 𝒕:
𝑰𝒕 = 𝟎. 𝟗𝑰𝒕−𝟏 + 𝟔𝟎
𝑰𝒕 + (−𝟎.
⏟ 𝟗) 𝑰𝒕−𝟏 = 𝟔𝟎 ⏟
𝑎≠−1 𝑐
And investment at 𝒕 = 𝟎 is £𝟗𝟎𝟎𝒎. Find the time-path of investment.
Since 𝒂 = −𝟎. 𝟗 ≠ −𝟏, the definite time-path will be:
60 60
𝑰𝒕 𝑰𝟎 −𝒂 𝒕
⏞
𝒄 ⏞
𝒄
𝒚⏞𝒕 = (𝒚
⏞𝟎 − ⏞) +
)(𝒃
𝟏+ 𝒂⏟ 𝟏+ 𝒂⏟
−0.9 −0.9
𝟔𝟎 𝟔𝟎
𝑰𝒕 = (𝟗𝟎𝟎 − ) {−(−𝟎. 𝟗)}𝒕 +
𝟏 − 𝟎. 𝟗 𝟏 − 𝟎. 𝟗
𝑰𝒕 = (𝟗𝟎𝟎 − 𝟔𝟎𝟎)(𝟎. 𝟗)𝒕 + 𝟔𝟎𝟎
𝑰𝒕 = 𝟑𝟎𝟎(𝟎. 𝟗)𝒕 + 𝟔𝟎𝟎
Interpretation: 𝑰𝒕 = 𝟑𝟎𝟎(𝟎. 𝟗)𝒕 + 𝟔𝟎𝟎 is the definite time-path of the first order difference
equation 𝑰𝒕 − 𝟎. 𝟗𝒀𝒕−𝟏 = 𝟔𝟎 . Dynamic Stability of the time-path, mainly, depends on exponential
expression (𝟎. 𝟗)𝒕 .
Nature {⏟ ⏞
𝒃 }
|𝒃| = |−𝟎. 𝟗| < 𝟏
⏟ ⏟ = (𝟎. 𝟗) < 𝟎
𝒃
𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡 𝑁𝑜𝑛−𝑜𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦
Therefore, the time-path will be convergent in nature while it will move in a non-oscillatory pattern
as.
𝑰𝒄 (𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑓𝑟𝑜𝑚 𝑰∗ )
𝑰𝒄 →𝟎,𝑎𝑠 𝒕→∞
𝑰𝒕 = ⏞(𝟑𝟎𝟎)
⏟ ⏟ 𝟗)𝒕 + 𝟔𝟎𝟎
(𝟎. ⏟
𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 →𝟎 𝑰𝒑 (=𝑰∗ )
⏟ 𝑜𝑓 (𝟎.𝟗)𝒕 𝑎𝑠 𝒕→∞
𝑰𝒕 →(𝑰∗ =𝟔𝟎𝟎),𝑎𝑠 𝒕→∞
______________________________________________________________________________
TOPIC 169: PRICE DYNAMICS & MONEY MARKET USING DIFFERENCE EQUATIONS
Money supply (𝑴𝒔 ), is set by the government. Money demand (𝑳𝒕 ) depends on the price level
(𝑷𝒕 ) as well as on real output (𝒀). Inflation can be measured as the difference between the price
levels between consecutive time period (𝑷𝒕+𝟏 − 𝑷𝒕 ), and depends on excess money supply
(𝑴𝒔 − 𝑳𝒕 ) such that:
𝑳𝒕 = 𝜶𝑷𝒕 𝒀 (𝜶 > 𝟎)
𝑷𝒕+𝟏 − 𝑷𝒕 = 𝒋(𝑴𝒔 − 𝑳𝒕 ) (𝒋 > 𝟎)
Express the time path of the price level in the economy. Find its intertemporal equilibrium. How is
this equilibrium related to money supply and aggregate output?
Solution:
Plugging in the value of 𝑳𝒕 in 𝑷𝒕+𝟏 − 𝑷𝒕 = 𝒋(𝑴𝒔 − 𝑳𝒕 ):
𝑷𝒕+𝟏 − 𝑷𝒕 = 𝒋(𝑴𝒔 − 𝜶𝑷𝒕 𝒀 )
𝑷𝒕+𝟏 − 𝑷𝒕 = 𝒋(𝑴𝒔 − 𝜶𝑷𝒕 𝒀)
𝑷𝒕+𝟏 − 𝑷𝒕 + 𝜶𝒋𝑷𝒕 𝒀 = 𝒋𝑴𝒔
𝑷𝒕+𝟏 − (𝟏 − 𝜶𝒋𝒀)𝑷𝒕 = 𝒋𝑴𝒔
Recalling the Standard form of First Order Difference Equation:
𝑷
⏟ 𝒕+𝟏 + ⏟{−(𝟏 − 𝜶𝒋𝒀)} 𝑷
⏟𝒕 = 𝒋𝑴
⏟𝒔
𝑦𝑡+1 𝑎 𝑦𝑡 𝑐
=−𝟏
⏞ >𝟎 >𝟎 >𝟎
⏞ ⏞
Here; 𝒂 = − (𝟏 − ⏟
𝜶 𝒋 ⏞
𝒀 ), 𝒄 = 𝒋𝑴𝒔, 𝒚𝒕 = 𝑷𝒕 , 𝒚𝟎 = 𝑷𝟎 , 𝒃 = −𝒂.
⏟ >0
≠1
𝒂 ≠ −𝟏.
𝒕
(−𝒂)
𝒄 𝒄
𝒚𝒕 = (𝒚𝟎 − ⏞ ) +
)( 𝒃
𝟏+𝒂 𝟏+𝒂
______________________________________________________________________________
(𝒋𝑴𝒔 ) (𝒋𝑴𝒔 )
𝑷𝒕 = {𝑷𝟎 − } [−{−(𝟏 − 𝜶𝒋𝒀)}]𝒕 +
𝟏 + {−(𝟏 − 𝜶𝒋𝒀)} 𝟏 + {−(𝟏 − 𝜶𝒋𝒀)}
𝒋𝑴𝒔 𝒋𝑴𝒔
𝑷𝒕 = {𝑷𝟎 − } (𝟏 − 𝜶𝒋𝒀)𝒕 +
𝟏 − (𝟏 − 𝜶𝒋𝒀) 𝟏 − (𝟏 − 𝜶𝒋𝒀)
𝒋𝑴𝒔 𝒋𝑴𝒔
𝑷𝒕 = (𝑷𝟎 − ) (𝟏 − 𝜶𝒋𝒀)𝒕 +
𝟏 − 𝟏 + 𝜶𝒋𝒀 𝟏 − 𝟏 + 𝜶𝒋𝒀
𝒋𝑴𝒔 𝒋𝑴 𝒔
𝑷𝒕 = (𝑷𝟎 − ) (𝟏 − 𝜶𝒋𝒀)𝒕 +
𝑴𝒔 ⏞ 𝑴𝒔
𝑷𝒕 = (𝑷𝟎 − ) (𝟏 − 𝜶𝒋𝒀)𝒕 +
⏟ 𝜶𝒀 ⏟
𝜶𝒀
𝑃𝑐 𝑃𝑝
𝑴𝒔 𝑴𝒔
𝑷𝒕 = (𝑷𝟎 − ) (𝟏 − 𝜶𝒋𝒀)𝒕 +
𝜶𝒀 𝜶𝒀
𝑴 𝑴
Interpretation: 𝑷𝒕 = (𝑷𝟎 − 𝜶𝒀𝒔 ) (𝟏 − 𝜶𝒋𝒀)𝒕 + 𝜶𝒀𝒔 is the definite time-path of the first order
difference equation 𝑷𝒕+𝟏 − (𝟏 − 𝜶𝒋𝒀)𝑷𝒕 = 𝒋𝑴𝒔 . Dynamic Stability of the time-path, mainly,
depends on exponential expression (𝟏 − 𝜶𝒋𝒀)𝒕 .
Nature {⏟ ⏞
𝒃 }
|𝒃| = |𝟏 − 𝜶𝒋𝒀| 𝒃 = (𝟏 − 𝜶𝒋𝒀)
𝑖𝑓 (𝜶𝒋𝒀<𝟏) 𝑖𝑓 (𝜶𝒋𝒀<𝟏)
⋚𝟏 ⇒
⏟ |𝟏 − 𝜶𝒋𝒀| > 𝟏 ⋚𝟎 ⇒
⏟ (𝟏 − 𝜶𝒋𝒀) ≮ 𝟎
𝑻𝒊𝒎𝒆−𝒑𝒂𝒕𝒉 𝒘𝒊𝒍𝒍 𝒃𝒆 𝑪𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒕 𝑻𝒊𝒎𝒆−𝒑𝒂𝒕𝒉 𝒘𝒊𝒍𝒍 𝒃𝒆 𝑵𝒐𝒏−𝑶𝒔𝒄𝒊𝒍𝒍𝒂𝒕𝒐𝒓𝒚
Therefore, the time-path is convergent in nature if (𝜶𝒋𝒀 < 𝟏) while it will move in a non-
TOPIC 170: FISH STOCK ANALYSIS USING DIFFERENCE EQUATION
A fish population increases by 5% each generation, but a catch of 𝟖𝟎𝟎𝟎 is lost from each
generation. If 𝑷𝒕 represents the size of the population for generation 𝒕:
(a) Write the difference equation which expresses 𝑷𝒕 in terms of 𝑷𝒕 − 𝟏.
(b) Solve the difference equation, given 𝑷𝟎 = 𝟏𝟎𝟎𝟎𝟎𝟎. Hence, describe the time path.
Solution: Mathematizing the given information:
𝟓%↑ 𝑓𝑖𝑠ℎ 𝑐𝑎𝑡𝑐ℎ 𝑖𝑛
𝑖𝑛 𝟏𝟎𝟎 𝑒𝑣𝑒𝑟𝑦 𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑖𝑜𝑛
𝑭⏟𝒕 =⏞
𝟏. 𝟎𝟓 𝑭
⏟ 𝒕−𝟏
⏞
−𝟖𝟎𝟎𝟎
𝑐𝑢𝑟𝑟𝑒𝑛𝑡 𝑓𝑖𝑠ℎ 𝑠𝑡𝑜𝑐𝑘
𝑓𝑖𝑠ℎ 𝑠𝑡𝑜𝑐𝑘 𝑖𝑛 𝑙𝑎𝑠𝑡
𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑖𝑜𝑛
______________________________________________________________________________
𝑭 𝒕 = 𝟏. 𝟎𝟓𝑭 𝒕−𝟏 − 𝟖𝟎𝟎𝟎

Introducing 𝟏 generation lag:
Introducing 𝟏 generation lag:
𝑭𝒕 +𝟏 = 𝟏. 𝟎𝟓𝑭𝒕−𝟏 +𝟏 − 𝟖𝟎𝟎𝟎
𝑭 𝒕+𝟏 = 𝟏. 𝟎𝟓𝑭 𝒕 − 𝟖𝟎𝟎𝟎
𝑭𝒕+𝟏 − 𝟏. 𝟎𝟓𝑭𝒕 = −𝟖𝟎𝟎𝟎
𝑭𝒕+𝟏 + (−𝟏.
⏟ 𝟎𝟓) 𝑭𝒕 = (−𝟖𝟎𝟎𝟎)
⏟
𝑎≠−1 𝑐
Since 𝒂 = −𝟏. 𝟎𝟓 ≠ −𝟏, the definite time-path will be:
−8000 −8000
𝑭𝒕 𝑭𝟎 −𝑎 𝒕
⏞
𝒄 ⏞
𝒄
𝒚⏞𝒕 = (𝒚
⏞𝟎 − )(⏞
𝒃) +
𝟏+ 𝒂⏟ 𝟏+ 𝒂⏟
−1.05 −1.05
𝑭𝒕 = (𝟏𝟎𝟎𝟎𝟎𝟎 − 𝟏𝟔𝟎𝟎𝟎𝟎){−(−𝟏. 𝟎𝟓)}𝒕 + 𝟏𝟔𝟎𝟎𝟎𝟎
𝑭𝒕 = (−𝟔𝟎𝟎𝟎𝟎)(𝟏. 𝟎𝟓)𝒕 + 𝟏𝟔𝟎𝟎𝟎𝟎
𝑭𝒕 = −𝟔𝟎𝟎𝟎𝟎(𝟏. 𝟎𝟓)𝒕 + 𝟏𝟔𝟎𝟎𝟎𝟎
Interpretation: 𝑭𝒕 = −𝟔𝟎𝟎𝟎𝟎(𝟏. 𝟎𝟓)𝒕 + 𝟏𝟔𝟎𝟎𝟎𝟎 is the definite time-path of the first order
difference equation 𝑭𝒕+𝟏 − 𝟏. 𝟎𝟓𝑭𝒕 = −𝟖𝟎𝟎𝟎 . Dynamic Stability of the time-path, mainly,
depends on exponential expression (𝟏. 𝟎𝟓)𝒕 .
Nature {⏟ ⏞
𝒃 }
|𝒃| = |𝟏. 𝟎𝟓| ≮ 𝟏
⏟ 𝒃 = (𝟏. 𝟎𝟓) ≮ 𝟎
⏟
𝑁𝑜𝑛−𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡 𝑁𝑜𝑛−𝑜𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦
Therefore, the time-path will be non-convergent in nature while it will move in a non-oscillatory
pattern as.
𝑭𝒄 (𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑓𝑟𝑜𝑚 𝑭∗ )
𝑭𝒄 ↛𝟎,𝑎𝑠 𝒕→∞
𝑭𝒕 = ⏞
(−𝟔𝟎𝟎𝟎𝟎)
⏟ ⏟ 𝟎𝟓)𝒕 + 𝟏𝟔𝟎𝟎𝟎𝟎
(𝟏. ⏟
𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 →𝟎 𝑭𝒑 (=𝑭∗ )
⏟ 𝑜𝑓 (𝟏.𝟎𝟓)𝒕 𝑎𝑠 𝒕→∞
𝑭𝒕 ↛(𝑭∗ =𝟏𝟔𝟎𝟎𝟎𝟎),𝑎𝑠 𝒕→∞
______________________________________________________________________________
TOPIC 171: INITIAL AMOUNT DEPOSITED IN COMPOUND INTEREST
Derive the formula for the value 𝑷𝒕 of an initial amount of money 𝑷𝟎 deposited at 𝒊 interest for 𝒕
years when compounded annually.
When interest is compounded annually:
𝑷
⏟𝒕+𝟏 = 𝑷
⏟𝒕 + ⏟𝒊 × 𝑷 ⏟𝒕
𝑝𝑟𝑖𝑛𝑐𝑖𝑝𝑎𝑙 𝑖𝑛𝑖𝑡𝑖𝑎𝑙 𝑖𝑛𝑡𝑒𝑟𝑒𝑠𝑡 𝑖𝑛𝑖𝑡𝑖𝑎𝑙
+ 𝑎𝑚𝑜𝑢𝑛𝑡 ⏟𝑟𝑎𝑡𝑒 𝑎𝑚𝑜𝑢𝑛𝑡
𝑖𝑛𝑡𝑒𝑟𝑒𝑠𝑡 𝑐𝑜𝑚𝑝𝑜𝑢𝑛𝑑 𝑖𝑛𝑡𝑒𝑟𝑒𝑠𝑡
𝑷𝒕+𝟏 = (𝟏 + 𝒊)𝑷𝒕
𝑷𝒕+𝟏 − (𝟏 + 𝒊)𝑷𝒕 = 𝟎
Recalling the standard form of First Order Difference Equation:
𝑷
⏟ {−(𝟏 + 𝒊)} 𝑷
𝒕+𝟏 + ⏟ ⏟𝒕 = ⏟
𝟎
𝑦𝑡+1 𝑎≠−1 𝑦𝑡 𝑐
≠𝟎
{− (𝟏 + ⏞𝒊 )} : {−(𝟏 + 𝒊)} > −𝟏 ∴ {−(𝟏 + 𝒊)} ≠ −𝟏.
⏟
𝑎≠−1
Here; 𝒂 = {−(𝟏 + 𝒊)}, 𝒄 = 𝟎, 𝒚𝒕 = 𝑷𝒕 , 𝒚𝟎 = 𝑷𝟎 , 𝒃 = −𝒂.
𝒂 ≠ −𝟏.
______________________________________________________________________________
𝒕
(−𝒂)
𝒄 𝒄
𝒚𝒕 = (𝒚𝟎 − ⏞ ) +
)( 𝒃
𝟏+𝒂 𝟏+𝒂
𝟎 𝟎
𝑷𝒕 = {𝑷𝟎 − } (−𝒂)𝒕 +
𝟏 + {−(𝟏 + 𝒊)} 𝟏 + {−(𝟏 + 𝒊)}
𝑷𝒕 = {𝑷𝟎 }[−{−(𝟏 + 𝒊)}]𝒕
𝑷𝒕 = 𝑷𝟎 (𝟏 + 𝒊)𝒕
⏟𝟎 ⏞
𝑷𝒕 = 𝑷 (𝟏 + 𝒊)𝒕 + ⏟
𝟎
𝑌𝑐 𝑌𝑝
𝒕
Interpretation: 𝑷𝒕 = 𝑷𝟎 (𝟏 + 𝒊) is the definite time-path of the first order difference equation
𝑷𝒕+𝟏 − (𝟏 + 𝒊)𝑷𝒕 = 𝟎 . Dynamic Stability of the time-path, mainly, depends on exponential
expression (𝟏 + 𝒊)𝒕 .
Nature {⏟ ⏞
𝒃 }
𝑔𝑖𝑣𝑒𝑛 (𝒊>𝟎) 𝑔𝑖𝑣𝑒𝑛 (𝒂>𝟎)
|𝒃| = |𝟏 + 𝒊| ⇒
⏟ |𝟏 + 𝒊| ≮ 𝟏 𝒃 = (𝟏 + 𝒊) ⇒
⏟ (𝟏 + 𝒊) ≮ 𝟎
𝑻𝒊𝒎𝒆−𝒑𝒂𝒕𝒉 𝒘𝒊𝒍𝒍 𝒃𝒆 𝑵𝒐𝒏−𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒕 𝑻𝒊𝒎𝒆−𝒑𝒂𝒕𝒉 𝒘𝒊𝒍𝒍 𝒃𝒆 𝑵𝒐𝒏−𝑶𝒔𝒄𝒊𝒍𝒍𝒂𝒕𝒐𝒓𝒚
Therefore, the time-path is non-convergent in nature as (𝒂 > 𝒔) while it will move in a non-
TOPIC 172: HARROD MODEL OF GROWTH
The Harrod model attempts to explain the dynamics of growth in the economy. It assumes
𝑺𝒕 = 𝒔𝒀𝒕
where 𝒔 is a constant equal to both the 𝑴𝑷𝑺 and 𝑨𝑷𝑺. It also assumes the acceleration principle,
i.e., investment is proportional to the rate of change of national income over time.
𝑰𝒕 = 𝒂(𝒀𝒕 − 𝒀𝒕−𝟏 )
where 𝒂 is a constant equal to both the marginal and average capital-output ratios (𝑪𝑶𝑹).
Develop the time path of national income and interpret.
In equilibrium, 𝑰𝒕 = 𝑺𝒕 . Therefore 𝑰𝒕 = 𝒂(𝒀𝒕 − 𝒀𝒕−𝟏 ) becomes:
𝑺𝒕 = 𝒂(𝒀𝒕 − 𝒀𝒕−𝟏 )
𝒔𝒀𝒕 = 𝒂𝒀𝒕 − 𝒂𝒀𝒕−𝟏
𝒔𝒀𝒕 − 𝒂𝒀𝒕 + 𝒂𝒀𝒕−𝟏 = 𝟎
(𝒔 − 𝒂)𝒀𝒕 + 𝒂𝒀𝒕−𝟏 = 𝟎
Normalizing by dividing throughout by (𝒔 − 𝒂):
𝒂 𝟎
𝒀𝒕 + { } 𝒀𝒕−𝟏 =
(𝒔 − 𝒂) (𝒔 − 𝒂)
𝒂
𝒀𝒕 + ( )𝒀 =𝟎
𝒔 − 𝒂 𝒕−𝟏
Introducing a lag to make the given form comparable with the standard form:
𝒂
𝒀𝒕 +𝟏 + ( )𝒀 =𝟎
𝒔 − 𝒂 𝒕−𝟏 +𝟏
______________________________________________________________________________
𝒂
𝒀
⏟𝒕+𝟏 + ( )𝒀 = ⏟
𝟎
⏟𝒔 − 𝒂 ⏟𝒕
𝑦𝑡+1 𝑐
𝑦𝑡
𝑎≠−1
𝒂 𝒂 𝒂
(⏟𝒔−𝒂): (⏟𝒔−𝒂) = −𝟏 ∴ (⏟𝒔−𝒂) ≠ −𝟏 .
𝑎≠−1 𝑖𝑓 𝒔=𝟎,𝑏𝑢𝑡 0<𝒔 𝑎𝑠 𝑝𝑒𝑟 𝑒𝑐𝑜𝑛𝑜𝑚𝑖𝑐
𝑡ℎ𝑒𝑜𝑟𝑦
𝒂
Here; 𝒂 = (𝒔−𝒂), 𝒄 = 𝟎, 𝒚𝒕 = 𝒀𝒕 , 𝒚𝟎 = 𝒀𝟎 , 𝒃 = −𝒂.
𝒂 ≠ −𝟏.
𝒕
(−𝒂)
𝒄 𝒄
𝒚𝒕 = (𝒚𝟎 − )( ⏞
𝒃 ) +
𝟏+𝒂 𝟏+𝒂
𝟎 𝒂 𝒕 𝟎
𝒀𝒕 = {𝒀𝟎 − 𝒂 } {− ( )} + 𝒂
𝟏 + (𝒔 − 𝒂) 𝒔−𝒂 𝟏 + (𝒔 − 𝒂)
𝒂 𝒕
𝒀𝒕 = 𝒀𝟎 ( )
𝒂−𝒔
⏞ 𝒂 𝒕
𝒀𝒕 = 𝒀𝟎 ( ) +⏟
𝟎
⏟ 𝒂−𝒔
𝑌 𝑝
𝑌𝑐
𝒂 𝒕
Interpretation: 𝒀𝒕 = 𝒀𝟎 (𝒂−𝒔) is the definite time-path of the first order difference equation
𝒂
𝒀𝒕 +𝟏 + ( ) 𝒀𝒕−𝟏 +𝟏 = 𝟎 . Dynamic Stability of the time-path, mainly, depends on exponential
𝒔−𝒂
𝒂 𝒕
expression (𝒂−𝒔) . 𝒂 (the capital-output ratio) is usually greater than 𝟏 while 𝒔 (marginal propensity
to save) is lesser than 𝟏 (but also greater than zero):
Mathematically: ⏟(𝒂 > 𝟏)⋀(𝒔 < 𝟏). The inequality (𝒂 > 𝒔) shall be used in explaining the nature
𝒂>𝒔
and pattern of time-path.
Nature {⏟ ⏞
𝒃 }
𝒂 𝑔𝑖𝑣𝑒𝑛 (𝒂>𝒔) 𝒂 𝒂 𝑎𝑠 𝑔𝑖𝑣𝑒𝑛 (𝒂>𝒔) 𝒂
|𝒃| = | |⋚𝟏 ⇒ | |≮𝟏 𝒃=( )⋚𝟎 ⇒ ( )≮𝟎
𝒂−𝒔 ⏟ 𝒂−𝒔 𝒂−𝒔 ⏟ 𝒂−𝒔
𝑻𝒊𝒎𝒆−𝒑𝒂𝒕𝒉 𝒘𝒊𝒍𝒍 𝒃𝒆 𝑵𝒐𝒏−𝒄𝒐𝒏𝒗𝒆𝒓𝒈𝒆𝒏𝒕 𝑻𝒊𝒎𝒆−𝒑𝒂𝒕𝒉 𝒘𝒊𝒍𝒍 𝒃𝒆 𝑵𝒐𝒏−𝑶𝒔𝒄𝒊𝒍𝒍𝒂𝒕𝒐𝒓𝒚
Therefore, the time-path is non-convergent in nature as (𝒂 > 𝒔) while it will move in a non-
TOPIC 173: WARRANTED RATE OF GROWTH
In the Harrod model the warranted rate of growth is of pivotal importance. Where, ‘Warranted
Rate of Growth’ (i.e., the path the economy must follow to have equilibrium between saving and
investment each year).
______________________________________________________________________________
𝒂
Recalling the time-path of national income 𝒀𝒕 = ( ) 𝒀𝟎 . This time-path shows that current year’s
𝒂−𝒔
𝒂
national income is (𝒂−𝒔) times the national income of the previous period. e.g. for current period;
𝒂
𝒀𝟏 = (𝒂−𝒔) 𝒀 𝟎
𝒀𝟏 −𝒀𝟎
Intertemporal growth: 𝑮 = . Substituting the value of 𝒀𝟏 becomes:
𝒀𝟎
𝒀𝟏 − 𝒀𝟎
𝑮=
𝒀𝟎
𝒂
(𝒂 − 𝒔) 𝒀𝟎 − 𝒀𝟎
𝑮=
𝒀𝟎
𝒂
{( ) − 𝟏} 𝒀𝟎
𝑮= 𝒂−𝒔
𝒀𝟎
𝒂
𝑮= −𝟏
𝒂−𝒔
𝒂 − (𝒂 − 𝒔)
𝑮=
𝒂−𝒔
𝑊𝑎𝑟𝑟𝑎𝑛𝑡𝑒𝑑
𝐺𝑟𝑜𝑤𝑡ℎ 𝑟𝑎𝑡𝑒
𝒂−𝒂+𝒔
⏞
𝑮 =
𝒂−𝒔
𝒔
Warranted Growth rate: 𝑮𝒘 = 𝒂−𝒔
______________________________________________________________________________
Lesson 38
COBWEB MODEL AND DIFFERENCE EQUATIONS
TOPIC 174: COBWEB MODEL: FRAMEWORK
Using First Order Difference Equations, we can understand the price fluctuations of products with
lags involved in between production and sale. For instance, agricultural products, in which
planting and harvest occur with substantial gap.
The Model: Cobweb model is based on demand and supply forces:
Demand Side {𝑸𝒅(𝒕) }: The demand side of such products has no lags, since for a buyer:
𝑷𝒕 𝑸𝒅(𝒕)
⏞
𝐏𝐫𝐢𝐜𝐞 𝐨𝐟 𝐜𝐫𝐨𝐩 𝐢𝐧 𝐜𝐮𝐫𝐞𝐧𝐭 𝐭𝐢𝐦𝐞 𝐩𝐞𝐫𝐢𝐨𝐝 ⇒ ⏞
⏟ 𝐃𝐞𝐦𝐚𝐧𝐝 𝐨𝐟 𝐜𝐫𝐨𝐩 𝐢𝐧 𝐜𝐮𝐫𝐫𝐞𝐧𝐭 𝐭𝐢𝐦𝐞 𝐩𝐞𝐫𝐢𝐨𝐝
𝑏𝑢𝑦𝑒𝑟 𝑖𝑠 𝑐𝑜𝑛𝑐𝑒𝑟𝑛𝑒𝑑 𝑤𝑖𝑡ℎ 𝑡ℎ𝑒 𝑐𝑢𝑟𝑟𝑒𝑛𝑡 𝑝𝑟𝑖𝑐𝑒 𝑜𝑓 𝑎𝑔𝑟𝑖𝑐𝑢𝑙𝑡𝑢𝑟𝑎𝑙 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑠
Mathematizing this basic-psychology:
𝑸𝒅(𝒕) =
⏟ ⏟ 𝒕)
⏟ 𝑫(𝑷
𝑐𝑢𝑟𝑟𝑒𝑛𝑡 ⇐ 𝑐𝑢𝑟𝑟𝑒𝑛𝑡
𝑑𝑒𝑚𝑎𝑛𝑑 𝑝𝑟𝑖𝑐𝑒
𝑸𝒅(𝒕) = 𝑫(𝑷𝒕 )
Supply Side {𝑸𝒔(𝒕) }:
𝑷𝒕 𝑸𝒔(𝒕+𝟏)
𝐏𝐫𝐢𝐜𝐞 𝐨𝐟 𝐜𝐫𝐨𝐩 𝐢𝐧 𝐜𝐮𝐫𝐞𝐧𝐭 𝐭𝐢𝐦𝐞 𝐩𝐞𝐫𝐢𝐨𝐝 ⇒ ⏞

⏞
⏟ 𝐒𝐮𝐩𝐩𝐥𝐲(𝐎𝐮𝐭𝐩𝐮𝐭) 𝐨𝐟 𝐜𝐫𝐨𝐩 𝐢𝐧 𝐧𝐞𝐱𝐭 𝐭𝐢𝐦𝐞 𝐩𝐞𝐫𝐢𝐨𝐝
𝑖𝑡 𝑡𝑎𝑘𝑒𝑠 𝑎 𝑠𝑒𝑎𝑠𝑜𝑛 (𝑡𝑖𝑚𝑒 𝑝𝑒𝑟𝑖𝑜𝑑) 𝑓𝑜𝑟 𝑎 𝑐𝑟𝑜𝑝 𝑡𝑜 𝑔𝑒𝑡 𝑟𝑒𝑎𝑑𝑦
Mathematizing this agricultural phenomenon:
𝑸
⏟𝒔(𝒕+𝟏) = ⏟ 𝒕)
⏟ 𝑺(𝑷
𝑓𝑢𝑡𝑢𝑟𝑒 ⇐ 𝑐𝑢𝑟𝑟𝑒𝑛𝑡
𝑠𝑢𝑝𝑝𝑙𝑦 𝑝𝑟𝑖𝑐𝑒
𝑸𝒔 (𝒕+𝟏) = 𝑺(𝑷𝒕 )
⏟
𝑐𝑎𝑛′ 𝑡 𝑏𝑒 𝑒𝑞𝑢𝑎𝑡𝑒𝑑 𝑡𝑜 𝑸𝒅 (𝒕)
𝑑𝑢𝑒 𝑡𝑜 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡 𝑡𝑖𝑚𝑒 𝑝𝑒𝑟𝑖𝑜𝑑𝑠
Therefore, we convert 𝑸𝒔(𝒕+𝟏)into 𝑸𝒔(𝒕) by introducing a lag (−𝟏):
𝑸𝒔(𝒕+𝟏 −𝟏 ) = 𝑺 (𝑷𝒕 −𝟏 )
𝑸𝒔(𝒕) = 𝑺(𝑷𝒕−𝟏 )
Unlagged and Lagged Demand and Supply Functions:
Unlagged Demand Function Lagged Supply Function
𝑸𝒅(𝒕) = 𝑫(𝑷𝒕 ) 𝑸𝒔(𝒕) = 𝑺(𝑷𝒕−𝟏 )
Unlagged Demand Equation Lagged Supply Equation
𝑸𝒅(𝒕) = 𝜶 − 𝜷(𝑷𝒕 ) 𝑸𝒔(𝒕) = −𝜸 + 𝜹(𝑷𝒕−𝟏 )
For equilibrium in (agricultural) market: 𝑸𝒅 = 𝑸𝒔 .
𝑸𝒅(𝒕) = 𝑸𝒔(𝒕)
𝜶 − 𝜷(𝑷𝒕 ) = −𝜸 + 𝜹(𝑷𝒕−𝟏 )
−𝜷(𝑷𝒕 ) − 𝜹(𝑷𝒕−𝟏 ) = −𝜶 − 𝜸
𝜷(𝑷𝒕 ) + 𝜹(𝑷𝒕−𝟏 ) = 𝜶 + 𝜸
Normalizing by dividing throughout by 𝜷:
𝜹 (𝜶 + 𝜸)
(𝑷𝒕 ) + (𝑷𝒕−𝟏 ) =
𝜷 𝜷
______________________________________________________________________________
𝜹 𝜶+𝜸
𝑷𝒕 + (𝑷𝒕−𝟏 ) =
𝜷 𝜷
𝒚 𝒕+𝟏 − 𝒂𝒚 𝒕 = 𝒃
𝜹 𝜶+𝜸
𝑷 𝒕 + (𝑷 𝒕−𝟏 ) =
𝜷 𝜷
Boxed subscripts show discrepancy of time, so we introduce time lead in price equation.
𝜹 𝜶+𝜸
𝑷𝒕 +𝟏 + (𝑷𝒕−𝟏 +𝟏 ) =
𝜷 𝜷
𝜹 𝜶+𝜸
𝑷⏟𝒕+𝟏 + ( ) 𝑷⏟=
⏟𝜷 𝒕 ⏟𝜷
𝑦𝑡+1 𝑦𝑡
𝑎 𝑐
𝜹 𝜶+𝜸
Comparing with the standard form: 𝒂 = 𝜷, 𝒄 = 𝜷 , 𝒚𝒕+𝟏 = 𝑷𝒕+𝟏, 𝒚𝒕 = 𝑷𝒕 .
>0
⏞
𝜹 𝜹
( ) ⇒ ( ) ≠ −𝟏.
𝜷
⏟ 𝜷
⏟>0
𝑎>0
𝒂 ≠ −𝟏.
𝒕
(−𝒂)
𝒄 𝒄
𝒚𝒕 = (𝒚𝟎 − )( ⏞
𝒃 ) +
𝟏+𝒂 𝟏+𝒂
𝜶+𝜸 𝜶+𝜸
() ( )
𝜷 𝒕 𝜷
𝑷𝒕 = {𝑷𝟎 − } (−𝒂) +
𝜹 𝜹
(𝟏 + ) (𝟏 + )
𝜷 𝜷
𝜶+𝜸 𝜶+𝜸
( ) 𝜹 𝒕 ( 𝜷 )
𝜷
𝑷𝒕 = {𝑷𝟎 − } (− ) +
𝜷+𝜹 𝜷 𝜷+𝜹
( ) ( )
𝜷 𝜷
𝑠𝑡𝑎𝑡𝑖𝑜𝑛𝑎𝑟𝑦
(𝜶,𝜷,𝜸,𝜹):𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡𝑠
𝜶+𝜸 𝜹 𝒕 ⏞𝜶 + 𝜸
𝑷𝒕 = 𝑷𝟎 − ( ) (− ) + ( )
⏟𝜷 + 𝜹 𝜷 ⏟ 𝜷+𝜹
{ 𝑃𝑝 } 𝑃𝑝
𝒕
𝜹
𝑷𝒕 = (𝑷𝟎 − 𝑷⏟𝒑 ) (− 𝜷) + 𝑷
⏟𝒑
∗ 𝑃 ∗ 𝑃
𝜹 𝒕
𝑷𝒕 = (𝑷𝟎 − 𝑷∗ ) (− ) + 𝑷∗
𝜷
𝑃𝑐
⏞ 𝑃𝑝
𝜹 𝒕
∗ ⏞∗
(𝑷𝟎 − 𝑷 ) (− ) + 𝑷
𝑷𝒕 = ⏟
⏟ 𝜷
𝐴
𝑏
Role of 𝑨: It can be recalled adopting the following table:
𝑨 = (𝑷𝟎 − 𝑷∗ ) being a coefficient can affect 𝑷𝒑 in either or both of the two ways:
______________________________________________________________________________
Scale Effect: Absolute Value of |𝑷𝟎 − 𝑷∗ | on 𝑷𝒑

Magnitude
|𝑷𝟎 − 𝑷∗ | = 𝟎 |𝑷𝟎 − 𝑷∗ | < 𝟏 |𝑷𝟎 − 𝑷∗ | = 𝟏 |𝑷𝟎 − 𝑷∗ | > 𝟏
Nullify Minify Neutral Amplify
𝜹 𝒕 𝜹 𝒕 𝜹 𝒕 𝜹 𝒕
𝑷𝒄 = 𝟎 (− 𝜷) 𝑷𝒄 = 𝟎. 𝟓 (− 𝜷) 𝑷𝒄 = 𝟏 (− 𝜷) 𝑷𝒄 = 𝟐 (− 𝜷)
𝑷𝒄 = 𝟎 𝜹 𝒕 𝜹 𝒕 𝜹 𝒕
𝑷𝒄 = 𝟎. 𝟓 (− ) 𝑷𝒄 = (− ) 𝒚𝒄 = 𝟐 (− )
𝜷 𝜷 𝜷
- Noteworthy - Noteworthy
Mirror Effect: Original Value of (𝑷𝟎 − 𝑷∗ ) on 𝑷𝒑

Sign
(𝑷𝟎 − 𝑷∗ ) > 𝟎 (𝑷𝟎 − 𝑷∗ ) < 𝟎
Retain Reverse
𝒕
∗)
𝜹 ∗)
𝜹 𝒕
𝑷𝒄 = (𝑷
⏟𝟎 − 𝑷 (− ) 𝑷 𝒄 = (𝑷
⏟𝟎 − 𝑷 (− )
𝜷 𝜷
>0 <0
𝑷𝒄 > 𝟎 𝑷𝒄 < 𝟎
- Noteworthy
TOPIC 175: COBWEB MODEL: DIAGRAMMATICALLY
Using First Order Difference Equation, we found the definite time-path

𝜹 𝒕 𝜹 𝜶+𝜸
𝑷𝒕 = (𝑷𝟎 − 𝑷∗ ) (− ) + 𝑷∗ of price equation 𝑷𝒕+𝟏 + (𝑷𝒕 ) = .
𝜷 𝜷 𝜷
𝑃𝑐
𝑃𝑝
⏞
𝜹 𝒕
⏟ 𝟎 − 𝑷∗ )
𝑷𝒕 = (𝑷 (− 𝜷) ⏞∗
+𝑷
⏟
𝐴
⏟ 𝑏
𝜹 𝒕
Within the boxed term, exponential expression (− 𝜷) is critical for determining the nature and
𝜹
pattern of the time path of 𝑷𝒕 . We borrow the base (− 𝜷) of the expression for elaboration.
Nature {⏟ ⏞
𝒃 }
𝜹 𝜹
|𝒃| = |− | ≶ 𝟏 𝒃 = (− ) ≶ 𝟎
𝜷 𝜷
𝜹<𝜷 𝜹>𝜷 Both 𝜹 and 𝜷 are positive. (𝜹, 𝜷) > 𝟎
𝜹 𝜹 Regardless of (𝜹 ⋚ 𝜷), 𝒃
⏟< 𝟎
|− | < 𝟏 |− | > 𝟏
⏟ 𝜷 ⏟𝜷 𝑂𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦
𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡 𝑁𝑜𝑛−𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡
(𝐷𝑖𝑣𝑒𝑟𝑔𝑒𝑛𝑡)
Therefore, the time-path can be either convergent or non-convergent in nature as (𝜹 ≶ 𝜷), while
it will move in an oscillatory pattern.
Underlying Mechanism of Cobweb Model:
______________________________________________________________________________
No movement along the curves due to discrete nature of 𝑷 & 𝑸.
𝜹
{|− 𝜷| < 𝟏} Convergent and Oscillatory Time-path of Price
AKA: Damped Oscillations
Left Panel shows the converging dynamics in Right Panel plots the time-path of
market. The slope of supply curves should be price (𝑷𝒕 ). It is coverging towards
lesser than that of demand (𝜹 < 𝜷). equilibrium price {ℎ𝑒𝑟𝑒(𝒑𝒆 )}, while
oscillating above and below it.
𝜹
{|− 𝜷| > 𝟏} Non-Convergent (Divergent) and Oscillatory Time-path of Price
AKA: Explosive Oscillations
______________________________________________________________________________
Left Panel shows the diverging Right Panel plots the time-path of price
dynamics in market. The slope of supply (𝑷𝒕 ). It is diverging away from equilibrium
curves should be greater than that of price {ℎ𝑒𝑟𝑒(𝒑𝒆 )}, while oscillating above
demand (𝜹 > 𝜷). and below it.
Conclusion: The nature of time-path is primarily dependent on the relative slope of demand and
supply curves (𝜹 ≶ 𝜷).
TOPIC 176: COBWEB MODEL: RARE CASE
We have understood two common cases of convergence and divergence in cobweb model:
Common Cases: 𝜹 ≶ 𝜷
𝜹<𝜷 𝜹>𝜷
𝜹 𝜹
|− | < 𝟏 |− | > 𝟏
⏟𝜷 ⏟𝜷
𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡 𝑁𝑜𝑛−𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡
(𝐷𝑖𝑣𝑒𝑟𝑔𝑒𝑛𝑡)
Now considering a rare case in cobweb model where (𝜹 = 𝜷):
Rare Case: 𝜹 = 𝜷
𝜹
(− ) = −𝟏
𝜷
Nature of time-path of Rare Case:
Nature {⏟ ⏞
𝒃 }
𝜹=𝜷 𝜹=𝜷
𝜹 𝜹
|𝒃| = |− | = 𝟏 𝒃 = (− ) = −𝟏
𝜷 𝜷
𝜹 𝒃 = −𝟏 < 𝟎
⏟
|− | = 𝟏 ≮ 𝟏 𝑂𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦
⏟𝜷
𝑁𝑜𝑛−𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡
(𝑁𝑒𝑖𝑡ℎ𝑒𝑟)
𝜹
{|− 𝜷| = 𝟏} Non-Convergent (Uniform) and Oscillatory Time-path of Price
______________________________________________________________________________
Left Panel shows the equi-distant dynamics in Right Panel plots the time-path of price (𝑷𝒕 ). It
market. The slope of supply curves should be is unformly oscillating above and below the
equal to that of demand (𝜹 = 𝜷). equilibrium price {ℎ𝑒𝑟𝑒(𝒑𝒆 )}.
Conclusion: The existence of uniform oscillation case is rare event as the equality of slope is a
matter of chance (𝜹 = 𝜷). Any slight inequality can lead to convergence or divergence cases.
TOPIC 177: COBWEB MODEL: NUMERICAL
Given the demand function and Supply Functions:

𝑸𝒅𝒕 = 𝟏𝟖 − 𝟑𝑷𝒕 and 𝑸𝒔𝒕 = −𝟑 + 𝟒𝑷𝒕−𝟏 , find the time-path of price.
Solution:
Unlagged Demand Equation Lagged Supply Equation
𝑸𝒅(𝒕) = 𝜶 − 𝜷(𝑷𝒕 ) 𝑸𝒔(𝒕) = −𝜸 + 𝜹(𝑷𝒕−𝟏 )
𝑸𝒅𝒕 = 𝟏𝟖 − 𝟑(𝑷𝒕 ) 𝑸𝒔(𝒕) = −𝟑 + 𝟒(𝑷𝒕−𝟏 )
Here, 𝜶 = 𝟏𝟖, 𝜷 = 𝟑, 𝜸 = 𝟑 and 𝜹 = 𝟒
For equilibrium in market: 𝑸𝒅 = 𝑸𝒔 .
𝑸𝒅(𝒕) = 𝑸𝒔(𝒕)
𝟏𝟖 − 𝟑𝑷𝒕 = −𝟑 + 𝟒𝑷𝒕−𝟏
−𝟑𝑷𝒕 − 𝟒𝑷𝒕−𝟏 = −𝟑 − 𝟏𝟖
−𝟑𝑷𝒕 − 𝟒𝑷𝒕−𝟏 = −𝟐𝟏
𝟑𝑷𝒕 + 𝟒𝑷𝒕−𝟏 = 𝟐𝟏
Normalizing by dividing throughout by 𝟑:
𝟒
𝑷𝒕 + (𝑷𝒕−𝟏 ) = 𝟕
𝟑
𝟒
𝑷𝒕 + (𝑷𝒕−𝟏 ) = 𝟕
𝟑
𝒚 𝒕+𝟏 − 𝒂𝒚 𝒕 = 𝒃
𝟒
𝑷 𝒕 + (𝑷 𝒕−𝟏 ) = 𝟕
𝟑
Boxed subscripts show discrepancy of time, so we introduce time lead in price equation.
𝟒
𝑷𝒕 +𝟏 + (𝑷𝒕−𝟏 +𝟏 ) = 𝟕
𝟑
______________________________________________________________________________
𝟒
𝑷
⏟𝒕+𝟏 + ( ) 𝑷
⏟=⏟
𝟕
⏟𝟑 𝒕 𝑐
𝑦𝑡+1 𝑦𝑡
𝑎
𝟒
Comparing with the standard form: 𝒂
⏟ = 𝟑, 𝒄 = 𝟕, 𝒚𝒕+𝟏 = 𝑷𝒕+𝟏, 𝒚𝒕 = 𝑷𝒕 .
𝒂≠−𝟏
𝒂 ≠ −𝟏.
𝒕
(−𝒂)
𝒄 𝒄
𝒚𝒕 = (𝒚𝟎 − ⏞ ) +
)( 𝒃
𝟏+𝒂 𝟏+𝒂
𝟕 𝟕
𝑷𝒕 = {𝑷𝟎 − } (−𝒂)𝒕 +
𝟒 𝟒
(𝟏 + 𝟑) (𝟏 + 𝟑)
𝟕 𝟕
𝑷𝒕 = {𝑷𝟎 − } (−𝒂)𝒕 +
𝟕 𝟕
(𝟑) (𝟑)
𝟒 𝒕
𝑷𝒕 = (𝑷𝟎 − 𝟑) (− ) + 𝟑
𝟑
𝐴
𝒕
⏞
𝟒
𝑷𝒕 = (𝑷𝟎 − ⏟
𝟑 ) (− ) + ⏟
𝟑
⏟ 𝟑
𝑃𝑝 𝑃𝑝
⏟ 𝑏
𝑃𝑐
𝟒 𝒕
Interpretation: The time-path of in case of cobweb model is 𝑷𝒕 = (𝑷𝟎 − 𝟑) (− 𝟑) + 𝟑 , if 𝑸𝒅 =
𝟏𝟖 − 𝟑𝑷𝒕 and 𝑸𝒔 = −𝟑 + 𝟒𝑷𝒕−𝟏.
Nature {⏟ ⏞
𝒃 }
𝜹 𝟒 𝜹 𝟒
|𝒃| = |− | = |− | > 𝟏 𝒃 = (− ) = (− ) < 𝟎
⏟ 𝜷 𝟑 ⏟ 𝜷 𝟑
𝑁𝑜𝑛−𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡 (𝐷𝑖𝑣𝑒𝑟𝑔𝑒𝑛𝑡) 𝑂𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦
Hence the time-path is divergent in nature and oscillatory in pattern – dynamically unstable.
______________________________________________________________________________
TOPIC 178: DIFFERENCE EQUATIONS AND ADAPTIVE EXPECTATIONS
Consider the following market model:

𝑸𝒅𝒕 = 𝑸𝒔𝒕
𝑸𝒅𝒕 = −𝜶 + 𝜷𝑷𝒕
𝑸𝒔𝒕 = 𝜸 − 𝜹𝑷∗𝒕
∗
Here, 𝑷𝒕 is the expected price in current time period 𝒕.
Moreover, the type of expectations is Adaptive Expectations.
Mathematically; 𝑷∗𝒕 = 𝑷∗𝒕−𝟏 + 𝜼(𝑷𝒕−𝟏 − 𝑷∗𝒕−𝟏 ) (𝟎 < 𝜼 ≤ 𝟏)
𝑒𝑥𝑝𝑒𝑐𝑡𝑎𝑡𝑖𝑜𝑛−
𝑎𝑑𝑗𝑢𝑠𝑡𝑚𝑒𝑛𝑡
∗ (𝑷𝒕−𝟏 − 𝑷∗𝒕−𝟏 )
⏟∗𝒕
𝑷 = 𝑷
⏟𝒕−𝟏 + ⏞
𝜼 ⏟
𝑒𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑝𝑟𝑖𝑐𝑒 𝑒𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑝𝑟𝑖𝑐𝑒 𝑑𝑖𝑠𝑐𝑟𝑒𝑝𝑎𝑛𝑦 𝑏𝑒𝑡𝑤𝑒𝑒𝑛
𝑖𝑛 𝑐𝑢𝑟𝑟𝑒𝑛𝑡 𝑖𝑛 𝑝𝑟𝑒𝑣𝑖𝑜𝑢𝑠 𝑎𝑐𝑡𝑢𝑎𝑙 𝑎𝑛𝑑 𝑒𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑝𝑟𝑖𝑐𝑒𝑠
𝑡𝑖𝑚𝑒 𝑝𝑒𝑟𝑖𝑜𝑑 𝑡𝑖𝑚𝑒 𝑝𝑒𝑟𝑖𝑜𝑑 𝑖𝑛 𝑝𝑟𝑒𝑣𝑖𝑜𝑢𝑠 𝑡𝑖𝑚𝑒 𝑝𝑒𝑟𝑖𝑜𝑑
Effect of the Adaptive Price Expectations:

If (𝑷𝒕−𝟏 > 𝑷∗𝒕−𝟏 ) If (𝑷𝒕−𝟏 < 𝑷∗𝒕−𝟏 )
Price in Previous time period is greater Price in Previous time period is lesser
than that in Current than that in Current
𝑷𝒕−𝟏 = 𝟐𝟎, 𝑷∗𝒕−𝟏 = 𝟏𝟎
⏟ 𝑷𝒕−𝟏 = 𝟐𝟎, 𝑷∗𝒕−𝟏 = 𝟑𝟎
⏟
𝑢𝑛𝑑𝑒𝑟−𝑒𝑥𝑝𝑒𝑐𝑡𝑎𝑡𝑖𝑜𝑛 𝑜𝑣𝑒𝑟−𝑒𝑥𝑝𝑒𝑐𝑡𝑎𝑡𝑖𝑜𝑛
Also assume that 𝜼 = 𝟎. 𝟓
𝑷∗𝒕 = 𝑷∗𝒕−𝟏 + 𝜼(𝑷𝒕−𝟏 − 𝑷∗𝒕−𝟏 )
∗
𝑷𝒕 = 𝟐𝟎 + 𝟎. 𝟓(𝟐𝟎 − 𝟏𝟎) 𝑷∗𝒕 = 𝟐𝟎 + 𝟎. 𝟓(𝟐𝟎 − 𝟑𝟎)
∗
𝑷𝒕 = 𝟐𝟓 𝑷∗𝒕 = 𝟏𝟓
Due to under-expectations, price Due to over-expectations, price
expectations are revised upwards (from expectations are revised downwards
𝟐𝟎 to 𝟐𝟓). (from 𝟐𝟎 to 𝟏𝟓).
Role of 𝜼: Let 𝜼 = 𝟏
Assuming 𝜼 = 𝟏, we rewrite the price expectations equation as follows:
𝑷∗𝒕 = 𝑷∗𝒕−𝟏 + ⏟
𝜼 (𝑷𝒕−𝟏 − 𝑷∗𝒕−𝟏 )
=𝟏
𝑷∗𝒕 = 𝑷∗𝒕−𝟏 + 𝑷𝒕−𝟏 − 𝑷∗𝒕−𝟏
⏟∗𝒕
𝑷 = 𝑷⏟𝒕−𝟏
𝑒𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑝𝑟𝑖𝑐𝑒 𝑖𝑛 𝑎𝑐𝑡𝑢𝑎𝑙 𝑝𝑟𝑖𝑐𝑒 𝑖𝑛
𝑐𝑢𝑟𝑟𝑒𝑛𝑡 𝑡𝑖𝑚𝑒 𝑝𝑒𝑟𝑖𝑜𝑑 𝑝𝑟𝑒𝑣𝑖𝑜𝑢𝑠 𝑡𝑖𝑚𝑒 𝑝𝑒𝑟𝑖𝑜𝑑
Conclusion: It implies that seller is expecting the price this year to be same as that in last year.
Such situation is observed in Cobweb model. Therefore, we find that adaptive expectations
have a significant role in the working of Cobweb model.
Adaptive Expectations in Differential vs. Difference Equations:
Differential Difference Equations
Equations
𝒅𝝅 𝑷∗𝒕 = 𝑷∗𝒕−𝟏 + 𝜼(𝑷𝒕−𝟏 − 𝑷∗𝒕−𝟏 )
= 𝒋(𝒑 − 𝝅)
𝒅𝒕 𝑷∗𝒕 − 𝑷∗𝒕−𝟏 = 𝜼(𝑷𝒕−𝟏 − 𝑷∗𝒕−𝟏 )
Change 𝒅𝝅 𝑷∗𝒕 − 𝑷∗𝒕−𝟏
𝒅𝒕
Adjustment 𝒋 𝜼
coefficient
______________________________________________________________________________
Discrepancy term (𝒑 − 𝝅) (𝑷𝒕−𝟏 − 𝑷∗𝒕−𝟏 )

Observation: Both, Differential & Difference equations have the same structure
while including the Adaptive Expectations.
______________________________________________________________________________
Lesson 39
A MARKET MODEL WITH INVENTORY USING DIFFERENCE EQUATIONS
TOPIC 179: A MARKET MODEL WITH INVENTORY: FRAMEWORK
In Cobweb model, market was assumed to be cleared at every time period. It is possible either if
product is perishable which can’t be stored or are not stocked.
However, in real life there are many products that can be stored. For such products we develop
a model with possibility of inventory.
Premises of The Model
Let us assume the following:

I. Both the quantity demanded, (𝑸𝒅 )and the quantity (𝑸𝒔 ) are currently produced, are
unlagged linear functions of price (𝑷𝒕 ).
𝑙𝑖𝑛𝑒𝑎𝑟 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝑙𝑖𝑛𝑒𝑎𝑟 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
⏞ ⏞
𝑸𝒅𝒕 = 𝑫(𝑷𝒕 ) and ⏟
⏟ 𝑸𝒔𝒕 = 𝑺(𝑷𝒕 )
𝑢𝑛𝑙𝑎𝑔𝑔𝑒𝑑 𝑢𝑛𝑙𝑎𝑔𝑔𝑒𝑑
𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑐𝑒 𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑐𝑒
II. Adjustment of price is affected not through market clearance in every period, but through
a process of price-setting by the sellers (seller-determined price). At the beginning of each
period, the sellers set a price for that period after taking into consideration the inventory
situation.
𝑝𝑟𝑖𝑐𝑒 𝑎𝑑𝑗𝑢𝑠𝑡𝑚𝑒𝑛𝑡
⏟⏞
(𝑷𝒕 − 𝑷𝒕−𝟏 ) = 𝒇 ⏟
(𝑸𝒔 − 𝑸𝒅 )
≠𝒇(𝑸𝒅 =𝑸𝒔 ) 𝑖𝑛𝑣𝑒𝑛𝑡𝑜𝑟𝑦
Accumulated inventory in preceding-period drops the price in current-period, so the
merchandise could be ‘moved’.
Conversely, a decumulated inventory raises the current price higher than before.
III. The price adjustment made from period to period is inversely proportional to the observed
change in the inventory (stock).
Mathematizing the premises:

Premise – I 𝑸𝒅𝒕 = 𝜶 − 𝜷𝑷𝒕 (𝜶, 𝜷 > 𝟎)
𝑸𝒔𝒕 = −𝜸 − 𝜹𝑷𝒕 (𝜸, 𝜹 > 𝟎)
Premise – II & III 𝑷𝒕+𝟏 = 𝑷𝒕 − 𝝈(𝑸𝒔𝒕 − 𝑸𝒅𝒕 ) (𝝈 > 𝟎)
𝝈: stock-induced-price-adjustment coefficient.
Note: (𝑸𝒔𝒕 − 𝑸𝒅𝒕 ) is considered here as the inventory (stock), whereas (𝑸𝒅𝒕 − 𝑸𝒔𝒕 ) is considered
as excess demand.
TOPIC 180: A MARKET MODEL WITH INVENTORY: SOLUTION
Given the demand, supply and price-adjustment functions:

𝑸𝒅𝒕 = 𝜶 − 𝜷𝑷𝒕 (𝜶, 𝜷 > 𝟎)
𝑸𝒔𝒕 = −𝜸 + 𝜹𝑷𝒕 (𝜸, 𝜹 > 𝟎)
𝑷𝒕+𝟏 = 𝑷𝒕 − 𝝈(𝑸𝒔𝒕 − 𝑸𝒅𝒕 ) (𝝈 > 𝟎)
Substituting the values of 𝑸𝒅𝒕 and 𝑸𝒔𝒕 in price-adjustment equation:
𝑷𝒕+𝟏 = 𝑷𝒕 − 𝝈( 𝑸𝒔𝒕 − 𝑸𝒅𝒕 )
______________________________________________________________________________
𝑷𝒕+𝟏 = 𝑷𝒕 − 𝝈( −𝜸 + 𝜹𝑷𝒕 − 𝜶 − 𝜷𝑷𝒕 )

𝑷𝒕+𝟏 = 𝑷𝒕 − 𝝈(−𝜸 + 𝜹𝑷𝒕 − 𝜶 + 𝜷𝑷𝒕 )
𝑷𝒕+𝟏 = 𝑷𝒕 + 𝝈𝜸 − 𝝈𝜹𝑷𝒕 + 𝝈𝜶 − 𝝈𝜷𝑷𝒕
𝑷𝒕+𝟏 = 𝑷𝒕 − 𝝈𝜷𝑷𝒕 − 𝝈𝜹𝑷𝒕 + 𝝈𝜶 + 𝝈𝜸
𝑷𝒕+𝟏 − 𝑷𝒕 + 𝝈𝜷𝑷𝒕 + 𝝈𝜹𝑷𝒕 = 𝝈𝜶 + 𝝈𝜸
𝑷𝒕+𝟏 − (𝟏 − 𝝈𝜷 − 𝝈𝜹)𝑷𝒕 = 𝝈𝜶 + 𝝈𝜸
𝑷𝒕+𝟏 − {𝟏 − 𝝈(𝜷 + 𝜹)}𝑷𝒕 = 𝝈(𝜶 + 𝜸)
First Order Difference Equation representing a market model with inventory. Elaborating the
equation by comparing with the standard from:
𝑷⏟𝒕+𝟏 + [−{𝟏
⏟ − 𝝈(𝜷 + 𝜹)}] 𝑷
⏟𝒕 = ⏟
𝝈(𝜶 + 𝜸)
Here; 𝒂 = [−{𝟏 − 𝝈(𝜷 + 𝜹)}], 𝒄 = 𝝈(𝜶 + 𝜸), 𝒚𝒕+𝟏 = 𝑷𝒕+𝟏, 𝒚𝒕 = 𝑷𝒕
≠1
⏞ >0
⏞ >0 >0
⏞+⏞ 𝜹
𝒂= − 𝟏− 𝝈
⏟ (𝜷 𝜹) ⇒ (𝜷) ≠ −𝟏.
>0 ⏟
>0
⏟{ }
[ ≠−1 ]
𝒂 ≠ −𝟏.
𝒕
(−𝒂)
𝒄 ⏞ ) + 𝒄
𝒚𝒕 = (𝒚𝟎 − )( 𝒃
𝟏+𝒂 𝟏+𝒂
𝒕
(−𝒂)
𝝈(𝜶+𝜸) ⏞ ) + 𝝈(𝜶+𝜸)
𝑷𝒕 = [𝑷𝟎 − 𝟏+[−{𝟏−𝝈(𝜷+𝜹)}]] ( 𝒃 𝟏+[−{𝟏−𝝈(𝜷+𝜹)}]
𝝈(𝜶+𝜸) 𝒕 𝝈(𝜶+𝜸)
𝑷𝒕 = [𝑷𝟎 − ] [− [−{𝟏 − 𝝈(𝜷 + 𝜹)}]] +
𝟏−𝟏+𝝈(𝜷+𝜹) 𝟏−𝟏+𝝈(𝜷+𝜹)
𝝈(𝜶+𝜸) 𝝈(𝜶+𝜸)
𝑷𝒕 = [𝑷𝟎 − 𝝈(𝜷+𝜹)
] {𝟏 − 𝝈(𝜷 + 𝜹)}𝒕 + 𝝈(𝜷+𝜹)
𝜶+𝜸 𝜶+𝜸
𝑷𝒕 = [𝑷𝟎 − ( )] {𝟏 − 𝝈(𝜷 + 𝜹)}𝒕 + ( )
⏟𝜷+𝜹 ⏟𝜷+𝜹
𝑃∗ 𝑃∗
𝑷𝒕 = (𝑷𝟎 − 𝑷∗ ){𝟏 − 𝝈(𝜷 + 𝜹)}𝒕 + 𝑷∗

𝑃𝑐 𝑒𝑞𝑢𝑖𝑙𝑖𝑏𝑟𝑖𝑢𝑚 (𝑃𝑝 )
𝑷𝒕 = ⏞
(𝑷𝟎 − 𝑷∗ ){𝟏 − 𝝈(𝜷 + 𝜹)}𝒕 + ⏞∗
𝑷
𝑃𝑐 𝑃𝑝
⏞ {𝟏 − 𝝈(𝜷 + 𝜹)}𝒕 + 𝑷
(𝑷𝟎 − 𝑷∗ ) ⏟ ⏞∗
𝑷𝒕 = ⏟
𝐴 𝑏
Interpretation: The definite time-path of market model with inventory is obtained.
______________________________________________________________________________
TOPIC 181: A MARKET MODEL WITH INVENTORY: TYPES OF TIME PATH

𝑃𝑐 𝑃𝑝
⏞ {𝟏 − 𝝈(𝜷 + 𝜹)}𝒕 + 𝑷
(𝑷𝟎 − 𝑷∗ ) ⏟ ⏞∗
𝑷𝒕 = ⏟
𝐴 𝑏
The time-path of price in market with inventory is a First Order Difference Equation. 𝒃 =
{𝟏 − 𝝈(𝜷 + 𝜹)} which can make seven cases(regions) depending on value of 𝒃: (−𝟏 > 𝒃 > 𝟏).
Region – I: (𝒃 > 𝟏) ⇒ Divergent & Non-oscillatory
{𝟏 − 𝝈(𝜷 + 𝜹)} > 𝟏
{𝟏 −𝟏 − 𝝈(𝜷 + 𝜹)} > 𝟏 −𝟏
{−𝝈(𝜷 + 𝜹)} > 𝟎
Dividing throughout by −(𝜷 + 𝜹) & hence reversing the inequality.
𝝈(𝜷 + 𝜹) 𝟎
{− } <
−(𝜷 + 𝜹) −(𝜷 + 𝜹)
𝝈
⏟< 𝟎
𝐴𝑙𝑟𝑒𝑎𝑑𝑦 𝑔𝑖𝑣𝑒𝑛 𝑡ℎ𝑎𝑡
(𝝈>𝟎),𝑠𝑜 𝑡ℎ𝑖𝑠 𝑐𝑎𝑠𝑒 𝑖𝑠 𝑛𝑜𝑡
𝑎𝑐𝑐𝑒𝑝𝑡𝑎𝑏𝑙𝑒 𝑎𝑠 𝑝𝑒𝑟 𝑔𝑖𝑣𝑒𝑛
𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑟𝑖𝑐 𝑟𝑒𝑠𝑡𝑟𝑖𝑐𝑡𝑖𝑜𝑛
Therefore, we rule-out Region – I
Region – II: (𝒃 = 𝟏) ⇒ Non-convergent & Non-oscillatory

{𝟏 − 𝝈(𝜷 + 𝜹)} = 𝟏
−𝝈(𝜷 + 𝜹) = 𝟎
𝟎
𝝈=
−(𝜷 + 𝜹)
𝝈
⏟= 𝟎
𝐴𝑙𝑟𝑒𝑎𝑑𝑦 𝑔𝑖𝑣𝑒𝑛 𝑡ℎ𝑎𝑡
(𝝈>𝟎),𝑠𝑜 𝑡ℎ𝑖𝑠 𝑐𝑎𝑠𝑒 𝑖𝑠 𝑛𝑜𝑡
𝑎𝑐𝑐𝑒𝑝𝑡𝑎𝑏𝑙𝑒 𝑎𝑠 𝑝𝑒𝑟 𝑔𝑖𝑣𝑒𝑛
𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑟𝑖𝑐 𝑟𝑒𝑠𝑡𝑟𝑖𝑐𝑡𝑖𝑜𝑛
Therefore, we rule-out Region – II
Region – III: (𝟎 < 𝒃 < 𝟏) ⇒ Convergent & Non-oscillatory

𝟎 < {𝟏 − 𝝈(𝜷 + 𝜹)} < 𝟏
𝟎 −𝟏 < {𝟏 −𝟏 − 𝝈(𝜷 + 𝜹)} < 𝟏 −𝟏
−𝟏 < {−𝝈(𝜷 + 𝜹)} < 𝟎
−𝟏 < {−𝝈(𝜷 + 𝜹)} < 𝟎
−𝟏 𝝈(𝜷 + 𝜹) 𝟎
> {− } >
−(𝜷 + 𝜹) −(𝜷 + 𝜹) −(𝜷 + 𝜹)
𝟏
> {𝝈} > 𝟎
𝜷+𝜹
𝟏
𝟎<𝝈<
𝜷+𝜹
Region – IV: (𝒃 = 𝟎) ⇒ Pre-convergent & Non-oscillatory

{𝟏 − 𝝈(𝜷 + 𝜹)} = 𝟎
______________________________________________________________________________
𝟏
−𝝈 = −
(𝜷 + 𝜹)
𝟏
𝝈=
𝜷+𝜹
Region – V: (−𝟏 < 𝒃 < 𝟎) ⇒ Convergent & Oscillatory

−𝟏 < {𝟏 − 𝝈(𝜷 + 𝜹)} < 𝟎
−𝟏 −𝟏 < {𝟏 −𝟏 − 𝝈(𝜷 + 𝜹)} < 𝟏 −𝟏
−𝟐 < {−𝝈(𝜷 + 𝜹)} < 𝟎
𝟐 𝝈(𝜷 + 𝜹) 𝟎
− > {− } >
−(𝜷 + 𝜹) −(𝜷 + 𝜹) −(𝜷 + 𝜹)
𝟐
> {𝝈} > 𝟎
(𝜷 + 𝜹)
𝟐
𝟎<𝝈<
𝜷+𝜹
Region – VI: (𝒃 = −𝟏) ⇒ Non-Convergent & Oscillatory

{𝟏 − 𝝈(𝜷 + 𝜹)} = −𝟏
−𝝈(𝜷 + 𝜹) = −𝟏 − 𝟏
−𝝈(𝜷 + 𝜹) = −𝟐
Dividing throughout by −(𝜷 + 𝜹).
𝝈(𝜷 + 𝜹) 𝟐
− =−
−(𝜷 + 𝜹) −(𝜷 + 𝜹)
𝟐
𝝈=
𝜷+𝜹
Region – VII: (𝒃 < −𝟏) ⇒ Divergent & Oscillatory

{𝟏 − 𝝈(𝜷 + 𝜹)} < −𝟏
{𝟏 −𝟏 − 𝝈(𝜷 + 𝜹)} < −𝟏 −𝟏
{−𝝈(𝜷 + 𝜹)} < − 𝟐
𝝈(𝜷 + 𝜹) 𝟐
{− } >−
−(𝜷 + 𝜹) −(𝜷 + 𝜹)
𝟐
𝝈>
𝜷+𝜹
Conclusion: For market model with inventory, we experimented with seven cases(regions), out
of which last five (𝟓) are feasible. While, first two (𝟐) are ruled out. All cases relate stock-induced-
price-adjustment coefficient (𝝈) with slopes of demand and supply curves (𝜷 & 𝜹), respectively.
TOPIC 182: A MARKET MODEL WITH INVENTORY: GRAPHICALLY
In order to graphically depict the all possible cases of market model with inventory, we put them
in one table:
Nature
______________________________________________________________________________
𝒃 = {𝟏 − 𝝈(𝜷 + 𝜹)} 𝝈 Pattern

Divergent
I 𝒃>𝟏 𝝈 < 𝟎 (violating 𝝈 > 𝟎 )
Non-oscillatory
Non-convergent
II 𝒃=𝟏 𝝈 = 𝟎 (violating 𝝈 > 𝟎 )
Non-oscillatory
𝟏 Convergent
III 𝟎<𝒃<𝟏 𝟎<𝝈<
𝜷+𝜹 Non-oscillatory
𝟏 Pre-convergent
IV 𝒃=𝟎 𝝈 = 𝜷+𝜹 (Rectangular Hyperbola)
Non-oscillatory
𝟏 𝟐 Convergent
V −𝟏 < 𝒃 < 𝟎 <𝝈<
𝜷+𝜹 𝜷+𝜹 Oscillatory
𝟐 Non-Convergent
VI 𝒃 = −𝟏 𝝈= (Rectangular Hyperbola)
𝜷+𝜹 Oscillatory
𝟐 Divergent
VII 𝒃 < −𝟏 𝝈>
First two of the cases are ruled-out as they violate the parametric restriction (𝝈 > 𝟎).
We can try to plot cases (III-VII).
Both Case – IV and Case – VI, are equalities and can be plotted with ease.
Both of them are ‘rectangular hyperbola’ functions.
Moving from Case – III to Case – VII, we observe that 𝝈 is increasing.
𝟏 𝟏 𝟏 𝟐 𝟐 𝟐
𝟎<𝝈< 𝝈 = <𝝈< 𝝈 = 𝝈>
𝜷+𝜹 𝜷+𝜹 𝜷+𝜹 𝜷+𝜹 𝜷+𝜹 𝜷+𝜹
𝝈
𝟏 𝟏 𝟏 𝟐 𝟐 𝟐
between 𝟎 & 𝜷+𝜹 equals 𝜷+𝜹 between 𝜷+𝜹 & 𝜷+𝜹 equals 𝜷+𝜹 Greater than 𝜷+𝜹
Accordingly, the regions should be in an ascending order in the graph.
Moreover, 1st quadrant shall be used as both 𝝈 > 𝟎 and (𝜷 + 𝜹) > 𝟎.
Graph: Cases (Regions): III – VII
______________________________________________________________________________
Interpretation: Using Cases III – VII, we plotted a graph. It shows all feasible regions of Market
Model with Inventory in one diagram.
TOPIC 183: A MARKET MODEL WITH INVENTORY: NUMERICAL
Example # 1: Given the demand, supply and price-adjustment functions:

𝑸𝒅𝒕 = 𝟐𝟏 − 𝟐𝑷𝒕
𝑸𝒔𝒕 = −𝟑 + 𝟔𝑷𝒕
𝑷𝒕+𝟏 = 𝑷𝒕 − 𝟎. 𝟑(𝑸𝒔𝒕 − 𝑸𝒅𝒕 )
Find the time-path 𝑷𝒕 and determine its dynamic stability.
Solution: The stock-induced-price-adjustment coefficient (𝝈) is the critical parameter which is
𝝈 = 𝟎. 𝟑 > 𝟎.
⏟
𝑎𝑠 𝑝𝑒𝑟 𝑡ℎ𝑒𝑜𝑟𝑦
Substituting the values of 𝑸𝒅𝒕 and 𝑸𝒔𝒕 in price-adjustment equation:
𝑷𝒕+𝟏 = 𝑷𝒕 − 𝟎. 𝟑( 𝑸𝒔𝒕 − 𝑸𝒅𝒕 )
𝑷𝒕+𝟏 = 𝑷𝒕 − 𝟎. 𝟑( −𝟑 + 𝟔𝑷𝒕 − 𝟐𝟏 − 𝟐𝑷𝒕 )
𝑷𝒕+𝟏 = 𝑷𝒕 + 𝟎. 𝟗−𝟏. 𝟖𝑷𝒕 + 𝟔. 𝟑−𝟎. 𝟔𝑷𝒕
𝑷𝒕+𝟏 = 𝟕. 𝟐−𝟏. 𝟒𝑷𝒕
𝑷𝒕+𝟏 + 𝟏. 𝟒𝑷𝒕 = 𝟕. 𝟐
First Order Difference Equation representing a market model with inventory. Elaborating the
equation by comparing with the standard from:
𝑷
⏟𝒕+𝟏 + 𝟏.
⏟ 𝟒𝑷
⏟𝒕 = 𝟕.
⏟ 𝟐
Here; ⏟
𝒂 = 𝟏. 𝟒, 𝒄 = 𝟕. 𝟐, 𝒚𝒕+𝟏 = 𝑷𝒕+𝟏 , 𝒚𝒕 = 𝑷𝒕
𝒂≠𝟏
______________________________________________________________________________
𝒂 ≠ −𝟏.
𝒕
(−𝒂)
𝒄 𝒄
𝒚𝒕 = (𝒚𝟎 − )( ⏞
𝒃 ) +
𝟏+𝒂 𝟏+𝒂
𝟕. 𝟐 𝟕. 𝟐
𝑷𝒕 = (𝑷𝟎 − ) (−𝟏. 𝟒)𝒕 +
𝟏 + 𝟏. 𝟒 𝟏 + 𝟏. 𝟒
𝑷𝒕 = (𝑷𝟎 − 𝟑)(−𝟏. 𝟒)𝒕 + 𝟑
𝑃𝑐 𝑒𝑞𝑢𝑖𝑙𝑖𝑏𝑟𝑖𝑢𝑚 (𝑃𝑝 )
𝑷𝒕 = ⏞
(𝑷𝟎 − 𝟑)(−𝟏. 𝟒)𝒕 + ⏞
𝟑
𝑃𝑐 𝑃𝑝
𝑷𝒕 = ⏞
𝒕
(𝑷 ⏟ 𝟒) + ⏞
⏟ 𝟎 − 𝟑) (−𝟏. 𝟑
𝐴 𝑏
Interpretation: The definite time-path of market model with inventory is obtained.
Dynamic Stability using 𝒃:
Dynamic stability can be checked using the standard procedure used for First Order Difference
Equations.
Nature {⏟ ⏞
𝒃 }
|𝒃|
⏟ = |−𝟏. 𝟒| > 𝟏 𝒃 = −𝟏. 𝟒 < 𝟎
⏟
𝑁𝑜𝑛−𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡 (𝐷𝑖𝑣𝑒𝑟𝑔𝑒𝑛𝑡) 𝑂𝑠𝑐𝑖𝑙𝑙𝑎𝑡𝑜𝑟𝑦
Hence the time-path is divergent in nature and oscillatory in pattern – dynamically unstable.
Dynamic Stability using 𝝈, 𝜷 and 𝜹: Dynamic stability can also be verified using values of 𝝈, 𝜷
and 𝜹:
Nature
𝒃 = {𝟏 − 𝝈(𝜷 + 𝜹)} 𝝈
Pattern
Divergent
I 𝒃>𝟏 𝝈 < 𝟎 (violating 𝝈 > 𝟎 )
Non-oscillatory
Non-convergent
II 𝒃=𝟏 𝝈 = 𝟎 (violating 𝝈 > 𝟎 )
Non-oscillatory
𝟏 Convergent
III 𝟎<𝒃<𝟏 𝟎<𝝈<
𝟏 Pre-convergent
IV 𝒃=𝟎 𝝈=
𝟏 𝟐 Convergent
V −𝟏 < 𝒃 < 𝟎 <𝝈<
𝜷+𝜹 𝜷+𝜹 Oscillatory
𝟐 Non-Convergent
VI 𝒃 = −𝟏 𝝈=
𝟐 Divergent
VII 𝒃 < −𝟏 𝝈> Oscillatory
𝜷+𝜹
Choosing Case – VII through 𝒃 < −𝟏, and converting the condition to 𝝈, 𝜷, 𝜹.
𝝈 = 𝟎. 𝟑, 𝜷 = 𝟐, 𝜹 = 𝟔. 𝟐 Divergent & Oscillatory
⏟ >
𝝈
0.3
𝜷
⏟+ 𝜹
0.25
Therefore, we get same result of time-path either by using
______________________________________________________________________________
𝒃 or 𝝈, 𝜷, 𝜹.
______________________________________________________________________________
Lesson 40
SECOND ORDER DIFFERENCE EQUATIONS
TOPIC 184: SECOND-ORDER DIFFERENCE EQUATIONS: PARTICULAR INTEGRAL
Similar to Second Order Differential Equations, Difference Equations can also acquire second
order of difference(recurrence).
Understanding Second Difference:
∆𝟐 𝒚𝒕 shall be called second difference of 𝒚𝒕 .
∆𝟐 𝒚𝒕 = ∆(∆𝒚𝒕 )
∆𝟐 𝒚𝒕 = ∆ (𝒚𝒕+𝟏 − 𝒚𝒕 )
∆𝟐 𝒚𝒕 = ∆𝒚𝒕+𝟏 − ∆𝒚𝒕
= (𝒚𝒕+𝟐 − 𝒚𝒕+𝟏 ) − (𝒚𝒕+𝟏 − 𝒚𝒕 )
= 𝒚𝒕+𝟐 − 𝒚𝒕+𝟏 − 𝒚𝒕+𝟏 + 𝒚𝒕
∆𝟐 𝒚𝒕 = 𝒚𝒕+𝟐 − 𝟐𝒚𝒕+𝟏 + 𝒚𝒕
Second difference of 𝒚𝒕 has an elaborated expression involving lead/lag of 𝟐, lead/lag of 𝟏 and a
term with no lead/lag.
Standard Form of Second Order Difference Equation: A linear, non-homogeneous version
with constant coefficients (𝒂𝟏 , 𝒂𝟐 ) and constant term 𝒄.
𝒚𝒕+𝟐 + 𝒂𝟏 𝒚𝒕+𝟏 + 𝒂𝟐 𝒚𝒕 = 𝒄
𝒏𝒅 𝒏𝒅
Order: 𝟐 (due to 𝟐 difference)
Degree: Linear/First degree (due to highest power ‘𝟏’ of the 𝟐𝒏𝒅 difference) Constant Coefficients
(𝒂′ 𝑠) and Constant term (𝒄).
Delimitations:
6. Cross terms – product of function ′𝒚𝒕 ′ and its differences (leads/lags) 𝒚𝒕+𝟏 – are not
included.
Possible Cases: Depending on the values of 𝑐, we get 2 possible broad cases:
Main Cases based on value of (𝒄)
Identification 𝒄=𝟎 𝒄≠𝟎
Terminology Homogenous Case Non-Homogenous Case
Components of 𝒚𝒑 = 𝟎 𝒚𝒑 ≠ 𝟎
Solution
Sub-Cases based on values of (𝒂𝟏 ) and/or (𝒂𝟐 )
Identification 𝒂𝟐 ≠ 𝟎, 𝒂𝟐 = 𝟎, 𝒂𝟐 = 𝟎, 𝒂𝟐 ≠ 𝟎, 𝒂𝟐 = 𝟎, 𝒂𝟐 = 𝟎,
(𝒂𝟏 ≠ 𝟎) 𝒂𝟏 ≠ 𝟎 𝒂𝟏 = 𝟎 (𝒂𝟏 ≠ 𝟎) 𝒂𝟏 ≠ 𝟎 𝒂𝟏 = 𝟎
{ ⋁ } { ⋁ }
(𝒂𝟏 = 𝟎) (𝒂𝟏 = 𝟎)
Particular Solution (𝟎) (𝟎) 𝒚𝒑 = 𝒃 𝒃 𝒃
𝒚𝒑 = 𝒚𝒑 = 𝒕 𝒚𝒑 = 𝒂 𝒚𝒑 = 𝒂 𝒕 𝒚𝒑 = 𝟐 𝒕𝟐
𝒂𝟐 𝒂𝟏 (𝟎)
(𝒚𝒑 ) 𝒕𝟐
𝟐 𝟏
𝟐
Particular Integral (𝒚𝒑 ): Any particular solution (containing no arbitrary constants) of complete
equation. 𝒚𝒑 provides equilibrium value of 𝒚𝒕 , in intertemporal sense:
______________________________________________________________________________
𝒚⏟𝒕 = 𝒚𝒄
⏟ + 𝒚
⏟𝒑
𝑒𝑞𝑢𝑖𝑙𝑖𝑏𝑟𝑖𝑢𝑚 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝒚𝒕
Subcase – I: Since the particular integral (𝒚𝒑 ) can be any solution, i.e., any value of 𝒚 that
satisfies this nonhomogeneous equation, we should always try the simplest possible type;
namely, 𝒚 = 𝒂 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕. If 𝒚 = 𝒌, it follows that:
𝒚 = 𝒚𝒕 = 𝒌
𝒚𝒕+𝟏 = 𝒌
𝒚𝒕+𝟐 = 𝒌
Substituting the values of 𝒚𝒕 , 𝒚𝒕+𝟏 and 𝒚𝒕+𝟐 in the general form 𝒚𝒕+𝟐 + 𝒂𝟏 ∙ 𝒚𝒕+𝟏 + 𝒂𝟐 ∙ 𝒚𝒕 = 𝒄:
𝒌 + 𝒂𝟏 ∙ 𝒌 + 𝒂𝟐 ∙ 𝒌 = 𝒄
𝒌(𝟏 + 𝒂𝟏 + 𝒂𝟐 ) = 𝒄
𝒄
𝒌=
(𝟏 + 𝒂𝟏 + 𝒂𝟐 )
Since 𝒌 is a particular value of 𝒚, that satisfies the difference equation, we can call it Particular
Solution (𝒚𝒑 ):
𝒄
𝒚𝒑 =
𝟏 + 𝒂𝟏 + 𝒂𝟐
𝒄
𝒚𝒑 =
𝟏+⏟ 𝒂𝟏 + 𝒂𝟐
≠−1
Interpretation: Particular Integral has actually a constant value as 𝒄 and 𝒂𝟏 , 𝒂𝟐 are constants
and so will be their ratio.
Caveat: For a defined Particular Solution (𝒚𝒑 ), 𝒂𝟏 + 𝒂𝟐 ≠ −𝟏 .
𝒄
Subcase – II: If 𝒂𝟏 + 𝒂𝟐 = −𝟏 , the formula of particular solution (𝒚𝒑 = 𝟏+ 𝒂 ) shall give
𝟏 +𝒂𝟐
𝒄
undefined answer. i.e. (𝒚𝒑 = 𝟎
= ∞). Therefore, we need some other formula for (𝒚𝒑 ). For
another formula, we need another assumption regarding solution value of 𝒚𝒕 . Instead of
considering constant a solution 𝒚, we try a nonconstant (variable); i.e. 𝒚 = 𝒌𝒕.
𝒚𝒕 = 𝒌𝒕
𝒚𝒕+𝟏 = 𝒌(𝒕 + 𝟏)
𝒚𝒕+𝟐 = 𝒌(𝒕 + 𝟐)
Substituting the values of 𝒚𝒕 , 𝒚𝒕+𝟏 , 𝒚𝒕+𝟐 and in the general form 𝒚′′ (𝒕) + 𝒂𝟏 ∙ 𝒚′ (𝒕) + 𝒂𝟐 ∙ 𝒚(𝒕) =
𝒄:
𝒚𝒕+𝟐 + 𝒂𝟏 ∙ 𝒚𝒕+𝟏 + 𝒂𝟐 ∙ 𝒚𝒕 = 𝒄
𝒌(𝒕 + 𝟐) + 𝒂𝟏 ∙ 𝒌(𝒕 + 𝟏) + 𝒂𝟐 ∙ 𝒌𝒕 = 𝒄
𝒌 {(𝒕 + 𝟐) + 𝒂𝟏 ∙ (𝒕 + 𝟏) + 𝒂𝟐 ∙ 𝒕} = 𝒄
𝒌 {(𝒕 + 𝟐) + (𝒂𝟏 𝒕 + 𝒂𝟏 ) + 𝒂𝟐 𝒕} = 𝒄
𝒌 (𝒕 + 𝒂𝟏 𝒕 + 𝒂𝟐 𝒕 + 𝒂𝟏 + 𝟐) = 𝒄
=−1
𝒌 {(𝟏 + ⏞
𝒂𝟏 + 𝒂𝟐 ) 𝒕 + 𝒂𝟏 + 𝟐} = 𝒄
______________________________________________________________________________
𝒌 [{𝟏 + (−𝟏)}𝒕 + 𝒂𝟏 + 𝟐] = 𝒄
𝒌 [{𝟎}𝒕 + 𝒂𝟏 + 𝟐] = 𝒄
𝒌[𝒂𝟏 + 𝟐] = 𝒄
𝒄
𝒌 =
𝒂𝟏 + 𝟐
Since the assumed solution value of 𝒚 was 𝒚 = 𝒌 𝒕, we substitute the value of 𝒌 to get the
𝒄
𝒚= 𝒕
𝒂𝟏 + 𝟐
𝒄
𝒚𝒑 = ( )𝒕
𝒂
⏟𝟏 + 𝟐
≠−2
Interpretation: Particular Solution has actually a nonconstant value regardless of 𝒄, 𝒂𝟏 and 𝒂𝟐
𝒄
being constants. It is due to 𝒕 that gets multiplied with the ratio (𝒂 +𝟐). In this subcase;
𝟏
𝒂𝟏 + 𝒂𝟐 = −𝟏 is permitted.
Caveat: For a defined Particular Solution (𝒚𝒑 ), 𝒂𝟏 ≠ −𝟐 .
Subcase – III: If 𝒂𝟐 = −𝟐 and already 𝒂𝟏 + 𝒂𝟐 = −𝟏, we check if Subcase – I and Subcase – II

formula of particular solution are applicable:
Subcase – I Subcase – II
𝒄 𝒄
𝒚𝒑 = 𝒚𝒑 = 𝒕
𝟏 + 𝒂𝟏 + 𝒂𝟐 𝒂𝟏 + 𝟐
Substituting values of Subcase – III (𝒂𝟏 = −𝟐, 𝒂𝟏 + 𝒂𝟐 = −𝟏)
𝒄 𝒄
𝒚𝒑 = = ∞ 𝒚𝒑 = 𝒕 = ∞
𝟎 𝟎
Therefore, the particular integral using both Subcase – III give an undefined answer.
We need another formula to find the value of 𝒚𝒑 .
Building on the assumptions 𝒚 = 𝒌 and 𝒚 = 𝒌𝒕, we use 𝒚 = 𝒌𝒕𝟐 in Subcase – III.
𝒚𝒕 = 𝒌𝒕𝟐
𝒚𝒕+𝟏 = 𝒌(𝒕 + 𝟏)𝟐
𝒚𝒕+𝟐 = 𝒌(𝒕 + 𝟐)𝟐
Substituting the values of 𝒚𝒕 , 𝒚𝒕+𝟏 , 𝒚𝒕+𝟐 , and 𝒂𝟏 in the standard form: 𝒚𝒕+𝟐 + 𝒂𝟏 ∙ 𝒚𝒕+𝟏 + 𝒂𝟐 ∙
𝒚𝒕 = 𝒄:
𝒌(𝒕 + 𝟐)𝟐 + 𝒂𝟏 ∙ 𝒌(𝒕 + 𝟏)𝟐 + 𝒂𝟐 ∙ 𝒌𝒕𝟐 = 𝒄
𝒌(𝒕𝟐 + 𝟒𝒕 + 𝟒) + 𝒂𝟏 ∙ 𝒌(𝒕𝟐 + 𝟐𝒕 + 𝟏) + 𝒂𝟐 ∙ 𝒌𝒕𝟐 = 𝒄
𝒌 {(𝒕𝟐 + 𝟒𝒕 + 𝟒) + 𝒂𝟏 ∙ (𝒕𝟐 + 𝟐𝒕 + 𝟏) + 𝒂𝟐 ∙ 𝒕𝟐 } = 𝒄
𝒌(𝒕𝟐 + 𝒂𝟏 𝒕𝟐 + 𝒂𝟐 𝒕𝟐 + 𝟒𝒕 + 𝟐𝒂𝟏 𝒕 + 𝟒 + 𝒂𝟏 ) = 𝒄
=−𝟏 =−𝟐 =−𝟐
𝒌 {𝒕𝟐 (𝟏 + ⏞ ⏞𝟏 ) + 𝟒 + 𝒂
𝒂𝟏 + 𝒂𝟐 ) + 𝒕 (𝟒 + 𝟐 𝒂 ⏞𝟏 } = 𝒄
______________________________________________________________________________
𝒌 [𝒕𝟐 ⏟
{𝟏 + (−𝟏)} + 𝒕 ⏟
{𝟒 + 𝟐(−𝟐)} + ⏟
𝟒 + (−𝟐)] = 𝒄
0 0 𝟐
𝒌[𝟐] = 𝒄
𝒄
𝒌 =
𝟐
Since the assumed solution value of 𝒚 was 𝒚 = 𝒌 𝒕𝟐 , we substitute the value of 𝒌 to get the
𝒚 = 𝒌 𝒕𝟐
𝒄
𝒚𝒑 = ( ) 𝒕𝟐
𝟐
Interpretation: Particular Solution has actually a nonconstant value regardless of 𝒄, 𝒂𝟏 and 𝒂𝟐
𝒄
being constants. It is due to 𝒕𝟐 that gets multiplied with the ratio ( ). In this subcase; both
𝟐
𝒂𝟏 + 𝒂𝟐 = −𝟏 and 𝒂𝟏 = −𝟐 are permitted.
Caveat: Since (𝟐), a constant, is in the denominator, the Particular Solution (𝒚𝒑 ), shall have a
defined answer despite 𝒂𝟏 + 𝒂𝟐 = −𝟏 and 𝒂𝟏 = −𝟐.
TOPIC 185: SECOND-ORDER DIFFERENCE EQUATIONS: YP NUMERICALS
Non-Homogenous Subcase – I: 𝒄 ≠ 𝟎, 𝒂𝟏 + 𝒂𝟐 ≠ −𝟏.

Example: Consider a Second Order Difference Equation:
𝒚𝒕+𝟐 − 𝟑𝒚𝒕+𝟏 + 𝟒𝒚𝒕 = 𝟏𝟐
𝒚𝒕+𝟐 − 𝟑𝒚𝒕+𝟏 + 𝟒𝒚𝒕 = 𝟏𝟐
𝒚𝒕+𝟐 + (−𝟑)
⏟ × 𝒚𝒕+𝟏 + ⏟
𝟒 × 𝒚𝒕 = 𝟏𝟐
⏟
𝒂𝟏 𝒂𝟐 𝑐
𝑺𝒖𝒃𝒄𝒂𝒔𝒆−𝑰 𝑪𝒂𝒔𝒆
Here; ⏞
(𝒂𝟏 + 𝒂𝟏 = −𝟑 + 𝟒 = 𝟏 ≠ −𝟏), ⏞ 𝒄 = 𝟏𝟐 ≠ 𝟎 .
𝒄
Suitable Formula for Particular Solution: 𝒚𝒑 = 𝟏+𝒂 +𝒂
𝟏 𝟐
𝟏𝟐
𝒚𝒑 =
𝟏 + (−𝟑) + 𝟒
𝒚𝒑 = 𝟔
Interpretation: 𝒚𝒑 = 𝟔 is particular solution of the Non-Homogenous Subcase – I of Second
Order Difference Equation 𝒚𝒕+𝟐 − 𝟑𝒚𝒕+𝟏 + 𝟒𝒚𝒕 = 𝟏𝟐 .
Non-Homogenous Case – II: 𝒄 ≠ 𝟎, 𝒂𝟏 + 𝒂𝟐 = −𝟏, 𝒂𝟏 ≠ −𝟐.
Example: Consider a Second Order Difference Equation:
𝒚𝒕+𝟐 + 𝒚𝒕+𝟏 − 𝟐𝒚𝒕 = 𝟏𝟐
𝒚𝒕+𝟐 + (𝟏)
⏟ 𝒚𝒕+𝟏 + (−𝟐)
⏟ 𝒚𝒕 = 𝟏𝟐
⏟
𝑎1 𝑎2 𝑐
𝑺𝒖𝒃𝒄𝒂𝒔𝒆−𝑰𝑰 𝑪𝒂𝒔𝒆
Here; ⏞
(𝒂𝟏 = 𝟏), (𝒂𝟐 = −𝟐), ⏞ 𝒄 = 𝟏𝟐 ≠ 𝟎 .
𝒄
Suitable Formula for Particular Solution: 𝒚𝒑 = (𝒂 +𝟐) 𝒕
𝟏
______________________________________________________________________________
𝟏𝟐
𝒚𝒑 = ( )𝒕
𝟏+𝟐
𝒚𝒑 = 𝟒𝒕
Interpretation: 𝒚𝒑 = 𝟒𝒕 is particular solution of the Non-Homogenous Case – II of Second
Order Differential Equation 𝒚𝒕+𝟐 + 𝒚𝒕+𝟏 − 𝟐𝒚𝒕 = 𝟏𝟐 .
Non-Homogenous Case – III: 𝒄 ≠ 𝟎, 𝒂𝟏 + 𝒂𝟐 = −𝟏, 𝒂𝟏 = −𝟐.
𝒚𝒕+𝟐 − 𝟐𝒚𝒕+𝟏 + 𝒚𝒕 = 𝟏𝟐
𝒚𝒕+𝟐 + 𝒂𝟏 𝒚𝒕+𝟏 + 𝒂𝟐 𝒚𝒕 = 𝒃
𝒚𝒕+𝟐 + (−𝟐)
⏟ 𝒚𝒕+𝟏 + (𝟏) ⏟ 𝒚𝒕 = 𝟏𝟐⏟
𝑎1 𝑎2 𝑐
𝑺𝒖𝒃𝒄𝒂𝒔𝒆−𝑰𝑰𝑰 𝑪𝒂𝒔𝒆
Here; ⏞
(𝒂𝟏 = −𝟐), (𝒂𝟐 = 𝟏), ⏞ 𝒃 = 𝟏𝟐 ≠ 𝟎 .
𝒄
Suitable Formula for Particular Solution: 𝒚𝒑 = 𝟐 𝒕𝟐
𝟏𝟐
𝒚𝒑 = × 𝒕𝟐
𝟐
𝒚𝒑 = 𝟔𝒕𝟐
Interpretation: 𝒚𝒑 = 𝟔𝒕𝟐 is particular solution of the Non-Homogenous Case – III of Second
Order Differential Equation 𝒚𝒕+𝟐 − 𝟐𝒚𝒕+𝟏 + 𝒚𝒕 = 𝟏𝟐 .
𝒚𝒕+𝟐 − 𝟑𝒚𝒕+𝟏 + 𝟒𝒚𝒕 = 𝟏𝟐 𝒚𝒕+𝟐 + 𝒚𝒕+𝟏 − 𝟐𝒚𝒕 = 𝟏𝟐 𝒚𝒕+𝟐 − 𝟐𝒚𝒕+𝟏 + 𝒚𝒕 = 𝟏𝟐
𝒄 𝒄 𝒄
𝒚𝒑 = 𝒚𝒑 = 𝒕 𝒚𝒑 = 𝒕𝟐
𝟏 + 𝒂𝟏 + 𝒂𝟐 𝒂𝟏 + 𝟐 𝟐
In Non-Homogenous case, 𝒃 = 𝟏𝟐
𝟏𝟐 𝟏𝟐 𝟏𝟐 𝟐
𝟏 + (−𝟑) + 𝟒 𝟏+𝟐 𝟐
𝒚𝒑 = 𝟔 𝒚𝒑 = 𝟒𝒕 𝒚𝒑 = 𝟔𝒕𝟐
HOMOGENOUS SUBCASES
Since formulae of all subcases of 𝒚𝒑 have 𝒄 in numerator:
𝒚𝒕+𝟐 − 𝟑𝒚𝒕+𝟏 + 𝟒𝒚𝒕 = 𝟎 𝒚𝒕+𝟐 + 𝒚𝒕+𝟏 − 𝟐𝒚𝒕 = 𝟎 𝒚𝒕+𝟐 − 𝟐𝒚𝒕+𝟏 + 𝒚𝒕 = 𝟎
𝒄 𝒄 𝒄
𝒚𝒑 = 𝒚𝒑 = 𝒕 𝒚𝒑 = 𝒕𝟐
𝟏 + 𝒂𝟏 + 𝒂𝟐 𝒂𝟏 + 𝟐 𝟐
In Non-Homogenous case, 𝒃 = 𝟎
𝟎 𝟎 𝟎 𝟐
𝟏 + (−𝟑) + 𝟒 𝟏+𝟐 𝟐
𝒚𝒑 = 𝟎 𝒚𝒑 = 𝟎 𝒚𝒑 = 𝟎
All subcases have a zero value of particular solution; i.e. (𝒚𝒑 = 𝟎).
______________________________________________________________________________
TOPIC 186: SECOND-ORDER DIFFERENCE EQUATIONS: YC DISTINCT REAL ROOTS
Complementary Function (𝒚𝒄 ): General solution (containing arbitrary constants) of reduced

equation. 𝒚𝒄 provides deviation of time path 𝒚(𝒕) from equilibrium value, at every point in time.
𝒚(𝒕)
⏟ = 𝒚
⏟𝒄 + 𝒚
⏟𝒑
Adaptation of the Complementary Function (𝒚𝒄 ):

For the complementary function (𝑦𝑐 ), we can adopt from the second order differential equations.
The natural exponential expression is replaced with exponential function for this purpose:
Second Order Differential Equation Second Order Difference Equation

𝒓𝟏 𝒕 𝒓𝟐 𝒕 (𝒃𝟏 )𝒕 (𝒃𝟐 )𝒕
𝒚𝒄 = 𝑨𝟏 𝒆⏟ + 𝑨𝟐 𝒆⏟ 𝒚𝒄 = 𝑨𝟏 ⏟ + 𝑨𝟐 ⏟
𝑛𝑎𝑡𝑢𝑟𝑎𝑙 𝑛𝑎𝑡𝑢𝑟𝑎𝑙 𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙 𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙
𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙 𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
Numerical Example: Find the solution of 𝒚𝒕+𝟐 + 𝒚𝒕+𝟏 − 𝟐𝒚𝒕 = 𝟏𝟐.

Solution: Elaborating the given equation:
𝒚𝒕+𝟐 + (𝟏)
⏟ 𝒚𝒕+𝟏 + (−𝟐)
⏟ 𝒚𝒕 = 𝟏𝟐
⏟
𝑎1 𝑎2 𝑐
𝒄
𝒚𝒑 = ( )𝒕
⏟ 𝒂𝟏 + 𝟐
𝑎1 +𝑎2 =−1,𝑎1 =−2
𝟏𝟐
𝒚𝒑 = ( )𝒕
𝟏+𝟐
𝒚𝒑 = 𝟒𝒕
______________________________________________________________________________
{(𝟏)𝟐 − 𝟒(−𝟐)} ⋛ 𝟎
𝟗
⏟> 𝟎
𝟗
⏟> 𝟎
−𝒂𝟏 ± √𝒂𝟐𝟏 − 𝟒𝒂𝟐
𝒃𝟏 , 𝒃𝟐 =
𝟐
−(𝟏) ± √(𝟏)𝟐 − 𝟒(−𝟐)
𝒃𝟏 , 𝒃𝟐 =
𝟐
−𝟏 ± √𝟗
𝒃𝟏 , 𝒃𝟐 =
𝟐
−𝟏 + 𝟑 −𝟏 − 𝟑
𝒃𝟏 = , 𝒃𝟐 =
𝟐 𝟐
𝒃𝟏 = 𝟏 or 𝒃𝟐 = −𝟐
Complementary Function: 𝒚𝒄 = 𝑨𝟏 (𝒃𝟏 )𝒕 + 𝑨𝟐 (𝒃𝟐 )𝒕
𝒚𝒄 = 𝑨𝟏 (𝟏)𝒕 + 𝑨𝟐 (−𝟐)𝒕
𝒚𝒄 = 𝑨𝟏 + 𝑨𝟐 (−𝟐)𝒕
𝒚(𝒕) = ⏟ 𝑨𝟏 + 𝑨𝟐 (−𝟐)𝒕 + 𝟒𝒕
⏟
𝒚𝒄 𝒚𝒑
Definite Solution: For definite solution, we substitute 𝒕 = 𝟎 in the General Solution; 𝒚(𝒕) = 𝑨𝟏 +
𝑨𝟐 (−𝟐)𝒕 + 𝟒𝒕:
⏟ = 𝑨𝟏 + 𝑨𝟐 (−𝟐)(𝟎) + 𝟒(𝟎)
𝒚(𝟎)
⏟ = 𝑨𝟏 + 𝑨𝟐 (𝟏) + 𝟎
𝒚(𝟎)
4
𝟒 = 𝑨𝟏 + 𝑨 𝟐
𝟒 = 𝑨𝟏 + 𝑨 𝟐
As there are two unknowns 𝑨𝟏 and 𝑨𝟐 , we need two (simultaneous) equations to evaluate them.
So, develop Second equation putting next value of 𝒕 i.e. (𝒕 = 𝟏) in 𝒚𝒕 , as 𝒚𝟏 is given.
𝒚𝒕 = 𝑨𝟏 + 𝑨𝟐 (−𝟐)𝒕 + 𝟒𝒕
𝒚⏟𝟏 = 𝑨𝟏 + 𝑨𝟐 (−𝟐)(𝟏) + 𝟒(𝟏)
𝒚⏟𝒕 = 𝑨𝟏 − 𝟐𝑨𝟐 + 𝟒
5
𝟓 = 𝑨𝟏 − 𝟐𝑨𝟐 + 𝟒
𝟏 = 𝑨𝟏 − 𝟐𝑨𝟐
Simultaneously solving the two equations with 𝑨𝟏 and 𝑨𝟐 :
𝟒 = 𝑨𝟏 + 𝑨 𝟐
⏟
𝑨𝟏 =𝟒−𝑨𝟐
𝟏 = 𝑨𝟏 − 𝟐𝑨𝟐
𝟏 = 𝟒 − 𝑨𝟐 − 𝟐𝑨𝟐
𝟏 = 𝟒 − 𝟑𝑨𝟐
______________________________________________________________________________
𝑨𝟐 = 𝟏
For evaluating 𝑨𝟏 , put 𝑨𝟐 = 𝟏 in 𝑨𝟏 = 𝟒 − 𝑨𝟐
𝑨𝟏 = 𝟒 − 𝑨 𝟐
𝑨𝟏 = 𝟒 − 𝟏
𝑨𝟏 = 𝟑
Returning value of 𝑨𝟏 and 𝑨𝟐 in General Solution 𝒚(𝒕) = 𝟑 + 𝟏 (−𝟐)𝒕 + 𝟒𝒕 to get the Definite
Solution:
𝒚(𝒕) = 𝟑 + (−𝟐)𝒕 + 𝟒𝒕
Interpretation: Definite Solution of Second Order Difference Equation 𝒚𝒕+𝟐 + 𝒚𝒕+𝟏 − 𝟐𝒚𝒕 = 𝟏𝟐
with case of Distinct Real Roots is 𝒚(𝒕) = 𝟑 + (−𝟐)𝒕 + 𝟒𝒕 .
TOPIC 187: SECOND-ORDER DIFFERENCE EQUATIONS: YC REPEATED REAL ROOTS
Adaptation of the Complementary Function (𝒚𝒄 ):


𝒓𝟏 𝒕 𝒓𝟐 𝒕 (𝒃𝟏 )𝒕 (𝒃𝟐 )𝒕
𝒚𝒄 = 𝑨𝟑 𝒆⏟ + 𝑨𝟒 × 𝒕 × 𝒆⏟ 𝒚𝒄 = 𝑨𝟑 ⏟ + 𝑨𝟒 × 𝒕 × ⏟
𝑛𝑎𝑡𝑢𝑟𝑎𝑙 𝑛𝑎𝑡𝑢𝑟𝑎𝑙 𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙 𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙
𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙 𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
Numerical Example: Find the solution of 𝒚𝒕+𝟐 + 𝟔𝒚𝒕+𝟏 + 𝟗𝒚𝒕 = 𝟒.

𝒚𝒕+𝟐 + (𝟔)
⏟ 𝒚𝒕+𝟏 + (𝟗)
⏟ 𝒚𝒕 = ⏟
𝟒
𝑎1 𝑎2 𝑐
𝒄
𝒚𝒑 = ( )
⏟ 𝟏 + 𝒂𝟏 + 𝒂𝟐
𝑎1 +𝑎2 ≠−1
𝟒
𝒚𝒑 = ( )
𝟏+𝟔+𝟗
𝟏
𝒚𝒑 =
𝟒
______________________________________________________________________________
{(𝟔)𝟐 − 𝟒(𝟗)} ⋛ 𝟎
𝟎
⏟= 𝟎
𝟎
⏟= 𝟎
−𝒂𝟏 ± √𝒂𝟐𝟏 − 𝟒𝒂𝟐
𝒃𝟏 , 𝒃𝟐 =
𝟐
−(𝟔) ± √(𝟔)𝟐 − 𝟒(𝟗)
𝒃𝟏 , 𝒃𝟐 =
𝟐
−𝟔 ± √𝟎
𝒃𝟏 , 𝒃𝟐 =
𝟐
−𝟔
𝒃𝟏 =
𝟐
𝒃𝟏 = 𝟑 or 𝒃𝟐 = −𝟑 ⇒ 𝒃 = −𝟑
Complementary Function: 𝒚𝒄 = 𝑨𝟑 (𝒃)𝒕 + 𝑨𝟒 𝒕(𝒃)𝒕
𝒚𝒄 = 𝑨𝟑 (−𝟑)𝒕 + 𝑨𝟒 𝒕(−𝟑)𝒕
𝒚𝒄 = 𝑨𝟑 (−𝟑)𝒕 + 𝑨𝟒 𝒕(−𝟑)𝒕
𝟏
𝒚(𝒕) = ⏟𝑨𝟑 (−𝟑)𝒕 + 𝑨𝟒 𝒕(−𝟑)𝒕 +
⏟
𝟒
𝒚𝒄
𝒚𝒑
Interpretation: General Solution of Second Order Difference Equation 𝒚𝒕+𝟐 + 𝟔𝒚𝒕+𝟏 + 𝟗𝒚𝒕 = 𝟒
𝟏
with case of Repeated Real Roots is 𝒚(𝒕) = 𝑨𝟑 (−𝟑)𝒕 + 𝑨𝟒 𝒕(−𝟑)𝒕 + 𝟒 .
Note: Definite solution can also be found, if initial conditions 𝒚𝟎 and 𝒚𝟏 are given.
______________________________________________________________________________
TOPIC 188: SECOND-ORDER DIFFERENCE EQUATIONS: YC COMPLEX ROOTS

𝒚𝒄 = 𝒉𝒕
𝒆⏟ (𝑨𝟓 𝐜𝐨𝐬 𝝊𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝝊𝒕 ) 𝒚𝒄 = ⏟𝒕
𝑹 (𝑨𝟓 𝐜𝐨𝐬 𝜽𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝜽𝒕 )
𝑛𝑎𝑡𝑢𝑟𝑎𝑙 𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙
𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
𝟏
Numerical Example: Find the solution of 𝒚𝒕+𝟐 + 𝟒 𝒚𝒕 = 𝟓.
𝑠𝑒𝑙𝑓
𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑒𝑑
𝟏
𝒚𝒕+𝟐 + ⏞
(𝟎)
⏟ 𝒚𝒕+𝟏 + ( ) 𝒚𝒕 = ⏟
𝟓
⏟𝟒
𝑎1 𝑐
𝑎2
𝒄
𝒚𝒑 = ( )
⏟ 𝟏 + 𝒂𝟏 + 𝒂𝟐
𝑎1 +𝑎2 ≠−1
𝟓
𝒚𝒑 = ( )
𝟏
𝟏+𝟎+𝟒
𝒚𝒑 = 𝟒
______________________________________________________________________________
𝟏
{(𝟎)𝟐 − 𝟒 ( )} ⋛ 𝟎
𝟒
−𝟏 < 𝟎
⏟
−𝟏 < 𝟎
⏟
−𝒂𝟏 ± √𝒂𝟐𝟏 − 𝟒𝒂𝟐
𝒃𝟏 , 𝒃𝟐 =
𝟐
𝟏
−(𝟎) ± √(𝟎)𝟐 − 𝟒 (𝟒)
𝒃𝟏 , 𝒃𝟐 =
𝟐
𝟎 ± √−𝟏
𝒃𝟏 , 𝒃𝟐 =
𝟐𝜐
𝒉 ⏞𝟏
⏞
𝒃𝟏 , 𝒃𝟐 = 𝟎 ± ( ) 𝒊
𝟐
𝟏
𝒉 = 𝟎 and 𝝊 =
𝟐
Using 𝒉 and 𝝊, we find 𝑹 and 𝜽.
𝝊 𝝊
𝑹 = √𝒉𝟐 + 𝝊𝟐 𝒔𝒊𝒏 𝜽 = ⇒ 𝜽 = 𝒔𝒊𝒏−𝟏 ( )
𝑹 𝑹
𝟏 𝟐 𝟏
𝑹 = √(𝟎)𝟐 + ( ) 𝜽 = 𝒔𝒊𝒏−𝟏 ( 𝟐 )
𝟐 𝟏
𝟏 𝟐
𝑹= 𝜽 = 𝒔𝒊𝒏 −𝟏 (𝟏)
𝟐
𝜽 = 𝟗𝟎°
𝑹 and 𝜽 can be substituted in the formula of Complementary Function (𝒚𝒄 ):

Complementary Function:
𝒚𝒄 = 𝑹𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝜽𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝜽𝒕 )
𝟏 𝒕 𝝅 𝝅
𝒚𝒄 = ( ) {𝑨𝟓 𝐜𝐨𝐬 ( ) 𝒕 + 𝑨𝟔 𝐬𝐢𝐧 ( ) 𝒕 }
𝟐 𝟐 𝟐
General Solution: 𝒚𝒕 = 𝒚𝒄 + 𝒚𝒑
𝟏 𝒕 𝝅 𝝅
𝒚𝒕 = ( ) {𝑨𝟓 𝐜𝐨𝐬 ( ) 𝒕 + 𝑨𝟔 𝐬𝐢𝐧 ( ) 𝒕 } + ⏟
𝟒
⏟𝟐 𝟐 𝟐 𝑦 𝑝
𝑦𝑐
𝟏
Interpretation: General Solution of Second Order Difference Equation 𝒚𝒕+𝟐 + 𝟒 𝒚𝒕 = 𝟓 with case
𝟏 𝒕 𝝅 𝝅
of Complex Roots is 𝒚𝒕 = (𝟐) {𝑨𝟓 𝐜𝐨𝐬 ( 𝟐 ) 𝒕 + 𝑨𝟔 𝐬𝐢𝐧 ( 𝟐 ) 𝒕 } + 𝟒 .
Note: Definite solution can also be found, if initial conditions 𝒚𝟎 and 𝒚𝟏 are given.
______________________________________________________________________________
Lesson 41
DYNAMIC STABILITY: DISTINCT REAL ROOTS
TOPIC 189: DYNAMIC STABILITY: DISTINCT REAL ROOTS
Recalling the under Distinct Real Roots case:

⏟𝟏 (𝒃𝟏 )𝒕 + 𝑨𝟐 (𝒃𝟐 )𝒕 + 𝒚𝒑
𝒚𝒕 = 𝑨
𝑦𝑐
𝑨𝟏 (𝒃𝟏 )𝒕 + 𝑨𝟐 (𝒃𝟐 )𝒕 +
𝒚𝒕 = ⏟ 𝒚
⏟𝒑
⏞
⏟𝟏 (𝒃𝟏 )𝒕 + 𝑨𝟐 (𝒃𝟐 )𝒕
𝑨
𝑎𝑟𝑏𝑖𝑡𝑟𝑎𝑟𝑦 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 𝑎𝑟𝑏𝑖𝑡𝑟𝑎𝑟𝑦 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡
⏞𝟏 (𝒃𝟏 ) ⏞
𝒕 ⏞𝟐 (𝒃𝟐 ) ⏞𝒕
⏟𝑨 + 𝑨
𝑨𝟏 and 𝑨𝟐 are constant and their values depend on initial conditions. 𝒕 is independent variable
spanning (𝟎 − ∞). Deviation from equilibrium will depend on roots (𝒃𝟏 , 𝒃𝟐 ).
Hair Splitting:
𝑨𝟏 𝒆 𝒓⏞𝟏 (𝒕) + 𝑨𝟐 𝒆 𝒓⏞𝟐 (𝒕)

𝒃𝟏 & 𝒃𝟐 , determine the dynamic stability.
If both 𝒃𝟏 & 𝒃𝟐 are −𝒗𝒆, then the exponential terms {𝑨𝟏 (𝒃𝟏 )𝒕 & 𝑨𝟐 (𝒃𝟐 )𝒕 } will fade away due to
exponential decay. Automatically their sum also fades away:
−𝑣𝑒 𝒕 −𝑣𝑒 𝒕
⏞𝟏 )
𝑨𝟏 × ( 𝒃 ⏞𝟐 )
+ 𝑨𝟐 × ( 𝒃 → 𝟎 when 𝒕 → ∞, ∴ 𝒚𝒄 → 𝟎.
⏟ ⏟
𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙 𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙
{ 𝑑𝑒𝑐𝑎𝑦 𝑑𝑒𝑐𝑎𝑦 }
Caveat: Both 𝒃𝟏 and 𝒃𝟐 should be negative, else convergence may not take place.
Numerical Example:
𝒚(𝒕) = 𝟑(−𝟎. 𝟖)𝒕 + (𝟎. 𝟏)𝒕
𝒚(𝒕) = 𝟑(𝟏)𝒕 + (𝟐)𝒕 + 𝟒𝒕 𝒚(𝒕) = 𝟑(𝟎. 𝟓)𝒕 + (−𝟐)𝒕 + 𝟒𝒕 + 𝟒𝒕
𝒕 𝒕 𝒕 𝒕 𝒕
𝟑 (𝟏)
⏟ + (𝟐)⏟ 𝟑 (𝟎.
⏟ 𝟓) + (−𝟐)
⏟ 𝟑 (−𝟎.
⏟ 𝟖)
|𝒃𝟏 |≮𝟏 |𝒃𝟐 |≮𝟏 |𝒃𝟏 |<𝟏 |𝒃𝟐 |≮𝟏 |𝒃𝟏 |≮𝟏
𝒕
(𝟎. 𝟏)
+⏟
𝒚𝒄 𝒚𝒄 |𝒃𝟐 |≮𝟏 𝒚𝒄
𝑨𝟏 (𝒃𝟏 )𝒕 𝑨𝟐 (𝒃𝟐 )𝒕 𝑨𝟏 (𝒃𝟏 )𝒕 𝑨𝟐 (𝒃𝟐 )𝒕 𝑨𝟏 (𝒃𝟏)𝒕 𝑨𝟐 (𝒃𝟐 )𝒕
𝒕 𝒕 𝒕 𝒕 𝒕 𝒕
𝟑 (𝟏)
⏟ (𝟐)
⏟ 𝟑 (𝟎.
⏟ 𝟓) (−𝟐)
⏟ 𝟑 ⏟ 𝟖) (𝟎.
(−𝟎. ⏟ 𝟏)
𝒕 +𝒗𝒆 +𝒗𝒆 +𝒗𝒆 −𝒗𝒆 −𝒗𝒆 +𝒗𝒆
0 3 1 4 3 1 4 3 1 4
1 3 2 5 1.5 -2 -0.5 -2.4 0.1 -2.3
______________________________________________________________________________
2 3 4 7 0.75 4 4.8 0.011.9 1.9

3 3 8 11 0.375 -8 -7.6 0.001-1.5 -1.5
4 3 16 19 0.188 16 16.2 0.0001 1.2 1.2
3 32 35 0.094 -32 -31.9 -0.98
0.0000 -0.98
5 1
3 64 67 0.046 64 64.1 0.8 0.0000 0.79
6 01
𝑨𝟐 (𝒃𝟐 )𝒕 (𝒚𝒄 ) 𝑨𝟏 (𝒃𝟏 )𝒕 |𝑨𝟐 (𝒃𝟐 )𝒕 | |𝒚𝒄 | |𝑨𝟏 (𝒃𝟏 )𝒕 | 𝑨𝟐 (𝒃𝟐 )𝒕 |𝒚𝒄 |
⇆
𝑨𝟏 (𝒃𝟏 )𝒕 ↑ ↑ ↓ ↑ ↑ ↓ ↓ ↓
st st
1 exponential Deviat 1 exponential Deviat Both Deviati
term remains ion term increases, ion exponential on
constant, 2nd one increa 2nd one increa terms decrease decrea
increases ses decreases ses ses
Divergence Divergence Convergence
Dynamically Unstable Dynamically Unstable Dynamically Stable
Hence it is numerically confirmed that for Dynamic Stability, Absolute value of BOTH the
roots must be: (|𝒃𝟏 | ∧ |𝒃𝟐 |) < 𝟏.
TOPIC 190: DYNAMIC STABILITY: REPEATED REAL ROOTS
Recalling the under Repeated Real Roots case:

𝑨𝟑 (𝒃)𝒕 + 𝑨𝟒 𝒕(𝒃)𝒕 + 𝒚𝒑
𝒚𝒕 = ⏟
𝑦𝑐
𝑨𝟑 (𝒃)𝒕 + 𝑨𝟒 𝒕(𝒃)𝒕 +
𝒚𝒕 = ⏟ 𝒚
⏟𝒑
⏞
𝑨𝟑 (𝒃)𝒕 + 𝑨𝟒 𝒕(𝒃)𝒕
⏟
𝑎𝑟𝑏𝑖𝑡𝑟𝑎𝑟𝑦 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 𝑎𝑟𝑏𝑖𝑡𝑟𝑎𝑟𝑦 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡
𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒
⏞ ⏞
𝒕 ⏞ ⏞ ⏞
𝒕
⏟ 𝑨𝟑 (𝒃) + 𝑨𝟒 𝒕 (𝒃)
𝑨𝟑 and 𝑨𝟒 are constant and their values depend on initial conditions. 𝒕 is independent variable
spanning (𝟎 − ∞). Deviation from equilibrium will depend on root (𝒃).
Hair Splitting:
𝒕 𝒕
𝑨𝟑 (𝒃)
⏟ + 𝑨𝟒 𝒕 (𝒃)
⏟
−𝒗𝒆 −𝒗𝒆
𝒃, determine the dynamic stability.
If 𝒃 is −𝒗𝒆, then the exponential terms {𝑨𝟏 (𝒃)𝒕 & 𝑨𝟐 𝒕(𝒃)𝒕 } will fade away due to exponential decay.
Though, the second term (𝑨𝟐 𝒕 (𝒃)𝒕 ) has an additional term 𝒕 being multiplied, yet it doesn’t
reverse the dynamics of this exponential term. Eventually, sum of the two exponential terms also
fades away:
(0,1,2,…∞) 𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑐𝑎𝑛𝑑
𝑨𝟑 × ⏟𝒕
(𝒃) + 𝑨𝟒 × ⏟ ⏞ 𝒕 ⏞ (𝒃)
⏟𝒕 → 𝟎 when 𝒕 → ∞, ∴ 𝒚𝒄 → 𝟎.
𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙 𝑁𝑒𝑢𝑡𝑟𝑎𝑙 𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙
{ 𝑑𝑒𝑐𝑎𝑦 𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑒𝑟 𝑑𝑒𝑐𝑎𝑦 }
______________________________________________________________________________
Numerical Example:
𝟏 𝟏
𝒚𝒕 = 𝟐(𝟑)𝒕 + 𝒕(𝟑)𝒕 + 𝒚𝒕 = 𝟐(𝟎. 𝟑)𝒕 + 𝒕(𝟎. 𝟑)𝒕 +
𝟒 𝟒
𝟐(𝟑)𝒕 + 𝒕(𝟑)𝒕 𝟐(𝟎. 𝟑)𝒕 + 𝒕(𝟎. 𝟑)𝒕
𝑨𝟑 𝒃 𝒕 𝑨𝟒 𝒕𝒃𝒕 𝒚𝒄 𝑨𝟑 𝒃 𝒕 𝑨𝟒 𝒕𝒃𝒕 𝒚𝒄
𝒕 𝒕 𝒕 𝒕
𝟐 (𝟑)
⏟ 𝒕 (𝟑)
⏟ 𝟐 (𝟎.
⏟ 𝟑) 𝒕 (𝟎.
⏟ 𝟑)
𝒕 +𝒗𝒆 +𝒗𝒆 +𝒗𝒆 +𝒗𝒆
0 2 0 2 2 0 2
1 6 3 9 0.6 0.3 0.9
2 18 18 36 0.18 0.18 0.36
3 54 81 135 0.054 0.081 0.135
4 162 324 486 0.016 0.032 0.049
5 486 1215 1701 0.005 0.012 0.017
𝒕
(𝑨𝟑 𝒃 )↑ 𝒕
(𝑨𝟒 𝒕𝒃 )↑ (𝒚𝒄 )↑ 𝒕
(𝑨𝟑 𝒃 )↓ 𝒕
(𝑨𝟒 𝒕𝒃 )↓ (𝒚𝒄 )↓
⇆
Both exponential terms Deviation Both exponential terms Deviation

increase increases decrease decreases
Divergence Convergence
Dynamically Unstable Dynamically Stable
Hence it is numerically confirmed that for Dynamic Stability, the absolute value of repeated root
(𝒃) must be less than 𝟏: |𝒃| < 𝟏
TOPIC 191: DYNAMIC STABILITY: COMPLEX ROOTS
Recalling the under Complex Roots case:

𝑹𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝜽𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝜽𝒕) + 𝒚𝒑
𝒚𝒕 = ⏟
𝑦𝑐
Hair Splitting:
⏟𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝜽𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝜽𝒕) +
𝒚𝒕 = 𝑹 𝒚
⏟𝒑
⏞
𝑹𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝜽𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝜽𝒕)
⏟
______________________________________________________________________________
𝒎𝒖𝒍𝒕𝒊𝒑𝒍𝒊𝒄𝒂𝒏𝒅
⏞𝑎𝑟𝑏𝑖𝑡𝑟𝑎𝑟𝑦 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 𝑎𝑟𝑏𝑖𝑡𝑟𝑎𝑟𝑦 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡
𝒕
𝑹
⏟ ( ⏞𝟓
𝑨 𝐜𝐨𝐬 𝜽 ⏞
𝒕 + ⏞𝟔
𝑨 𝐬𝐢𝐧 𝜽 ⏞
𝒕 )
𝒎𝒖𝒍𝒕𝒊𝒑𝒍𝒊𝒆𝒓
⏟
⏟
Stepped Fluctuation: The multiplicand (𝑨𝟓 𝐜𝐨𝐬 𝜽𝒕 + 𝑨𝟔 𝐬𝐢𝐧 𝜽𝒕) has the knack of having a wave-
like graph. But due to discrete nature of 𝒕, the smoothness of the graph is ‘stepped’ as shown
above. This shall have uniform amplitude above and below the equilibrium. So, convergence shall
not take place. However, multiplier 𝑹𝒕 can mitigate the amplitude of the graph if it’s value declines
𝑹𝒕 → 𝟎 as 𝒕 → ∞.
with time (𝒕): i.e. ⏟
𝑹<𝟏
______________________________________________________________________________
Numerical Example:
𝒚𝒕
(𝟐)𝒕 {𝐜𝐨𝐬(𝟏)𝒕 + 𝟑 𝐬𝐢𝐧(𝟏)𝒕} + 𝟒 𝒕
(𝟏) {𝐜𝐨𝐬(𝟏)𝒕 𝟏 𝒕
+ 𝟑 𝐬𝐢𝐧(𝟏)𝒕} + 𝟒 ( ) {𝐜𝐨𝐬(𝟏)𝒕 + 𝟑 𝐬𝐢𝐧(𝟏)𝒕}
𝟐
+𝟒
(𝟐)𝒕 {𝐜𝐨𝐬(𝟏)𝒕 (𝟏)𝒕 {𝐜𝐨𝐬(𝟏)𝒕 𝒕
𝒚𝒄 𝒚𝒄 𝟏
+ 𝟑 𝐬𝐢𝐧(𝟏)𝒕} + 𝟑 𝐬𝐢𝐧(𝟏)𝒕} ( ) {𝐜𝐨𝐬(𝟏)𝒕
𝟐
+ 𝟑 𝐬𝐢𝐧(𝟏)𝒕}
𝑹𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝜽𝒕 𝑹𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝜽𝒕 𝑹𝒕 (𝑨𝟓 𝐜𝐨𝐬 𝜽𝒕
𝒚𝒄
+ 𝑨𝟔 𝐬𝐢𝐧 𝜽𝒕) + 𝑨𝟔 𝐬𝐢𝐧 𝜽𝒕) + 𝑨𝟔 𝐬𝐢𝐧 𝜽𝒕)
𝒕 (𝟐)
⏟
𝒕 {𝐜𝐨𝐬(𝟏)𝒕 (𝟏)
⏟
𝒕 {𝐜𝐨𝐬(𝟏)𝒕 𝟏 𝒕 {𝐜𝐨𝐬(𝟏)𝒕
+ 𝟑 𝐬𝐢𝐧(𝟏)𝒕} + 𝟑 𝐬𝐢𝐧(𝟏)𝒕} ( ) + 𝟑 𝐬𝐢𝐧(𝟏)𝒕}
≮𝟏 ≮𝟏 ⏟𝟐
<𝟏
0 1 0 0 1 0 0 1 0 0
1 2 3.06 6.13 1 3.06 3.06 0.5 3.06 1.53
2 4 2.31 9.25 1 2.31 2.31 0.25 2.31 0.58
3 8 -0.57 -4.53 1 -0.57 -0.57 0.13 -0.57 -0.07
4 16 -2.92 -46.78 1 -2.92 -2.92 0.06 -2.92 -0.18
5 32 -2.59 -82.98 1 -2.59 -2.59 0.03 -2.59 -0.08
6 64 0.12 7.80 1 0.12 0.12 0.02 0.12 0.002
7 128 2.72 348.78 1 2.72 2.72 0.01 2.72 0.021
8 256 2.82 722.58 1 2.82 2.82 0.003 2.82 0.011
9 512 0.33 166.52 1 0.33 0.33 0.002 0.33 0.001
1 1024 -2.47 - -2.47 -2.47 0.001 -2.47 -0.002
0 2530.44 1
(𝑹𝒕 ) +𝒗𝒆 +𝒗𝒆 +𝒗𝒆 +𝒗𝒆 𝒕 +𝒗𝒆 +𝒗𝒆
( ) ( ) (𝑹𝒕 ) ( ) ( ) (𝑹 ) ( ) ( )
⇆
↑ −𝒗𝒆 −𝒗𝒆 −𝒗𝒆 −𝒗𝒆 ↓ −𝒗𝒆 −𝒗𝒆

↓↑ ↑ ↓↑ ↓↑ ↓↑ ↓
Increa con Dampe
Fluctuati Explosi Fluctuati Unifo Decre Fluctuati
sing sta d
on ve on rm asing on
nt
Explosive Fluctuation Uniform Fluctuation Damped Fluctuation
& Divergent & Neither Divergent & Convergent
nor Convergent
Dynamically Unstable Neither Dynamically Stable
Hence it is numerically confirmed that for Dynamic Stability, in Complex root happens
if; (𝑹 < 𝟏).
______________________________________________________________________________
Lesson 42
SAMUELSON MULTIPLIER-ACCELERATION INTERACTION MODEL
TOPIC 192: SAMUELSON MULTIPLIER-ACCELERATION INTERACTION MODEL:

FRAMEWORK
o Attributed to Professor Paul A. Samuelson (the first Economist to win

the Nobel Prize).
o An application of Second Order Difference Equations in Economics.

o It seeks to explore the ‘dynamic’ process of income determination when the acceleration
principle is in operation along with the Keynesian multiplier.
Among other things, that model shows that mere interaction of the multiplier and the
accelerator is capable of generating cyclical fluctuations endogenously.
𝒐𝒓
𝑹𝒆𝒄𝒐𝒗𝒆𝒓𝒚 𝑩𝒐𝒐𝒎
𝒂𝒏𝒅
⏞ ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ (↑ ∆𝒀) ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ (↑ ∆𝒀) … ⇒
(↑ ∆𝑰𝒐 ) ⏟
𝑀𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑒𝑟 𝐴𝑐𝑐𝑒𝑙𝑒𝑟𝑎𝑡𝑜𝑟 (↑ ∆𝑰𝒊 ) ⏟
⏟ 𝑀𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑒𝑟
𝑀𝑃𝐶 (𝐼𝑛𝑣𝑒𝑠𝑡𝑜𝑟𝑠′ 𝑀𝑃𝐶
𝑟𝑒𝑔𝑎𝑟𝑑𝑖𝑛𝑔
𝑐𝑜𝑛𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛)
or
o And Vice Versa i.e. Recession and depression.
𝒐𝒓
𝑹𝒆𝒄𝒆𝒔𝒔𝒊𝒐𝒏 𝑫𝒆𝒑𝒓𝒆𝒔𝒔𝒊𝒐𝒏
𝒂𝒏𝒅
⏞ ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ (↓ ∆𝒀) 𝐴𝑐𝑐𝑒𝑙𝑒𝑟𝑎𝑡𝑜𝑟
⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ (↓ ∆𝒀) … ⇒
(↓ ∆𝑰𝒐 ) ⏟
𝑀𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑒𝑟 ⏟ (↓ ∆𝑰𝒊 ) ⏟
𝑀𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑒𝑟
𝑀𝑃𝐶 (𝐼𝑛𝑣𝑒𝑠𝑡𝑜𝑟𝑠′ 𝑀𝑃𝐶
𝑟𝑒𝑔𝑎𝑟𝑑𝑖𝑛𝑔
𝑐𝑜𝑛𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛)
______________________________________________________________________________
National Income (𝒀𝒕 ) has three component expenditure streams: consumption (𝑪𝒕 ), investment
(𝑰𝒕 ), and government expenditure (𝑮𝒕 ).
Consumption (𝑪𝒕 ):
Consumption is envisaged as a function not of current income (𝒀𝒕 ) but of the income of the prior
period (𝒀𝒕−𝟏 ).
For simplicity, assume that (𝑪𝒕 ) is strictly proportional to (𝒀𝒕−𝟏 ):
𝑪𝒕 = 𝜸 𝒀𝒕−𝟏 where (𝟎 < 𝜸 < 𝟏) & is actually 𝑴𝑷𝑪.
⏟
𝑛𝑜 𝑖𝑛𝑡𝑒𝑟𝑐𝑒𝑝𝑡
Investment (𝑰𝒕 ):
(Induced) investment (𝑰𝒊 ), is a function of the prevailing trend of consumer spending (𝑪𝒕 ):
⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ (↑ ∆𝒀).
(↑ ∆𝑰𝒊 ) ⏟
𝑀𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑒𝑟
𝑀𝑃𝐶
This induced investment triggers the acceleration mechanism in the model.
Role of consumption:
Investment (𝑰𝒕 ): It is assumed that investment is induced on the basis on consumption patterns.
𝑰𝒕 = 𝜶 ⏟ (∆𝑪𝒕−𝟏 )
Consumption
increment
(∆𝑪𝒕−𝟏 )
𝑰𝒕 = 𝜶 ⏟
𝑪𝒕 − 𝑪𝒕−𝟏
𝑰𝒕 = 𝜶(𝑪𝒕 − 𝑪𝒕−𝟏 )
𝜶 > 𝟎, as the relationship between investment and consumption pattern is positive.
Economic Significance of 𝜶: It is actually the ‘Acceleration Coefficient’.

Moreover, it is considered a Fixed Ratio
𝑰𝒕 = 𝜶(𝑪𝒕 − 𝑪𝒕−𝟏 )
𝑰𝒕
=𝜶
𝑰𝒕
𝜶=
𝑰𝒕
𝜶= (𝑓𝑖𝑥𝑒𝑑)
∆𝑪𝒕
Government Expenditure (𝑮𝒕 ):
Third component is Government expenditure (𝑮𝒕 ), which is considered exogenous (𝑮𝒐 ).
Therefore, the framework of Samuelson Multiplier-Acceleration Interaction Model is composed of
following system of equations:
𝑪𝒕 = 𝜸 𝒀𝒕−𝟏
𝑰𝒕 = 𝜶(𝑪𝒕 − 𝑪𝒕−𝟏 )
𝑮𝒕 = 𝑮𝟎

SOLUTION
Borrowing the framework of Samuelson Multiplier-Acceleration Interaction Model which is

composed of following system of equations:
𝑪𝒕 = 𝜸 𝒀𝒕−𝟏
𝑰𝒕 = 𝜶(𝑪𝒕 − 𝑪𝒕−𝟏 )
______________________________________________________________________________
𝑮𝒕 = 𝑮𝟎
Combining all three components in Expenditure approach Equation:1
𝒀𝒕 = 𝑪𝒕 + 𝑰𝒕 + 𝑮𝒕
𝒀𝒕 = 𝜸 𝒀𝒕−𝟏 + 𝜶(𝑪𝒕 − 𝑪𝒕−𝟏 ) + 𝑮𝒐
Introducing a lag in 𝑪𝒕 = 𝜸 𝒀𝒕−𝟏 , we get, 𝑪𝒕 −𝟏 = 𝜸 𝒀𝒕−𝟏 −𝟏
𝑪𝒕−𝟏 = 𝜸 𝒀𝒕−𝟐
𝒀𝒕 equation becomes:
𝒀𝒕 = 𝜸 𝒀𝒕−𝟏 + 𝜶(𝜸 𝒀𝒕−𝟏 − 𝜸 𝒀𝒕−𝟐 ) + 𝑮𝒐
𝒀𝒕 = 𝜸 𝒀𝒕−𝟏 + 𝜶𝜸 𝒀𝒕−𝟏 − 𝜶𝜸 𝒀𝒕−𝟐 + 𝑮𝒐
𝒀𝒕 = 𝜸(𝟏 + 𝜶)𝒀𝒕−𝟏 − 𝜶𝜸 𝒀𝒕−𝟐 + 𝑮𝒐
𝒀𝒕 − 𝜸(𝟏 + 𝜶)𝒀𝒕−𝟏 + 𝜶𝜸 𝒀𝒕−𝟐 = 𝑮𝒐
Introducing a lead of 𝟐 on both sides:
𝒀𝒕 +𝟐 − 𝜸(𝟏 + 𝜶)𝒀𝒕−𝟏 +𝟐 + 𝜶𝜸 𝒀𝒕−𝟐 +𝟐 = 𝑮𝒐
𝒀𝒕+𝟐 − 𝜸(𝟏 + 𝜶)𝒀𝒕+𝟏 + 𝜶𝜸 𝒀𝒕 = 𝑮𝒐
comparable with second order difference equation (𝒚𝒕+𝟐 + 𝒂𝟏 𝒚𝒕+𝟏 + 𝒂𝟐 𝒚𝒕 = 𝒄):
𝒀
⏟ {−𝜸(𝟏 + 𝜶)} 𝒀
𝒕+𝟐 + ⏟ ⏟𝒕+𝟏 + 𝜶𝜸
⏟ 𝒀⏟𝒕 = 𝑮
⏟𝒐
𝑦𝑡+2 𝑎1 𝑦𝑡+1 𝑎2 𝑦𝑡 𝑐
Here; 𝒚 = 𝒀, 𝒂𝟏 = −𝜸(𝟏 + 𝜶), 𝒂𝟐 = 𝜶𝜸 & 𝒄 = 𝑮𝒐
Particular Integral (𝒀𝒑 ): 𝒚𝒑 depends on 𝒂𝟏 & 𝒂𝟐 .
𝒂𝟏 + 𝒂𝟐 = −𝜸(𝟏 + 𝜶) + 𝜶𝜸
𝒂𝟏 + 𝒂𝟐 = −𝜸 − 𝜶𝜸 + 𝜶𝜸
𝒂𝟏 + 𝒂𝟐 = −𝜸 ≠ −𝟏
Recalling the formula of 𝒚𝒑 with assumption 𝒂𝟏 + 𝒂𝟐 ≠ −𝟏:
𝒄
𝒚𝒑 =
𝟏 + 𝒂𝟏 + 𝒂𝟐
𝑮𝒐
𝒀𝒑 =
𝟏 − 𝜸(𝟏 + 𝜶) + 𝜶𝜸
𝑮𝒐
𝒀𝒑 =
𝟏 − 𝜸 − 𝜶𝜸 + 𝜶𝜸
𝒐 𝑮
𝒀𝒑 = 𝟏−𝜸
⏟
𝑌∗
Hence, we have obtained the intertemporal equilibrium of national income (𝒀).
Complementary Solution (𝒀𝒄 ): 𝒚𝒄 depends on 𝒂𝟐𝟏 ⋛ 𝟒𝒂𝟐 .

Case – I: Distinct Real Roots (𝒂𝟐𝟏 > 𝟒𝒂𝟐 )
{−𝜸(𝟏 + 𝜶)}𝟐 > 𝟒(𝜶𝜸)
𝜸𝟐 (𝟏 + 𝜶)𝟐 > 𝟒𝜶𝜸
𝜸(𝟏 + 𝜶)𝟐 > 𝟒𝜶
𝟒𝜶
𝜸>
(𝟏 + 𝜶)𝟐
Case – II: Repeated Real Roots (𝒂𝟐𝟏 = 𝟒𝒂𝟐 )
{−𝜸(𝟏 + 𝜶)}𝟐 = 𝟒(𝜶𝜸)
1 Letting 𝐼𝑡 = 0, shall lead to p-547, Example no. 2 situation.

______________________________________________________________________________
𝜸𝟐 (𝟏 + 𝜶)𝟐 = 𝟒𝜶𝜸
𝜸(𝟏 + 𝜶)𝟐 = 𝟒𝜶
𝟒𝜶
𝜸=
(𝟏 + 𝜶)𝟐
Case – III: Complex Roots (𝒂𝟐𝟏 < 𝟒𝒂𝟐 )
{−𝜸(𝟏 + 𝜶)}𝟐 < 𝟒(𝜶𝜸)
𝜸𝟐 (𝟏 + 𝜶)𝟐 < 𝟒𝜶𝜸
𝜸(𝟏 + 𝜶)𝟐 < 𝟒𝜶
𝟒𝜶
𝜸<
(𝟏 + 𝜶)𝟐
Interpretation: In terms of 𝜶 and 𝜸, we have conditions of all three cases of complementary
functions (𝒀𝒄 ). Remaining calculations can be done using their values in a Second Order
Difference Equation.

NUMERICALS
Under Samuelson Multiplier-Acceleration Interaction Model, the characteristic equation is similar

to standard form of SODE.
𝒚𝒕+𝟐 + 𝒂𝟏 𝒚𝒕+𝟏 + 𝒂𝟐 𝒚𝒕+𝟏 = 𝒄
Reduced form:
𝒚𝒕+𝟐 + 𝒂𝟏 𝒚𝒕+𝟏 + 𝒂𝟐 𝒚𝒕+𝟏 = 𝟎
Trial Solutions: (𝒚𝒕 = 𝑨𝒃 ) ⇒ (𝒚𝒕+𝟏 = 𝑨𝒃𝒕+𝟏 )&(𝒚𝒕+𝟐 = 𝑨𝒃𝒕+𝟐 ).
𝒕
𝑨𝒃𝒕+𝟐 + 𝒂𝟏 𝑨𝒃𝒕+𝟏 + 𝒂𝟐 𝑨𝒃𝒕 = 𝟎

𝑨𝒃𝒕 (𝒃𝟐 + 𝒂𝟏 𝒃𝟏 + 𝒂𝟐 ) = 𝟎
−𝒂𝟏 ± √𝒂𝟐𝟏 − 𝟒𝒂𝟐

𝟐
𝒃 + 𝒂𝟏 𝒃 + 𝒂𝟐 = 𝟎
⏟ ⇒ 𝒃𝟏 , 𝒃𝟐 =
𝑐ℎ𝑎𝑟𝑎𝑐𝑡𝑒𝑟𝑖𝑠𝑡𝑖𝑐 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛
𝟐
𝑓𝑜𝑟 𝑆𝑂𝐷𝐸
Recalling 𝒂𝟏 and 𝒂𝟐 from Samuelson Multiplier-Acceleration Interaction Model:
𝒂𝟏 = −𝜸(𝟏 + 𝜶), 𝒂𝟐 = 𝜶𝜸.
−{−𝜸(𝟏 + 𝜶)} ± √{−𝜸(𝟏 + 𝜶)}𝟐 − 𝟒(𝜶𝜸)
𝒃𝟏 , 𝒃𝟐 =
𝟐
𝜸(𝟏 + 𝜶) ± √𝜸𝟐 (𝟏 + 𝜶)𝟐 − 𝟒𝜶𝜸
𝒃𝟏 , 𝒃𝟐 =
⏟ 𝟐
𝑀𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑒𝑟−𝐴𝑐𝑐𝑒𝑙𝑒𝑟𝑎𝑡𝑜𝑟 𝑀𝑜𝑑𝑒𝑙−𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑐 𝑐ℎ𝑎𝑟𝑎𝑐𝑡𝑒𝑟𝑖𝑠𝑡𝑖𝑐 𝑟𝑜𝑜𝑡𝑠
Numerical # 1: For Samuelson Multiplier-Acceleration Interaction Model, if 𝜶 = 𝟑. 𝟓 and 𝜸 = 𝟎. 𝟖,

then find characteristic roots and judge the dynamic stability of the time-path.
𝜸(𝟏 + 𝜶) ± √𝜸𝟐 (𝟏 + 𝜶)𝟐 − 𝟒𝜶𝜸
𝒃𝟏 , 𝒃𝟐 =
⏟ 𝟐
(𝟎. 𝟖)(𝟏 + 𝟑. 𝟓) ± √(𝟎. 𝟖)𝟐 (𝟏 + 𝟑. 𝟓)𝟐 − 𝟒(𝟑. 𝟓)(𝟎. 𝟖)
𝒃𝟏 , 𝒃𝟐 =
𝟐
𝟑. 𝟔 ± √𝟏. 𝟕𝟔
𝒃𝟏 , 𝒃𝟐 =
𝟐
______________________________________________________________________________
𝟑. 𝟔 ± 𝟏. 𝟑𝟐𝟔𝟔
𝒃𝟏 , 𝒃𝟐 =
𝟐
𝟑. 𝟔 + 𝟏. 𝟑𝟐𝟔𝟔 𝟑. 𝟔 − 𝟏. 𝟑𝟐𝟔𝟔
𝒃𝟏 , 𝒃𝟐 = 𝑎𝑛𝑑
𝟐 𝟐
𝒃𝟏 = 𝟐. 𝟒𝟔 𝑎𝑛𝑑 𝒃𝟐 = 𝟏. 𝟏𝟒
Interpretation: Both |𝒃𝟏 | > 𝟏 and |𝒃𝟐 | > 𝟏, therefore, the Interaction time-path shall be divergent
& hence dynamically unstable.
Numerical # 2: For Samuelson Multiplier-Acceleration Interaction Model, if 𝜶 = 𝟎. 𝟐 and 𝜸 = 𝟎. 𝟗,

then find characteristic roots and judge the dynamic stability of the time-path.
𝜸(𝟏 + 𝜶) ± √𝜸𝟐 (𝟏 + 𝜶)𝟐 − 𝟒𝜶𝜸
𝒃𝟏 , 𝒃𝟐 =
⏟ 𝟐
(𝟎. 𝟗)(𝟏 + 𝟎. 𝟐) ± √(𝟎. 𝟗)𝟐 (𝟏 + 𝟎. 𝟐)𝟐 − 𝟒(𝟎. 𝟐)(𝟎. 𝟗)
𝒃𝟏 , 𝒃𝟐 =
𝟐
𝟏. 𝟎𝟖 ± √𝟎. 𝟒𝟒𝟔𝟒
𝒃𝟏 , 𝒃𝟐 =
𝟐
𝟏. 𝟎𝟖 ± 𝟎. 𝟔𝟔𝟖𝟏
𝒃𝟏 , 𝒃𝟐 =
𝟐
𝟏. 𝟎𝟖 + 𝟎. 𝟔𝟔𝟖𝟏 𝟏. 𝟎𝟖 − 𝟎. 𝟔𝟔𝟖𝟏
𝒃𝟏 , 𝒃𝟐 = 𝑎𝑛𝑑
𝟐 𝟐
𝒃𝟏 = 𝟎. 𝟖𝟕 𝑎𝑛𝑑 𝒃𝟐 = 𝟎. 𝟐𝟏
Interpretation: Both |𝒃𝟏 | < 𝟏 and |𝒃𝟐 | < 𝟏, therefore, the Interaction time-path shall be
convergent & hence dynamically stable.
______________________________________________________________________________
Lesson 43
SECOND ORDER DIFFERENCE EQUATION WITH VARIABLE TERM
TOPIC 195: SECOND-ORDER DIFFERENCE EQUATION WITH VARIABLE TERM
Second Order Difference Equation with constant coefficient and constant term:
𝒚𝒕+𝟐 + 𝒂
⏟𝟏 𝒚𝒕+𝟏 + 𝒂
⏟𝟐 𝒚𝒕 = ⏟
𝒄
𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑡𝑒𝑟𝑚
What if: instead of 𝒄, there is a variable term {𝒇(𝒕)} in terms of 𝒕.
𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡−𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒−𝑡𝑒𝑟𝑚 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛
⏞𝒚
𝒕+𝟐 + 𝒂
⏟𝟏 𝒚𝒕+𝟏 + 𝒂
⏟𝟐 𝒚𝒕 = 𝒇(𝒕)
⏟
𝒄𝒎𝒕 :𝑐∧𝑚≠0
⏟ 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡
(𝒚𝒑 ) 𝑠ℎ𝑎𝑙𝑙
𝑎𝑠 𝑖𝑡 𝑎𝑙𝑤𝑎𝑦𝑠 𝑑𝑒𝑝𝑒𝑛𝑑𝑠 𝑜𝑛 𝒃𝟐 +𝒂𝟏 𝒃+𝒂𝟐 =𝟎 .
Coefficients.
successive differences together contain only distinct types of expressions:
Function & its Differences Distinct Type of Expressions
𝒚𝒕 𝟕𝒕
∆𝒚𝒕 ∆𝟕 = 𝟕𝒕+𝟏 − 𝟕𝒕
𝒕
= (𝟕 − 𝟏)𝟕𝒕
∆𝟕𝒕 = 𝟔(𝟕𝒕 )
∆𝟐 𝒚𝒕 ∆𝟐 (𝟕𝒕 ) = ∆(∆𝟕𝒕 )
∆𝟐 (𝟕𝒕 ) = ∆{𝟔(𝟕𝒕 )}
∆𝟐 (𝟕𝒕 ) = 𝟔{∆(𝟕𝒕 )}
∆𝟐 (𝟕𝒕 ) = 𝟔(𝟕𝒕+𝟏 − 𝟕𝒕 )
∆𝟐 (𝟕𝒕 ) = 𝟔(𝟕 − 𝟏)𝟕𝒕
∆𝟐 (𝟕𝒕 ) = 𝟑𝟔(𝟕𝒕 )

𝒚𝒕+𝟐 + 𝒚𝒕+𝟏 − 𝟑𝒚𝒕 = 𝟕𝒕
𝒚𝒕+𝟐 + (𝟏)
⏟ 𝒚𝒕+𝟏 + (−𝟑)
⏟ 𝒚𝒕 = 𝟕⏟𝒕
𝑎1 𝑎2 𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙
𝑡𝑒𝑟𝑚
𝒇(𝒕)
Here, 𝒂𝟏 = 𝟏, 𝒂𝟐 = −𝟑 and 𝒇(𝒕) = 𝟕𝒕.
Here, particular integral (𝒚𝒑 ) is calculated differently due to change in 𝑳. 𝑯. 𝑺.
𝑚𝑢𝑠𝑡
𝐿.𝐻.𝑆 𝑏𝑒 𝑒𝑞𝑢𝑎𝑙 𝑅.𝐻.𝑆
⏞𝒚𝒕+𝟐 + 𝒚𝒕+𝟏 − 𝟑 𝒚𝒕 =
⏞ ⏞ 𝟕⏟𝒕
⏟
𝒚𝒕+𝟐 ,𝒚𝒕+𝟏 & 𝒚𝒕 𝑖𝑛 𝑡𝑒𝑟𝑚𝑠 𝑜𝑓 (𝒕)
𝑡𝑒𝑟𝑚 𝒇(𝒕)
Both sides are in terms of (𝒕), however, the polynomial on 𝑳. 𝑯. 𝑺 is not explicit i.e. functions;
𝒚𝒕+𝟐 , 𝒚𝒕+𝟏 & 𝒚𝒕 .
______________________________________________________________________________
Equality (=) requires that if 𝑹. 𝑯. 𝑺 is exponential function in terms of (𝒕), then 𝑳. 𝑯. 𝑺 should also
be exponential in terms of (𝒕). So, we assume a (trial) general exponential function that could
satisfy 𝒚𝒕 , 𝒚𝒕+𝟏 & 𝒚𝒕+𝟐 on 𝑳. 𝑯. 𝑺 to satisfy equality.
As found in table above:
𝒚𝒕 = 𝟕𝒕
∆𝒚𝒕 = 𝟔(𝟕𝒕 )
∆𝟐 𝒚𝒕 = 𝟑𝟔(𝟕𝒕 )
A constant grows with 𝟕𝒕
Observed pattern: 𝑩(𝟕𝒕 )
𝒕)
Generally speaking, 𝒚𝒕 = 𝑩(𝟕 can serve as exponential type of solution, which produces
distinct type of expressions. Therefore, we use it as a trial (general) solution of the given
equation.
𝒚𝒕 = 𝑩(𝟕𝒕 )
Generating 𝒚𝒕+𝟏 and 𝒚𝒕+𝟐 for the given equation ( 𝒚𝒕+𝟐 + 𝒚𝒕+𝟏 − 𝟑 𝒚𝒕 = 𝟕𝒕 ).
𝒚𝒕+𝟏 = 𝑩(𝟕𝒕+𝟏 )
𝒚𝒕+𝟐 = 𝑩(𝟕𝒕+𝟐 )
Substituting, 𝒚𝒕 , 𝒚𝒕+𝟏 and 𝒚𝒕+𝟐 in the given equation ( 𝒚𝒕+𝟐 + 𝒚𝒕+𝟏 − 𝟑 𝒚𝒕 = 𝟕𝒕 ).
𝑩(𝟕𝒕+𝟐 ) + 𝑩(𝟕𝒕+𝟏 ) − 𝟑 𝑩(𝟕𝒕 ) = 𝟕𝒕
𝑩(𝟕𝒕+𝟐 ) + 𝑩(𝟕𝒕+𝟏 ) − 𝟑𝑩(𝟕𝒕 ) = 𝟕𝒕
𝑩{(𝟕𝟐 ) + (𝟕) − 𝟑}(𝟕𝒕 ) = 𝟕𝒕
𝑩{𝟓𝟖}(𝟕𝒕 ) = 𝟕𝒕
𝟓𝟖𝑩(𝟕𝒕 ) = 𝟕𝒕
𝟓𝟖𝑩 𝟕𝒕 = {𝟏} 𝟕𝒕
⏟
𝑐𝑜𝑚𝑝𝑎𝑟𝑖𝑛𝑔 𝑡ℎ𝑒 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡𝑠 𝑜𝑓 𝟕𝒕 𝑜𝑛 𝑏𝑜𝑡ℎ 𝑠𝑖𝑑𝑒𝑠
𝟓𝟖𝑩 = {𝟏}
𝟓𝟖𝑩 = 𝟏
𝟏
𝑩=
𝟓𝟖
The trial (general) solution was: 𝒚𝒕 = 𝑩(𝟕𝒕 ), so substitute value of 𝑩:
𝟏
𝒚𝒕 = ( ) (𝟕𝒕 )
𝟓𝟖
𝟏
Interpretation: 𝒚𝒕 = (𝟓𝟖) (𝟕𝒕 ) is the particular integral (𝒚𝒑 ) of given equation
𝒚𝒕+𝟐 + 𝒚𝒕+𝟏 − 𝟑𝒚𝒕 = 𝟕𝒕 which is found by deploying the Method of Undetermined Coefficients.
Whereas, (𝒚𝒄 ) can be found using methods same as ‘constant-coefficient and constant-term
second order difference equation’.
TOPIC 196: SECOND-ORDER DIFFERENCE EQUATION WITH VARIABLE TERM (CTN)
We solved the Second Order Difference Equation with constant coefficient and variable term
(𝒄𝒎𝒕 ):
What if: instead of (𝒄𝒎𝒕 ), there is a variable term (𝒄𝒕𝒏 ).
______________________________________________________________________________
𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡−𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒−𝑡𝑒𝑟𝑚 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛

⏞𝒚𝒕+𝟐 + 𝒂
⏟𝟏 𝒚𝒕+𝟏 + 𝒂
⏟𝟐 𝒚𝒕 = 𝒇(𝒕)
⏟
𝒄𝒕𝒏 :𝑐∧𝑛≠0
⏟ 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 (𝒚𝒑 ) 𝑠ℎ𝑎𝑙𝑙
𝑎𝑠 𝑖𝑡 𝑎𝑙𝑤𝑎𝑦𝑠 𝑑𝑒𝑝𝑒𝑛𝑑𝑠 𝑜𝑛 𝒃𝟐 +𝒂𝟏 𝒃+𝒂𝟐 =𝟎 .
Comparison of (𝒄𝒎𝒕 ) and (𝒄𝒕𝒏 ): (𝒄) is a common coefficient in both, however:

(𝒎𝒕 ) (𝒕𝒏 )
Exponential Function Power Function
Coefficients.
successive differences together contain only distinct types of expressions:
Function & its Differences Distinct Type of Expressions
𝒚𝒕 𝒕𝟐
∆𝒚𝒕 ∆𝒕𝟐 = (𝒕 + 𝟏)𝟐 − 𝒕𝟐
∆𝒕𝟐 = 𝒕𝟐 + 𝟐𝒕 + 𝟏 − 𝒕𝟐
∆𝒕𝟐 = 𝟐𝒕 + 𝟏
∆𝟐 𝒚𝒕 ∆𝟐 (𝒕𝟐 ) = ∆(∆𝒕𝟐 )
∆𝟐 (𝒕𝟐 ) = ∆(𝟐𝒕 + 𝟏)
= {(𝟐(𝒕 + 𝟏) + 𝟏) − (𝟐𝒕 + 𝟏)}
= {𝟐𝒕 + 𝟐 + 𝟏 − 𝟐𝒕 − 𝟏}
∆𝟐 (𝒕𝟐 ) = 𝟐
Numerical Example: Find the Particular Integral of: 𝒚𝒕+𝟐 + 𝟓𝒚𝒕+𝟏 + 𝟐𝒚𝒕 = 𝒕𝟐
𝒚𝒕+𝟐 + ⏟𝟓 𝒚𝒕+𝟏 + ⏟𝟐 𝒚𝒕 = 𝒕⏟𝟐
𝒇(𝒕)=𝒄𝒕𝒏
Here, 𝒂𝟏 = 𝟓, 𝒂𝟐 = 𝟐 and 𝒇(𝒕) = 𝒕𝟐 .
Particular integral (𝒚𝒑 ) is calculated differently due to change in 𝑳. 𝑯. 𝑺.
𝑚𝑢𝑠𝑡
𝐿.𝐻.𝑆 𝑏𝑒 𝑒𝑞𝑢𝑎𝑙 𝑅.𝐻.𝑆
⏞
𝒚𝒕+𝟐 + 𝟓 𝒚𝒕+𝟏 + 𝟐 𝒚𝒕 =
⏞ ⏞ 𝒕⏟𝟐
⏟
𝒚𝒕 ,𝒚𝒕+𝟏 & 𝒚𝒕+𝟐 𝑖𝑛 𝑡𝑒𝑟𝑚𝑠 𝑜𝑓 (𝒕)
𝑖𝑛 𝑡𝑒𝑟𝑚𝑠 𝑜𝑓 (𝒕)
in terms of (𝒕). So, we assume a (trial) general quadratic function that could satisfy 𝒚𝒕 , 𝒚𝒕+𝟏 & 𝒚𝒕+𝟐
on 𝑳. 𝑯. 𝑺 to satisfy equality:
𝒚𝒕 = 𝑩𝟏 𝒕𝟐 + 𝑩𝟐 𝒕 + 𝑩𝟑 .
Developing 𝒚𝒕+𝟏 and 𝒚𝒕+𝟐 for the given ( 𝒚𝒕+𝟐 + 𝟓 𝒚𝒕+𝟏 + 𝟐 𝒚𝒕 = 𝒕𝟐 ).
𝒚𝒕+𝟏 = 𝑩𝟏 (𝒕 + 𝟏)𝟐 + 𝑩𝟐 (𝒕 + 𝟏) + 𝑩𝟑
𝒚𝒕+𝟏 = 𝑩𝟏 (𝒕𝟐 + 𝟐𝒕 + 𝟏) + 𝑩𝟐 𝒕 + 𝑩𝟐 + 𝑩𝟑
𝒚𝒕+𝟏 = 𝑩𝟏 𝒕𝟐 + 𝟐𝑩𝟏 𝒕 + 𝑩𝟐 𝒕 + 𝑩𝟏 + 𝑩𝟐 + 𝑩𝟑
𝒚𝒕+𝟏 = 𝑩𝟏 𝒕𝟐 + (𝟐𝑩𝟏 + 𝑩𝟐 )𝒕 + 𝑩𝟏 + 𝑩𝟐 + 𝑩𝟑
𝒚𝒕+𝟐 = 𝑩𝟏 (𝒕 + 𝟐)𝟐 + 𝑩𝟐 (𝒕 + 𝟐) + 𝑩𝟑
______________________________________________________________________________
𝒚𝒕+𝟐 = 𝑩𝟏 (𝒕𝟐 + 𝟒𝒕 + 𝟒) + 𝑩𝟐 𝒕 + 𝟐𝑩𝟐 + 𝑩𝟑

𝒚𝒕+𝟐 = 𝑩𝟏 𝒕𝟐 + 𝟒𝑩𝟏 𝒕 + 𝟒𝑩𝟏 + 𝑩𝟐 𝒕 + 𝟐𝑩𝟐 + 𝑩𝟑
𝒚𝒕+𝟐 = 𝑩𝟏 𝒕𝟐 + 𝟒𝑩𝟏 𝒕 + 𝑩𝟐 𝒕 + 𝟒𝑩𝟏 + 𝟐𝑩𝟐 + 𝑩𝟑
𝒚𝒕+𝟐 = 𝑩𝟏 𝒕𝟐 + (𝟒𝑩𝟏 + 𝑩𝟐 )𝒕 + 𝟒𝑩𝟏 + 𝟐𝑩𝟐 + 𝑩𝟑
Since 𝒚𝒕 , 𝒚𝒕+𝟏 and 𝒚𝒕+𝟐 are now symbolized, we substitute them on 𝑳. 𝑯. 𝑺:
𝒚𝒕+𝟐 + 𝟓 𝒚𝒕+𝟏 + 𝟐 𝒚𝒕 = 𝒕⏞𝟐
⏞
𝑩𝟏 𝒕𝟐 + (𝟒𝑩𝟏 + 𝑩𝟐 )𝒕 + 𝟒𝑩𝟏 + 𝟐𝑩𝟐 + 𝑩𝟑 + 𝟓 𝑩𝟏 𝒕𝟐 + (𝟐𝑩𝟏 + 𝑩𝟐 )𝒕 + 𝑩𝟏 + 𝑩𝟐 + 𝑩𝟑
+ 𝟐 𝑩𝟏 𝒕𝟐 + 𝑩𝟐 𝒕 + 𝑩𝟑 = 𝒕𝟐
𝑩𝟏 𝒕𝟐 + (𝟒𝑩𝟏 + 𝑩𝟐 )𝒕 + 𝟒𝑩𝟏 + 𝟐𝑩𝟐 + 𝑩𝟑 + 𝟓𝑩𝟏 𝒕𝟐 + 𝟓(𝟐𝑩𝟏 + 𝑩𝟐 )𝒕 + 𝟓𝑩𝟏 + 𝟓𝑩𝟐 + 𝟓𝑩𝟑
+ 𝟐𝑩𝟏 𝒕𝟐 + 𝟐𝑩𝟐 𝒕 + 𝟐𝑩𝟑 = 𝒕𝟐
𝑩𝟏 𝒕𝟐 + 𝟓𝑩𝟏 𝒕𝟐 + 𝟐𝑩𝟏 𝒕𝟐 + (𝟒𝑩𝟏 + 𝑩𝟐 )𝒕 + 𝟓(𝟐𝑩𝟏 + 𝑩𝟐 )𝒕 + 𝟐𝑩𝟐 𝒕 + 𝟒𝑩𝟏 + 𝟐𝑩𝟐 + 𝑩𝟑

+ 𝟓𝑩𝟏 + 𝟓𝑩𝟐 + 𝟓𝑩𝟑 + 𝟐𝑩𝟑 = 𝒕𝟐
𝑩𝟏 + 𝟓𝑩𝟏 + 𝟐𝑩𝟏 𝒕𝟐 + {𝟒𝑩𝟏 + 𝑩𝟐 + 𝟏𝟎𝑩𝟏 + 𝟓𝑩𝟐 + 𝟐𝑩𝟐 }𝒕 + 𝟒𝑩𝟏 + 𝟓𝑩𝟏 + 𝟐𝑩𝟐 + 𝟓𝑩𝟐 + 𝑩𝟑
+ 𝟓𝑩𝟑 + 𝟐𝑩𝟑 = 𝒕𝟐
𝑳.𝑯.𝑺 𝑹.𝑯.𝑺
⏞ ⏞
𝟖𝑩𝟏 𝒕𝟐 + (𝟏𝟒𝑩𝟏 + 𝟖𝑩𝟐 )𝒕 + 𝟗𝑩𝟏 + 𝟕𝑩𝟐 + 𝟖𝑩𝟑 = 𝟏 𝒕𝟐 + (𝟎)𝒕
⏟ + (𝟎)
𝑠𝑒𝑙𝑓−𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑒𝑑
𝑓𝑜𝑟 𝑐𝑜𝑚𝑝𝑎𝑟𝑖𝑠𝑜𝑛
Comparing the coefficients of 𝒕𝟐 & 𝒕 and the 𝐜𝐨𝐧𝐬𝐭𝐚𝐧𝐭𝐬 on both sides of equation:
𝒕𝟐 𝒕 𝐜𝐨𝐧𝐬𝐭𝐚𝐧𝐭𝐬
𝟖𝑩𝟏 = 𝟏 (𝟏𝟒 𝑩𝟏 + 𝟖𝑩𝟐 ) = (𝟎) 𝟗 𝑩 + 𝟕𝑩 + 𝟖𝑩 = (𝟎)
𝟏 𝟐 𝟑
𝟏 𝟕
𝑩𝟏 = + 𝟖𝑩𝟐 = 𝟎 𝟏
𝟖 𝟒 𝟗 + 𝟕𝑩𝟐 + 𝟖𝑩𝟑 = 𝟎
𝟖
𝟕 𝟗 𝟕
𝑩𝟐 = − + 𝟕 (− ) + 𝟖𝑩𝟑 = 𝟎
𝟑𝟐 𝟖 𝟑𝟐
𝟏𝟑
𝑩𝟑 =
𝟐𝟓𝟔
So, the Undetermined coefficients are now determined.
𝟏 𝟕 𝟏𝟑
𝒚⏟𝒕 = ( ) 𝒕𝟐 + (− ) 𝒕 + ( )
𝟖 𝟑𝟐 𝟐𝟓𝟔
𝑦𝑝
𝟏 𝟐 𝟕 𝟏𝟑
𝒚𝒑 = 𝒕 − 𝒕+
𝟖 𝟑𝟐 𝟐𝟓𝟔
Interpretation: So, the particular integral of the ‘constant-coefficient and variable-term second
order difference equation’ 𝒚𝒕+𝟐 + 𝟓𝒚𝒕+𝟏 + 𝟐𝒚𝒕 = 𝒕𝟐 is
𝟏 𝟕 𝟏𝟑
𝒚𝒑 = 𝟖 𝒕𝟐 − 𝟑𝟐 𝒕 + 𝟐𝟓𝟔 . Method used to find (𝒚𝒑 ) is Method of Undetermined Coefficients.
Whereas, (𝒚𝒄 ) can be found using methods same as ‘constant-coefficient and constant-term
second order difference equation’.
TOPIC 197: HIGHER-ORDER LINEAR DIFFERENCE EQUATION
______________________________________________________________________________
After comprehending the solution of second order difference equations, we extend our analysis
to Higher (𝒏 > 𝟐) Order Difference Equations.
Standard Form of Higher Order Difference Equation: A linear, non-homogeneous version with
‘𝒏’ constant coefficients (𝒂𝟏 , 𝒂𝟐 , … , 𝒂𝒏 ) and constant term 𝒄.
(=1)
⏞𝟎 𝒚𝒕+𝒏 + 𝒂𝟏 𝒚𝒕+(𝒏−𝟏) + 𝒂𝟐 𝒚𝒕+(𝒏−𝟐) + ⋯ + 𝒂𝒏−𝟏 𝒚𝒕+𝟏 + 𝒂𝒏 𝒚𝒕 = 𝒄
𝒂
←
𝑜𝑟𝑑𝑒𝑟 𝑜𝑓 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑖𝑛𝑐𝑟𝑒𝑎𝑠𝑒𝑠 {𝒚𝒕 ,𝒚𝒕+𝟏 𝑡𝑜 𝒚𝒕+𝒏 }
⏞
𝒂
⏟𝟎 𝒚𝒕+𝒏 + 𝒂𝟏 𝒚𝒕+(𝒏−𝟏) + 𝒂𝟐 𝒚𝒕+(𝒏−𝟐) + ⋯ + 𝒂𝒏−𝟏 𝒚𝒕+𝟏 + 𝒂𝒏 𝒚𝒕 = 𝒄
𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡−𝑠𝑢𝑏𝑠𝑐𝑟𝑖𝑝𝑡 𝑖𝑛𝑐𝑟𝑒𝑎𝑠𝑒𝑠 (𝒂𝟎 𝑡𝑜 𝒂𝒏 )
→
𝒕𝒉 𝒕𝒉
Order: 𝒏 (due to 𝒏 difference)
Degree: Linear/First degree (due to highest power ‘𝟏’ of the 𝒏𝒕𝒉 difference) Constant Coefficients
(𝒂′ 𝑠) and Constant term (𝒄).
Delimitations:
3. Cross terms – product of function ′𝒚𝒕 ′ and its differences (leads/lags) 𝒚𝒕+𝟏 ,… ,𝒚𝒕+𝒏 – are
not included.
Solution: Solution of Higher Order Difference equations is merely the extension of solution of
Second Order Difference equations.
Guidelines:
1. Particular Solution has same buildup of cases. e.g.:
Case – I has same condition; (𝒂𝟏 + 𝒂𝟐 + ⋯ + 𝒂𝒏 ≠ −𝟏)
2. 𝒏𝒕𝒉 order implies 𝒏 number of characteristic roots.
3. A hybrid complementary solution can exist, e.g.: out of 𝒏 roots, if first 𝟑 are repeated real
and last two are complex roots, then
𝒚𝒕 = ⏟𝑨𝟏 𝒃𝒕𝟏 + 𝑨𝟏 𝒕𝒃𝒕𝟏 + 𝑨𝟑 𝒕𝟐 𝒃𝒕𝟏 +, … , + ⏟
𝑹𝒕 (𝑨𝒏−𝟏 𝒄𝒐𝒔 𝜽𝒕 + 𝑨𝒏 𝒔𝒊𝒏 𝜽𝒕)
1𝑠𝑡 𝟑 𝑎𝑟𝑒 𝑟𝑒𝑝𝑒𝑎𝑡𝑒𝑑 𝑟𝑒𝑎𝑙 𝑟𝑜𝑜𝑡𝑠 𝑙𝑎𝑠𝑡 𝟐 𝑎𝑟𝑒 𝑐𝑜𝑚𝑝𝑙𝑒𝑥 𝑟𝑜𝑜𝑡𝑠
𝑵𝒖𝒎𝒃𝒆𝒓 𝒐𝒇:
4. ⏞
Characteristic Roots (𝒃𝒏 ) = arbitrary constants (𝑨𝒏 ) = ⏟
𝒏 initial conditions
𝑓𝑜𝑟 𝑑𝑒𝑓𝑖𝑛𝑖𝑡𝑒 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛
Numerical Example: Find the General Solution of third-order difference equation:

𝟕 𝟏 𝟏
𝒚𝒕+𝟑 − 𝒚𝒕+𝟐 + 𝒚𝒕+𝟏 + 𝒚 =𝟗
𝟖 𝟖 𝟑𝟐 𝒕
𝟕 𝟏 𝟏
𝒚𝒕+𝟑 + (− ) 𝒚𝒕+𝟐 + ( ) 𝒚𝒕+𝟏 + ( ) 𝒚𝒕 = ⏟ 𝟗
⏟ 𝟖 ⏟𝟖 ⏟𝟑𝟐 𝑐
𝑎1 𝑎2 𝑎3
𝟕 𝟏 𝟏
Here; 𝒂𝟏 = − 𝟖, 𝒂𝟐 = 𝟖
, 𝒂𝟑 = 𝒂𝒏 = 𝟑𝟐
and 𝒄=𝟗≠𝟎
⏟ .
𝑐𝑎𝑠𝑒
Particular Integral (𝒚𝒑 ): Testing for Case – I:
𝒂𝟏 + 𝒂𝟐 + 𝒂𝟑 ≠ −𝟏
𝟕 𝟏 𝟏 𝟐𝟑
− + + =− ≠ −𝟏
𝟖 𝟖 𝟑𝟐 𝟑𝟐
Recalling the formula of 𝒚𝒑 with assumption 𝒂𝟏 + 𝒂𝟐 + 𝒂𝟑 ≠ −𝟏:
𝒄
𝒚𝒑 =
⏟ 𝟏 + 𝒂 𝟏 + 𝒂𝟐 + 𝒂𝟑
𝑠𝑖𝑚𝑖𝑙𝑎𝑟 𝑓𝑜𝑟𝑚𝑢𝑙𝑎;𝑒𝑥𝑡𝑒𝑛𝑡𝑒𝑑 𝑡𝑜 𝒏=𝟑
______________________________________________________________________________
𝟗
𝒚𝒑 = [ ]
𝟕 𝟏 𝟏
{𝟏 + (− 𝟖) + 𝟖 + 𝟑𝟐}
𝒚𝒑 = 𝒚∗ = 𝟑𝟐
⏟
𝑠𝑡𝑎𝑡𝑖𝑐 𝑒𝑞𝑢𝑖𝑙𝑖𝑏𝑟𝑖𝑢𝑚
Complementary Function (𝒚𝒄 ): For Second Order Difference equation, the characteristic
equation was:
𝒃𝟐 + 𝒂𝟏 𝒃 + 𝒂𝟐 = 𝟎
Accordingly, for Third Order Difference equation, the characteristic equation will be:
𝒃𝟑 + 𝒂𝟏 𝒃𝟐 + 𝒂𝟐 𝒃 + 𝒂𝟑 = 𝟎
To find the roots (𝒃), we need to solve the characteristic equation, therefore, factorizing
𝟕 𝟏 𝟏
𝒃𝟑 − 𝒃𝟐 + 𝒃 + = 𝟎, we get:
𝟖 𝟖 𝟑𝟐
𝟏 𝟏 𝟏
(𝒃 − ) × (𝒃 − ) × (𝒃 + ) = 𝟎
𝟐 𝟐 𝟖
Either;
𝟏 𝟏 𝟏
(𝒃𝟏 − ) = 𝟎 (𝒃𝟏 − ) = 𝟎 (𝒃𝟑 + ) = 𝟎
𝟐 𝟐 𝟖
⇓ ⇓ ⇓
𝟏 𝟏 𝟏
𝒃𝟏 = 𝒃𝟐 = 𝒃𝟑 = −
𝟐 𝟐 𝟖
Summarizing the (hybrid) group of roots:
𝟏
𝒃=
𝟐
⏞ 𝟏 𝟏 𝟏
𝒃𝟏 = , 𝒃𝟐 = , 𝒃 𝟑 = −
⏟ 𝟐 𝟐 ⏟ 𝟖
𝑟𝑒𝑝𝑒𝑎𝑡𝑒𝑑 𝑟𝑒𝑎𝑙 𝑟𝑜𝑜𝑡𝑠 𝑑𝑖𝑠𝑡𝑖𝑛𝑐𝑡 𝑟𝑒𝑎𝑙
𝑟𝑜𝑜𝑡
Accordingly, the complementary function will also be hybrid:
𝒚𝒄 = ⏞ ⏟
𝑨𝟏 (𝒃𝟏 )𝒕 + 𝑨𝟐 𝒕(𝒃𝟐 )𝒕 + 𝑨𝟑 (𝒃𝟑 )𝒕
⏟
𝑅𝑒𝑝𝑒𝑎𝑡𝑒𝑑 𝑅𝑒𝑎𝑙 𝑟𝑜𝑜𝑡𝑠 𝑠𝑒𝑔𝑚𝑒𝑛𝑡 𝐷𝑖𝑠𝑡𝑖𝑛𝑐𝑡 𝑟𝑒𝑎𝑙 𝑟𝑜𝑜𝑡 𝑠𝑒𝑔𝑚𝑒𝑛𝑡
𝑤ℎ𝑒𝑟𝑒 𝒃𝟐 =𝒃𝟐 =𝒃
⏞ 𝟏 𝒕 𝟏 𝒕 𝟏 𝒕
𝒚𝒄 = 𝑨𝟏 ( ) + 𝑨𝟐 𝒕 ( ) + 𝑨𝟑 (− )
⏟ 𝟐 𝟐 ⏟ 𝟖
𝑅𝑒𝑝𝑒𝑎𝑡𝑒𝑑 𝑅𝑒𝑎𝑙 𝑟𝑜𝑜𝑡𝑠 𝑠𝑒𝑔𝑚𝑒𝑛𝑡 𝐷𝑖𝑠𝑡𝑖𝑛𝑐𝑡 𝑟𝑒𝑎𝑙 𝑟𝑜𝑜𝑡 𝑠𝑒𝑔𝑚𝑒𝑛𝑡
𝑤ℎ𝑒𝑟𝑒 𝒃𝟐 =𝒃𝟐 =𝒃
General Solution (𝒚𝒕 ): Like any other difference equation:

Substituting the values of 𝒚𝒄 , 𝒚𝒑 we get:
𝟏 𝒕 𝟏 𝒕 𝟏 𝒕
𝒚𝒕 = 𝑨𝟏 ( ) + 𝑨𝟐 𝒕 ( ) + 𝑨𝟑 (− ) + 𝟑𝟐
⏟
⏟ 𝟐 𝟐 𝟖 𝑦 𝑝
𝑦𝑐
______________________________________________________________________________
𝟏 𝒕 𝟏 𝒕 𝟏 𝒕
Interpretation: 𝒚𝒕 = 𝑨𝟏 (𝟐) + 𝑨𝟐 𝒕 (𝟐) + 𝑨𝟑 (− 𝟖) + 𝟑𝟐 is the general solution of Higher (Third)
𝟕 𝟏 𝟏
Order Difference Equation 𝒚𝒕+𝟑 − 𝒚𝒕+𝟐 + 𝒚𝒕+𝟏 + 𝒚 = 𝟗 . Complementary Function is hybrid
𝟖 𝟖 𝟑𝟐 𝒕
with a combination of Repeated Real roots and Distinct Real root.
Note: If 𝒏(= 𝟑) number of initial conditions are given then, we can also find the definite solution
from general solution which has 𝒏(= 𝟑) number arbitrary constants 𝑨𝒏 .
Dynamic Stability: Similar to any other Difference Equation, the roots should be less than 𝟏 in
terms of their magnitude. Here,
𝟏 𝟏
{(|𝒃𝟏 | = |𝒃𝟐 | = | |) ∧ (|𝒃𝟑 | = |− |)} < 𝟏
⏟ 𝟐 𝟖
𝐶𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡 𝑡𝑖𝑚𝑒−𝑝𝑎𝑡ℎ 𝑜𝑓 𝑡ℎ𝑒 3𝑟𝑑 𝑂𝑟𝑑𝑒𝑟 𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛
TOPIC 198: CONVERGENCE AND THE SCHUR THEOREM
After comprehending the solution of Higher order difference equations, we realize that it can be a
cumbersome quantitative exercise to find 𝒏 number of roots.
(=1)
⏞𝟎 𝒚𝒕+𝒏 + 𝒂𝟏 𝒚𝒕+(𝒏−𝟏) + 𝒂𝟐 𝒚𝒕+(𝒏−𝟐) + ⋯ + 𝒂𝒏−𝟏 𝒚𝒕+𝟏 + 𝒂𝒏 𝒚𝒕 = 𝒄
𝒂
𝒕𝒉
𝒏 order implies 𝒏 number of characteristic roots.
An easier qualitative method is available – Schur Theorem.
Without having to perform quantitative solution of higher order difference equations, one can find
the dynamic stability of the time path.
Determinants are developed and evaluated using the coefficients in given difference equations.
Number of determinants = Order of Equation
For second-order difference equation (SODE) to be dynamically stable:
𝒂𝟎 𝒂𝟐
∆𝟏 = |𝒂 𝒂 | ⋚ 𝟎
𝟐 𝟎
𝒂𝟎 𝟎 𝒂𝟐 𝒂𝟏
𝒂 𝒂𝟎 𝟎 𝒂𝟐
∆𝟐 = | 𝟏 |⋚𝟎
𝒂𝟐 𝟎 𝒂𝟎 𝒂𝟏
𝒂𝟏 𝒂𝟐 𝟎 𝒂𝟎
Interpretation
(∆𝟏 < 𝟎) ∧ (∆𝟐 > 𝟎)
(∆𝟏 ∧ ∆𝟐 ) < 𝟎 { ⋁ } (∆𝟏 ∧ ∆𝟐 ) > 𝟎
(∆𝟏 > 𝟎) ∧ (∆𝟐 < 𝟎)
Divergent Neither Convergent
Numerical Example # 1: Find the Dynamic stability of the higher (second) order difference
equation without solving it quantitatively:
𝒚𝒕+𝟐 + 𝟑𝒚𝒕+𝟏 + 𝟐𝒚𝒕 = 𝟏𝟐
𝟏 𝒚𝒕+𝟐 + (𝟑)
⏟ ⏟ 𝒚𝒕+𝟏 + (𝟐)
⏟ 𝒚𝒕 = 𝟏𝟐
⏟
𝑎0 𝑎1 𝑎2 𝑐
Here; 𝒂𝟎 = 𝟏, 𝒂𝟏 = 𝟑, 𝒂𝟐 = 𝟐, and 𝒄 = 𝟏𝟐 ≠ 𝟎 .
⏟
𝑐𝑎𝑠𝑒
Recalling two Schur Theorem determinants for second-order difference equation:
𝒂𝟎 𝒂𝟐 However, we evaluate the second
∆𝟏 = |𝒂 𝒂 | ⋚ 𝟎
𝟐 𝟎 determinant (∆𝟐 ) for the sake of
rehearsal.
______________________________________________________________________________
𝟏 𝟐 𝟏 𝟎 𝟐 𝟑
∆𝟏 = | | = −𝟑
𝟐 𝟏 𝟑 𝟏 𝟎 𝟐
∆𝟏 < 𝟎 ∆𝟐 = | |=𝟎
𝟐 𝟎 𝟏 𝟑
First Determinant (∆𝟏 ) is already found to be negative. It 𝟑 𝟐 𝟎 𝟏
violates the requirement Convergence in Schur ∆𝟐 = 𝟎
Theorem which requires all of the determinants to be Second determinant (∆𝟐 ) also defies
positive. the requirement of Convergence in
Schur Theorem.
Interpretation
(∆𝟏 < 𝟎) ∧ (∆𝟐 > 𝟎)
(∆𝟏 ∧ ∆𝟐 ) < 𝟎 { ⋁ } (∆𝟏 ∧ ∆𝟐 ) ≯ 𝟎
(∆𝟏 > 𝟎) ∧ (∆𝟐 < 𝟎)
DIVERGENT Not applicable Not applicable
Verification: It can be done by finding the roots via (quantitative) solution and comparing their
absolute values with < 𝟏. The roots of the given Higher Order Difference equation are −𝟏, −𝟐,
absolute value of both of which are not lesser than 𝟏. Hence it is verified that Schur Theorem
gave us correct interpretation.
𝟏 𝟏
Numerical Example # 2: Test the convergence of the path of 𝒚𝒕+𝟐 + 𝟔 𝒚𝒕+𝟏 − 𝟔 𝒚𝒕 = 𝟐
𝟏 𝟏
⏟
𝟏 𝒚𝒕+𝟐 + ( ) 𝒚𝒕+𝟏 + (− ) 𝒚𝒕 = ⏟𝟐
⏟𝟔 ⏟ 𝟔
𝑎 0 𝑐
𝑎1 𝑎2
𝟏 𝟏
Here; 𝒂𝟎 = 𝟏, 𝒂𝟏 = , 𝒂𝟐 = − , and 𝒄=𝟐≠𝟎
⏟ .
𝟔 𝟔
𝑐𝑎𝑠𝑒
Recalling two Schur Theorem determinants for second-order difference equation:
𝒂𝟎 𝒂𝟐 Therefore, we evaluate the second
∆𝟏 = |𝒂 𝒂 | ⋚ 𝟎
𝟐 𝟎 determinant (∆𝟐 ) in hope of another
𝟏 positive value.
𝟏 − 𝟑𝟓
∆𝟏 = | 𝟔 |= 𝟏 𝟏
𝟏 𝟑𝟔 𝟏 𝟎 −
− 𝟏 𝟔 𝟔
𝟔 | 𝟏 𝟏|
∆𝟏 > 𝟎 𝟏 𝟎 −
∆𝟐 = | 𝟔 𝟔|
First Determinant (∆𝟏 ) is found to be positive. It 𝟏 𝟏
follows the requirement Convergence in Schur − 𝟎 𝟏
| 𝟔 𝟔 |
Theorem which requires all of the determinants to 𝟏 𝟏
be positive. − 𝟎 𝟏
𝟔 𝟔
𝟏𝟏𝟕𝟔 𝟒𝟗
∆𝟐 = = >𝟎
𝟏𝟐𝟗𝟔 𝟓𝟒
Second determinant (∆𝟐 ) also follows the
requirement of Convergence in Schur
Theorem.
______________________________________________________________________________
Interpretation
(∆𝟏 < 𝟎) ∧ (∆𝟐 > 𝟎)
(∆𝟏 ∧ ∆𝟐 ) ≮ 𝟎 { ⋁ } (∆𝟏 ∧ ∆𝟐 ) > 𝟎
(∆𝟏 > 𝟎) ∧ (∆𝟐 < 𝟎)
Not applicable Not applicable CONVERGENT
Limitations:
1. Dynamic stability is investigated but the characteristic roots remain unknown.
2. For 𝒏 > 𝟐, the determinants become huge – contrary to our expectation of ease.
______________________________________________________________________________
Lesson 44
LINEAR AND NON-LINEAR PROGRAMMING
TOPIC 199: LINEAR PROGRAMMING: GLASSWARE MAKING
“A problem well stated is a problem half solved.”

Charles Franklin Kettering
Major Contributors were Leonid Kantorovich,
George Dantzig, John von Neumann during 1939
to 1947.
In addition to calculus-based optimization, algebra-based & graphical optimization can also be

done – Linear Programming. Linear programming addresses real life problems of allocating finite
resources such as labor, machine time or raw materials to maximize profits or minimize costs.
Process: Mathematize the given information into Decision Variables, Objective Function, and
(multiple) Constraints.
Limitation: It only deals with linear objective and constraint functions.
Generally, done in three steps.
Numerical Example: Glassware making involve processes like; glassblowing, annealing and
addition of ingredients like silica sand. Profit of the glassware firm depends on decision regarding
production of bowls (𝒙) and plates (𝒚):
𝒙 = number of bowls
𝒚 = number of plates
STEP – I: Mathematization of Given Information

𝝅 = 𝒇(𝒙, 𝒚)
Profit per bowl = $𝟏𝟓𝟎 ⇒ Total Profit from bowls = $𝟏𝟓𝟎 × 𝒙 = 𝟏𝟓𝟎𝒙
Profit per plate = $𝟏𝟎𝟎 ⇒ Total Profit from bowls = $𝟏𝟎𝟎 × 𝒚 = 𝟏𝟎𝟎𝒚
Total profit from both:
𝑂𝑏𝑗𝑒𝑐𝑡𝑖𝑣𝑒 𝐹𝑢𝑛𝑐𝑡𝑖𝑜𝑛
𝑡𝑜 𝑏𝑒 𝑚𝑎𝑥𝑖𝑚𝑖𝑧𝑒𝑑.
⏞
𝝅 = 𝟏𝟓𝟎𝒙
⏟ + 𝟏𝟎𝟎𝒚
⏟
𝜋𝑏𝑜𝑤𝑙𝑠 𝜋𝑝𝑙𝑎𝑡𝑒𝑠
Limited resources put constraints:
1st Constraint (Glassblowing Time): Total time available for glassblowing is 𝟕𝟎 hours per week.
A bowl takes 2 hours in making while a plate takes 1 hour:
𝑠ℎ𝑜𝑢𝑙𝑑 𝑏𝑒
𝑡𝑖𝑚𝑒 𝑐𝑜𝑛𝑠𝑢𝑚𝑒𝑑 𝑓𝑜𝑟 𝑔𝑙𝑎𝑠𝑠𝑏𝑙𝑜𝑤𝑖𝑛𝑔 ( 𝑙𝑒𝑠𝑠 𝑜𝑟 )
( )
𝑜𝑓 𝑏𝑜𝑤𝑙𝑠 𝑎𝑛𝑑 𝑝𝑙𝑎𝑡𝑒𝑠 𝑒𝑞𝑢𝑎𝑙
⏞ 𝟐𝒙
⏟ + 𝒚
⏟ ⏞
≤ 𝟕𝟎
⏟
𝑡𝑖𝑚𝑒 𝑛𝑒𝑒𝑑𝑒𝑑 𝑡𝑖𝑚𝑒 𝑛𝑒𝑒𝑑𝑒𝑑 𝑎𝑣𝑎𝑖𝑙𝑎𝑏𝑙𝑒
( 𝑓𝑜𝑟 𝑏𝑜𝑤𝑙𝑠 ) ( 𝑓𝑜𝑟 𝑝𝑙𝑎𝑡𝑒𝑠 ) (ℎ𝑜𝑢𝑟𝑠 𝑝𝑒𝑟 )
𝑤𝑒𝑒𝑘
2nd Constraint (Annealing Time): Total time available for Annealing is 𝟏𝟑𝟎 hours per week. A
bowl takes 4 hours in making while a plate takes 1 hour:
______________________________________________________________________________
𝑡𝑖𝑚𝑒 𝑐𝑜𝑛𝑠𝑢𝑚𝑒𝑑 𝑓𝑜𝑟 𝑎𝑛𝑛𝑒𝑎𝑙𝑖𝑛𝑔 ( 𝑙𝑒𝑠𝑠 𝑜𝑟 )
( )
𝑜𝑓 𝑏𝑜𝑤𝑙𝑠 𝑎𝑛𝑑 𝑝𝑙𝑎𝑡𝑒𝑠 𝑒𝑞𝑢𝑎𝑙
⏞ 𝟒𝒙
⏟ + 𝒚
⏟ ⏞
≤ 𝟏𝟑𝟎
⏟
𝑡𝑖𝑚𝑒 𝑛𝑒𝑒𝑑𝑒𝑑 𝑡𝑖𝑚𝑒 𝑛𝑒𝑒𝑑𝑒𝑑 𝑎𝑣𝑎𝑖𝑙𝑎𝑏𝑙𝑒
( 𝑓𝑜𝑟 𝑏𝑜𝑤𝑙𝑠 ) ( 𝑓𝑜𝑟 𝑝𝑙𝑎𝑡𝑒𝑠 ) (ℎ𝑜𝑢𝑟𝑠 𝑝𝑒𝑟 )
𝑤𝑒𝑒𝑘
3rd Constraint (mass of Silica Sand): Total amount available of Silica Sand is 𝟒𝟓 𝐤𝐠:
𝑚𝑎𝑠𝑠 𝑐𝑜𝑛𝑠𝑢𝑚𝑒𝑑 𝑜𝑓𝑆𝑖𝑙𝑖𝑐𝑎 ( 𝑙𝑒𝑠𝑠 𝑜𝑟 )
( )
𝑠𝑎𝑛𝑑 𝑓𝑜𝑟 𝑏𝑜𝑤𝑙𝑠 𝑎𝑛𝑑 𝑝𝑙𝑎𝑡𝑒𝑠 𝑒𝑞𝑢𝑎𝑙
⏞ ⏟
𝒙 + 𝒚
⏟ ⏞
≤ 𝟒𝟓
⏟
𝑚𝑎𝑠𝑠 𝑚𝑎𝑠𝑠 𝑎𝑣𝑎𝑖𝑙𝑎𝑏𝑙𝑒
( 𝑛𝑒𝑒𝑑𝑒𝑑 ) ( 𝑛𝑒𝑒𝑑𝑒𝑑 ) (ℎ𝑜𝑢𝑟𝑠 𝑝𝑒𝑟 )
𝑓𝑜𝑟 𝑏𝑜𝑤𝑙𝑠 𝑓𝑜𝑟 𝑏𝑜𝑤𝑙𝑠 𝑤𝑒𝑒𝑘
Since production cannot be negative, we develop 2 additional constraints:
Non-negativity Constraints:
4th Constraint (minimum one bowl):

⏟
𝒙 ≥
⏟ 𝟎
𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛 𝑐𝑎𝑛′ 𝑡 𝑏𝑒
( ) (𝑛𝑒𝑔𝑎𝑡𝑖𝑣𝑒 )
𝑜𝑓 𝑏𝑜𝑤𝑙𝑠
5th Constraint (minimum one plate):

𝒚
⏟ ≥
⏟ 𝟎
𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛 ( 𝑐𝑎𝑛′ 𝑡 𝑏𝑒 )
( ) 𝑛𝑒𝑔𝑎𝑡𝑖𝑣𝑒
𝑜𝑓 𝑝𝑙𝑎𝑡𝑒𝑠
Tabulating all the extracted information:

Objective (Profit) Function (to be maximized) 𝝅𝒎𝒂𝒙 = 𝟏𝟓𝟎𝒙 + 𝟏𝟎𝟎𝒚
Subject to:
1st Constraint (Glassblowing Time) 𝟐𝒙 + 𝒚 ≤ 𝟕𝟎
2nd Constraint (Annealing Time) 𝟒𝒙 + 𝒚 ≤ 𝟏𝟑𝟎
3rd Constraint (mass of Silica Sand) 𝒙 + 𝒚 ≤ 𝟒𝟓
th
4 Constraint (minimum one bowl) 𝒙≥𝟎
5th Constraint (minimum one plate) 𝒚≥𝟎
Plotting the Equalities from Inequalities Constraints: Reduce the inequalities to equalities that
are more meaningful to plot: ≥ ⇒ =
We use the intercepts to plot these linear functions:
Constraints Equalities/ 𝒙-intercept 𝒚-intercept
Inequalities
1st (Glassblowing Time) 𝟐𝒙 + 𝒚 = 𝟕𝟎 𝟐𝒙 + (𝟎) = 𝟕𝟎 𝟐(𝟎) + 𝒚 = 𝟕𝟎
≤ ⇒ 𝑏𝑒𝑙𝑜𝑤 𝑡ℎ𝑒 𝑐𝑢𝑟𝑣𝑒 𝒙 = 𝟑𝟓 ⇒ (𝟑𝟓, 𝟎) 𝒚 = 𝟕𝟎 ⇒ (𝟎, 𝟕𝟎)
2nd (Annealing Time) 𝟒𝒙 + 𝒚 = 𝟏𝟑𝟎 𝟒𝒙 + (𝟎) = 𝟏𝟑𝟎 𝟒(𝟎) + 𝒚 = 𝟏𝟑𝟎
≤ ⇒ 𝑏𝑒𝑙𝑜𝑤 𝑡ℎ𝑒 𝑐𝑢𝑟𝑣𝑒 𝒙 = 𝟑𝟐. 𝟓 ⇒ (𝟑𝟐. 𝟓, 𝟎) 𝒚 = 𝟏𝟑𝟎 ⇒ (𝟎, 𝟏𝟑𝟎)
3rd (mass of Silica Sand) 𝒙 + 𝒚 = 𝟒𝟓 𝒙 + (𝟎) = 𝟒𝟓 (𝟎) + 𝒚 = 𝟒𝟓
≤ ⇒ 𝑏𝑒𝑙𝑜𝑤 𝑡ℎ𝑒 𝑐𝑢𝑟𝑣𝑒 𝒙 = 𝟒𝟓 ⇒ (𝟒𝟓, 𝟎) 𝒚 = 𝟒𝟓 ⇒ (𝟎, 𝟒𝟓)
th st
4 (minimum one bowl) 𝒙≥𝟎 1 Quadrant
______________________________________________________________________________
5th (minimum one plate) 𝒚≥𝟎
STEP – II: Graphing of Constraints
STEP – III: Selection of Optimum

The highlighted feasible region satisfies all five given constraints. So, we choose the points from
boundary to find maximum profit.
Points (𝒙, 𝒚) Profit (𝝅 = 𝟏𝟓𝟎𝒙 + 𝟏𝟎𝟎𝒚)
(𝟎, 𝟎) 𝟏𝟓𝟎(𝟎) + 𝟏𝟎𝟎(𝟎) = 𝟎
(𝟎, 𝟒𝟓) 𝟏𝟓𝟎(𝟎) + 𝟏𝟎𝟎(𝟒𝟓) = 𝟒𝟓𝟎𝟎
(𝟐𝟓, 𝟐𝟎) 𝟏𝟓𝟎(𝟐𝟓) + 𝟏𝟎𝟎(𝟐𝟎) = 𝟓𝟕𝟓𝟎
(𝟑𝟎, 𝟏𝟎) 𝟏𝟓𝟎(𝟑𝟎) + 𝟏𝟎𝟎(𝟏𝟎) = 𝟓𝟓𝟎𝟎
(𝟑𝟐. 𝟓, 𝟎) 𝟏𝟓𝟎(𝟑𝟐. 𝟓) + 𝟏𝟎𝟎(𝟎) = 𝟒𝟖𝟕𝟓
Observation: While following all constraints, $𝟓𝟕𝟓𝟎 is the maximum amount of profit earned
by producing (𝒙∗ , 𝒚∗ ) = (𝟐𝟓, 𝟐𝟎) of bowls and plates, respectively.
TOPIC 200: LINEAR PROGRAMMING: EMPLOYMENT COST MINIMIZATION
After understanding maximization problem, we execute a minimization problem:
______________________________________________________________________________
Numerical Example: An insurance company employs full- and part-time staff, who work 𝟒𝟎 and
𝟐𝟎 hours per week, respectively. Full-time staff are paid $𝟖𝟎𝟎 per week and part-time staff $𝟑𝟐𝟎.
In addition, it is company policy that the number of part-time staff should not exceed one-third of
the number of full-time staff.
If the number of worker-hours per week required to deal with the company’s work is 𝟗𝟎𝟎, how
many workers of each type should be employed in order to complete the workload at minimum
cost?
Solution: Weekly Employment Cost of the insurance company depends full- and part-time staff:
𝒙 = full-time staff
𝒚 = part-time staff
STEP – I: Mathematization of Given Information
𝑪 = 𝒇(𝒙, 𝒚)
Wage per full-time worker = $𝟖𝟎𝟎 ⇒ Total Wage of full-time worker = $𝟖𝟎𝟎 × 𝒙 = 𝟖𝟎𝟎𝒙
Wage per part-time worker = $𝟑𝟐𝟎 ⇒ Total Wage of part-time worker = $𝟑𝟐𝟎 × 𝒚 = 𝟑𝟐𝟎𝒚
Total Wage Bill (Cost) from both types of workers:
𝑂𝑏𝑗𝑒𝑐𝑡𝑖𝑣𝑒 𝐹𝑢𝑛𝑐𝑡𝑖𝑜𝑛
𝑡𝑜 𝑏𝑒 𝑚𝑖𝑛𝑖𝑚𝑖𝑧𝑒𝑑.
⏞
𝝅 = 𝟖𝟎𝟎𝒙
⏟ + 𝟑𝟐𝟎𝒚
⏟
𝐶𝑓𝑢𝑙𝑙 𝐶𝑝𝑎𝑟𝑡
Limited resources put constraints:
1st Constraint (Total Worker Hours): Total time available for work is at least 𝟕𝟎 hours per week.
Full-time workers work 40 hours and Part-time workers work 20 hours:
𝑡𝑖𝑚𝑒 𝑐𝑜𝑛𝑠𝑢𝑚𝑒𝑑 𝑤𝑜𝑟𝑘𝑖𝑛𝑔 𝑏𝑦 ( 𝑚𝑜𝑟𝑒 𝑜𝑟 )
( )
𝑓𝑢𝑙𝑙− 𝑎𝑛𝑑 𝑝𝑎𝑟𝑡−𝑡𝑖𝑚𝑒 𝑤𝑜𝑟𝑘𝑒𝑟𝑠 𝑒𝑞𝑢𝑎𝑙
⏞ 𝟒𝟎𝒙
⏟ + 𝟐𝟎𝒚
⏟ ⏞
≥ 𝟕𝟎
⏟
𝑡𝑖𝑚𝑒 𝑐𝑜𝑛𝑠𝑢𝑚𝑒𝑑 𝑤𝑜𝑟𝑘𝑖𝑛𝑔 𝑡𝑖𝑚𝑒 𝑐𝑜𝑛𝑠𝑢𝑚𝑒𝑑 𝑤𝑜𝑟𝑘𝑖𝑛𝑔 𝑟𝑒𝑞𝑢𝑖𝑟𝑒𝑑
( ) ( ) (ℎ𝑜𝑢𝑟𝑠 𝑝𝑒𝑟 )
𝑏𝑦 𝑓𝑢𝑙𝑙−𝑡𝑖𝑚𝑒 𝑤𝑜𝑟𝑘𝑒𝑟𝑠 𝑏𝑦 𝑝𝑎𝑟𝑡−𝑡𝑖𝑚𝑒 𝑤𝑜𝑟𝑘𝑒𝑟𝑠
𝑤𝑒𝑒𝑘
2nd Constraint (Proportion of Part-time workers): Number of Part-time workers (𝒚) should be
𝟏
exceed 𝒓𝒅 Number of Full-time workers (𝒙):
𝟑
( 𝑙𝑒𝑠𝑠 𝑜𝑟 )
𝑒𝑞𝑢𝑎𝑙
𝟏
𝒚
⏟ ⏞
≤ × ⏟
𝒙
⏟
𝟑 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓
( 𝑃𝑎𝑟𝑡−𝑡𝑖𝑚𝑒 ) (𝑜𝑛𝑒−𝑡ℎ𝑖𝑟𝑑 𝑜𝑓) ( 𝐹𝑢𝑙𝑙−𝑡𝑖𝑚𝑒 )
𝑤𝑜𝑟𝑘𝑒𝑟𝑠 𝑤𝑜𝑟𝑘𝑒𝑟𝑠
Since worker cannot be negative, we develop 2 additional constraints:
Non-negativity Constraints:
3rd Constraint (minimum one Full-time worker):
⏟
𝒙 ≥
⏟ 𝟎
′
(𝐹𝑢𝑙𝑙−𝑡𝑖𝑚𝑒) ( 𝑐𝑎𝑛 𝑡 𝑏𝑒 )
𝑤𝑜𝑟𝑘𝑒𝑟𝑠 𝑛𝑒𝑔𝑎𝑡𝑖𝑣𝑒
4th Constraint (minimum one Part-time worker):
______________________________________________________________________________
𝒚
⏟ ≥
⏟ 𝟎
′
(𝑃𝑎𝑟𝑡−𝑡𝑖𝑚𝑒) ( 𝑐𝑎𝑛 𝑡 𝑏𝑒 )
𝑤𝑜𝑟𝑘𝑒𝑟𝑠 𝑛𝑒𝑔𝑎𝑡𝑖𝑣𝑒
Tabulating all the extracted information:

Objective (Cost) Function (to be minimized) 𝑪𝒎𝒊𝒏 = 𝟖𝟎𝟎𝒙 + 𝟑𝟐𝟎𝒚
Subject to
1st Constraint (Total Worker Hours) 𝟒𝟎𝒙 + 𝟐𝟎𝒚 ≥ 𝟗𝟎𝟎
nd
2 Constraint (Proportion of Part-time workers) 𝟏
𝒚≤ 𝒙
𝟑
3rd Constraint (minimum one Full-time worker) 𝒙≥𝟎
4th Constraint (minimum one Part-time worker) 𝒚≥𝟎
Plotting the Equalities from Inequalities Constraints: Reduce the inequalities to equalities that
are more meaningful to plot: ≥ & ≤ ⇒ =
We use the intercepts to plot these linear functions:
Constraints Equalities/ 𝒙-intercept 𝒚-intercept
Inequalities
1st Constraint (Total Worker 𝟒𝟎𝒙 𝟒𝟎𝒙 + 𝟐𝟎(𝟎) 𝟒𝟎(𝟎) + 𝟐𝟎𝒚
Hours) + 𝟐𝟎𝒚 = 𝟗𝟎𝟎 = 𝟗𝟎𝟎 = 𝟗𝟎𝟎
≥ ⇒ 𝑎𝑏𝑜𝑣𝑒 𝑡ℎ𝑒 𝑐𝑢𝑟𝑣𝑒 𝒙 = 𝟐𝟐. 𝟓 𝒚 = 𝟒𝟓
⇒ (𝟐𝟐. 𝟓, 𝟎) ⇒ (𝟑𝟓, 𝟎)
2nd
Constraint (Proportion of 𝑙𝑖𝑛𝑒𝑎𝑟 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 Positively sloped linear function:
Part-time workers) ⏞ 𝟏 𝟏 𝟐 𝟒
(𝟎, 𝟎), (𝟏, ), (𝟐, ), (𝟑, 𝟏), (𝟒, ) as
≤ ⇒ 𝑏𝑒𝑙𝑜𝑤 𝑡ℎ𝑒 𝑐𝑢𝑟𝑣𝑒 𝒚 = 𝒙 𝟑 𝟑 𝟑
⏟ 𝟑 shown in figure.
𝑦=𝑚𝑥
3rd Constraint (minimum one 𝒙≥𝟎
Full-time worker)
1st Quadrant
4th Constraint (minimum one 𝒚≥𝟎
Part-time worker)
STEP – II: Graphing of Constraints
______________________________________________________________________________
(? , ? ) are the coordinates of feasible area which can be found by simultaneously solving the two
equations, as (? , ? ) they intersect each other.
𝟏 𝟏
𝒚 = 𝟑 𝒙 and 𝟒𝟎𝒙 + 𝟐𝟎 ⏟ 𝒚 = 𝟗𝟎𝟎 ⇒ 𝟒𝟎𝒙 + 𝟐𝟎 (𝟑) 𝒙 = 𝟗𝟎𝟎
𝟏
= 𝒙
𝟑
𝟏𝟑𝟓 𝟏 𝟏 𝟏𝟑𝟓 𝟒𝟓
𝒙= 𝟕
& 𝒚 = (𝟑) ⏟
𝒙 ⇒ 𝒚 = (𝟑) 𝟕
= 𝟕
𝟏𝟑𝟓
=
𝟕
(𝒙, 𝒚) = (𝟏𝟗. 𝟑, 𝟔. 𝟒)
STEP – III: Selection of Optimum

The feasible region satisfies all five given constraints. So, we choose the points from boundary to
find minimum cost.
Points (𝒙, 𝒚) Employment Cost (𝑪 = 𝟖𝟎𝟎𝒙 + 𝟑𝟐𝟎𝒚)
(𝟎, 𝟎) 𝟖𝟎𝟎(𝟎) + 𝟑𝟐𝟎(𝟎) = 𝟎
(𝟏𝟗. 𝟑, 𝟔. 𝟒) 𝟖𝟎𝟎(𝟏𝟗. 𝟑) + 𝟑𝟐𝟎(𝟔. 𝟒) = 𝟏𝟕𝟒𝟖𝟓. 𝟕
(𝟐𝟐. 𝟓, 𝟎) 𝟖𝟎𝟎(𝟐𝟐. 𝟓) + 𝟑𝟐𝟎(𝟎) = 𝟏𝟖𝟎𝟎𝟎
Observation: While following all constraints, $𝟏𝟕𝟒𝟖𝟓. 𝟕 is the minimum amount of
employment cost incurred by insurance company (𝟏𝟗. 𝟑, 𝟔. 𝟒) of Full- and Part-time workers,
respectively.
Need for Integer Programming:

Workers cannot be in fractions and (𝟏𝟗. 𝟑, 𝟔. 𝟒) seem somewhat awkward. For this, one can
experiment WHOLE numbers like (𝟐𝟎, 𝟓), (𝟐𝟎, 𝟔), (𝟐𝟏, 𝟓), (𝟐𝟏, 𝟔) and (𝟐𝟏, 𝟕) among others.
______________________________________________________________________________
Integer Programming
Points (𝒙, 𝒚) Employment Cost (𝑪 = 𝟖𝟎𝟎𝒙 + 𝟑𝟐𝟎𝒚)

(𝟐𝟎, 𝟓) 𝟖𝟎𝟎(𝟐𝟎) + 𝟑𝟐𝟎(𝟓) = 𝟏𝟕𝟔𝟎𝟎
(𝟐𝟎, 𝟔) 𝟖𝟎𝟎(𝟐𝟎) + 𝟑𝟐𝟎(𝟔) = 𝟏𝟕𝟗𝟐𝟎
(𝟐𝟏, 𝟓) 𝟖𝟎𝟎(𝟐𝟏) + 𝟑𝟐𝟎(𝟓) = 𝟏𝟖𝟒𝟎𝟎
(𝟐𝟏, 𝟔) 𝟖𝟎𝟎(𝟐𝟏) + 𝟑𝟐𝟎(𝟔) = 𝟏𝟖𝟕𝟐𝟎
(𝟐𝟏, 𝟕) 𝟖𝟎𝟎(𝟐𝟏) + 𝟑𝟐𝟎(𝟕) = 𝟏𝟗𝟎𝟒𝟎
Interpretation: In this case (𝟐𝟎, 𝟓) shall give the minimum value of cost, making (𝟐𝟎, 𝟓) even
more meaningful than (𝟏𝟗. 𝟑, 𝟔. 𝟒). This additional step is known as Integer Programming.
𝑖𝑛𝑡𝑒𝑔𝑒𝑟 𝑝𝑟𝑜𝑔𝑟𝑎𝑚𝑚𝑖𝑛𝑔 𝑙𝑖𝑛𝑒𝑎𝑟 𝑝𝑟𝑜𝑔𝑟𝑎𝑚𝑚𝑖𝑛𝑔
Though ⏟ ⏞
(𝑪 = 𝟏𝟕𝟔𝟎𝟎) > ⏟⏞
(𝑪 = 𝟏𝟕𝟒𝟖𝟓. 𝟕) , yet this compromise was necessary as workers
𝑎𝑡 (𝟐𝟎,𝟓) 𝑎𝑡 (𝟏𝟗.𝟑,𝟔.𝟒)
cannot be in fractions.
TOPIC 201: NONLINEAR PROGRAMMING: UTILITY MAXIMIZATION
For nonlinear objective and constraint functions, linear programming cannot be used – need for
Nonlinear Programming. It can handle Nonlinear Objective Functions as well as nonlinear
inequality constraints.
Process: Nonlinear programming has a similar procedure. However, it differs in terms of
conditions known as Kuhn-Tucker conditions (1951).
______________________________________________________________________________
Kuhn-Tucker conditions are similar to first order conditions.

Kuhn-Tucker Conditions for Maximization: For 𝒋 number of variables, there are 𝒋 number of
conditions. For 𝒏 number of variables, there are 𝒏 number of conditions:
Variable-Based Conditions
𝝏𝒁
𝝏𝒁 ⇒⏟ 𝒙𝒋 ( )=𝟎
( ) ≤ 𝟎 𝒙𝒋 ≥ 𝟎 𝑒𝑖𝑡ℎ𝑒𝑟 𝑜𝑟 ⏟ 𝝏𝒙𝒋 Where 𝒋 = 𝟏, 𝟐, … , 𝒏
𝝏𝒙𝒋 𝑏𝑜𝑡ℎ (=0) 𝑐𝑜𝑚𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑟𝑦
𝑠𝑙𝑎𝑐𝑘𝑛𝑒𝑠𝑠
Multiplier-Based Conditions
𝝏𝒁
𝝏𝒁 ⇒
⏟ 𝝀𝒊 ( ) = 𝟎
( )≥𝟎 𝝀𝒊 ≥ 𝟎 𝑒𝑖𝑡ℎ𝑒𝑟 𝑜𝑟 ⏟ 𝝏𝝀𝒊 Where 𝒊 = 𝟏, 𝟐, … , 𝒎
𝝏𝝀𝒊 𝑏𝑜𝑡ℎ (=0) 𝑐𝑜𝑚𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑟𝑦
Numerical Example:
Objective Function 𝒑𝒓𝒐𝒅𝒖𝒄𝒕 𝒕𝒆𝒓𝒎−𝒏𝒐𝒏𝒍𝒊𝒏𝒆𝒂𝒓
⏞
𝑼 = 𝒙𝒚
𝒙 + 𝒚 ≤ 𝟏𝟎𝟎
General Constraints
𝒙 ≤ 𝟒𝟎
Nonnegativity Constraints 𝒙, 𝒚 ≥ 𝟎
STEP – I: Lagrangian Function:

Introduce the two constraints and hence two multipliers.
1𝑠𝑡 2𝑛𝑑
𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑒𝑟 𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑒𝑟
𝒁(𝒙, 𝒚, 𝝀𝟏 , 𝝀𝟐 ) = 𝒙𝒚
⏟ + ⏞𝟏
𝝀 (𝟏𝟎𝟎
⏟ − 𝒙 − 𝒚) + ⏞𝟐
𝝀 (𝟒𝟎
⏟ − 𝒙)
𝑜𝑏𝑗𝑒𝑐𝑡𝑖𝑣𝑒 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 1𝑠𝑡 𝑐𝑜𝑛𝑠𝑡𝑟𝑎𝑖𝑛𝑡 2𝑛𝑑 𝑐𝑜𝑛𝑠𝑡𝑟𝑎𝑖𝑛𝑡
STEP – II: Kuhn-Tucker Conditions – Minimization:

For two variables (𝒙, 𝒚) there are two conditions and for two multipliers (𝝀𝟏 , 𝝀𝟐 ) there are also two
conditions.
𝝏𝒁 ⇒
⏟ 𝝏𝒁
( ) ≤ 𝟎 𝒙 ≥ 𝟎 𝑒𝑖𝑡ℎ𝑒𝑟 𝑜𝑟 𝒙( ) = 𝟎
𝝏𝒙 𝑏𝑜𝑡ℎ (=0) 𝝏𝒙
𝝏𝒁 ⇒
⏟ 𝝏𝒁
( ) ≤ 𝟎 𝒚 ≥ 𝟎 𝑒𝑖𝑡ℎ𝑒𝑟 𝑜𝑟 𝒚( ) = 𝟎
𝝏𝒚 𝑏𝑜𝑡ℎ (=0) 𝝏𝒚
𝝏𝒁 ⇒
⏟ 𝝏𝒁
( ) ≥ 𝟎 𝝀𝟏 ≥ 𝟎 𝑒𝑖𝑡ℎ𝑒𝑟 𝑜𝑟 𝝀𝟏 ( )=𝟎
𝝏𝝀𝟏 𝑏𝑜𝑡ℎ (=0) 𝝏𝝀𝟏
𝝏𝒁 ⇒
⏟ 𝝏𝒁
( ) ≥ 𝟎 𝝀𝟐 ≥ 𝟎 𝑒𝑖𝑡ℎ𝑒𝑟 𝑜𝑟 𝝀𝟐 ( )=𝟎
𝝏𝝀𝟐 𝑏𝑜𝑡ℎ (=0) 𝝏𝝀𝟐
Developing First Order Conditions:

𝝏𝒁 𝜹 𝝏𝒁
= {𝒙𝒚 + 𝝀𝟏 (𝟏𝟎𝟎 − 𝒙 − 𝒚) + 𝝀𝟐 (𝟒𝟎 − 𝒙)} ≤ 𝟎 ⇒ = 𝒚 − 𝝀𝟏 − 𝝀 𝟐 ≤ 𝟎
𝝏𝒙 𝜹𝒙 𝝏𝒙
= {𝒙𝒚 + 𝝀𝟏 (𝟏𝟎𝟎 − 𝒙 − 𝒚) + 𝝀𝟐 (𝟒𝟎 − 𝒙)} ≤ 𝟎 ⇒ = 𝒙 − 𝝀𝟏 ≤ 𝟎
𝝏𝒚 𝜹𝒚 𝝏𝒚
______________________________________________________________________________
= {𝒙𝒚 + 𝝀𝟏 (𝟏𝟎𝟎 − 𝒙 − 𝒚) + 𝝀𝟐 (𝟒𝟎 − 𝒙)} ≥ 𝟎 ⇒ = 𝟏𝟎𝟎 − 𝒙 − 𝒚 ≥ 𝟎
𝝏𝝀𝟏 𝜹𝝀𝟏 𝝏𝝀𝟏
= {𝒙𝒚 + 𝝀𝟏 (𝟏𝟎𝟎 − 𝒙 − 𝒚) + 𝝀𝟐 (𝟒𝟎 − 𝒙)} ≥ 𝟎 ⇒ = 𝟒𝟎 − 𝒙 ≥ 𝟎
𝝏𝝀𝟐 𝜹𝝀𝟐 𝝏𝝀𝟐
Converting the inequalities into equalities:
𝝏𝒁
= 𝒚 − 𝝀𝟏 − 𝝀 𝟐 = 𝟎
𝝏𝒙
𝝏𝒁
= 𝒙 − 𝝀𝟏 = 𝟎
𝝏𝒚
𝝏𝒁
= 𝟏𝟎𝟎 − 𝒙 − 𝒚 = 𝟎
𝝏𝝀𝟏
𝝏𝒁
= 𝟒𝟎 − 𝒙 = 𝟎
𝝏𝝀𝟐
STEP – III: Trial Solution Cases:

We try two trial solutions:
Trial Solution Case – I: 𝝀𝟏 , 𝝀𝟐 > 𝟎 Trial Solution, Case – II: 𝒙, 𝒚 > 𝟎
𝜹𝒁
(𝝀𝟏 > 𝟎) ∧ 𝝀𝟏 ( ) = 𝟎
𝜹𝒁 (𝒙 > 𝟎) ∧ 𝒙 (⏟) = 𝟎
𝜹𝒙
⏟𝜹𝝀𝟏
𝑡ℎ𝑒𝑛
𝑡ℎ𝑒𝑛 𝜹𝒁
𝜹𝒁
( )=𝟎 { ( )=𝟎
𝜹𝒙 }
{ 𝜹𝝀𝟏 }
𝜹𝒁 𝜹𝒁
(𝝀𝟐 > 𝟎) ∧ 𝝀𝟐 ( ) = 𝟎 (𝒚 > 𝟎) ∧ 𝒚 ( ) = 𝟎
⏟𝜹𝝀𝟐 ⏟𝜹𝒚
𝑡ℎ𝑒𝑛 𝑡ℎ𝑒𝑛
𝜹𝒁 𝜹𝒁
( )=𝟎 { ( )=𝟎 }
{ 𝜹𝝀𝟐 } 𝜹𝒚
𝜹𝒁 𝜹𝒁 𝜹𝒁 𝜹𝒁
𝝀𝟏 , 𝝀𝟐 > 𝟎 ⇒ {( ),( )} = 𝟎 𝒙, 𝒚 > 𝟎 ⇒ {( ) , ( )} = 𝟎
𝜹𝝀𝟏 𝜹𝝀𝟐 𝜹𝒙 𝜹𝒚
𝜹𝒁 𝜹𝒁
Trial Solution Case – I: 𝝀𝟏 , 𝝀𝟐 > 𝟎 ⇒ {(𝜹𝝀 ) , (𝜹𝝀 )} = 𝟎
𝟏 𝟐
𝜹𝒁 𝜹𝒁
{( ),( )} = 𝟎
𝜹𝝀𝟏 𝜹𝝀𝟐
𝜹𝒁 𝜹𝒁
( ) =𝟎 ( ) =𝟎
𝜹𝝀𝟏 𝜹𝝀𝟏
(𝟏𝟎𝟎 − 𝒙 − 𝒚) = 𝟎
⏟ (𝟒𝟎 − 𝒙) = 𝟎
⏟
(𝑦=100−𝑥) 𝒙=𝟒𝟎
𝒚 = 𝟏𝟎𝟎 − 𝒙
𝒚 = 𝟏𝟎𝟎 − 𝟒𝟎
𝒚 = 𝟔𝟎
Trial Solution Case – I gives us values of both of the variables (𝒙∗ , 𝒚∗ ) = (𝟒𝟎, 𝟔𝟎). Multipliers
can also be found.
______________________________________________________________________________
𝜹𝒁 𝜹𝒁
For values of 𝝀𝟏 , 𝝀𝟐 substitute 𝒙∗ and 𝒚∗ in (𝜹𝒙) and (𝜹𝒚).
𝜹𝒁 𝜹𝒁
( )=𝟎 ( )=𝟎
⏟𝜹𝒚 ⏟𝜹𝒙
𝜹𝒁 𝜹𝒁
𝑎𝑠 𝒚>𝟎,( )>𝟎 𝑎𝑠 𝒙>𝟎,( )>𝟎
𝜹𝒚 𝜹𝒙
𝒙⏟− 𝝀𝟏 = 𝟎 ⇒ 𝝀𝟏 = 𝟒𝟎 𝒚 −𝝀
⏟ ⏟𝟏 − 𝝀𝟐 = 𝟎 ⇒ 𝝀𝟐 = 𝟐𝟎
(𝑥=40)
⏟ 60 40
𝜹𝒁
( =𝟎)
𝜹𝝀𝟐
(𝝀𝟏 , 𝝀𝟐 ) = (𝟒𝟎, 𝟐𝟎)
Verification: We found (𝒙, 𝒚) = (𝟒𝟎, 𝟔𝟎) and (𝝀𝟏 , 𝝀𝟐 ) = (𝟒𝟎, 𝟐𝟎). Now we check if the values
satisfy all of the functions:
Objective Function 𝑼 = 𝒙𝒚 𝑼 = (𝟒𝟎)(𝟔𝟎) = 𝟐𝟒𝟎𝟎
𝒙 + 𝒚 ≤ 𝟏𝟎𝟎 𝟒𝟎 + 𝟔𝟎 ≤ 𝟏𝟎𝟎
General Constraints
𝒙 ≤ 𝟒𝟎 𝟒𝟎 ≤ 𝟒𝟎
Nonnegativity Constraints 𝒙, 𝒚 ≥ 𝟎 𝟒𝟎, 𝟔𝟎 ≥ 𝟎
All constraints are satisfied. (𝒙, 𝒚) = (𝟒𝟎, 𝟔𝟎) are the undisputed values at equilibrium that
maximize the utility (𝟐𝟒𝟎𝟎).
TOPIC 202: NONLINEAR PROGRAMMING: COST MINIMIZATION
Kuhn-Tucker Conditions for Minimization: For 𝒋 number of variables, there are 𝒋 number of
conditions. For 𝒏 number of variables, there are 𝒏 number of conditions:
𝝏𝒁
𝝏𝒁 ⇒
⏟ 𝒙𝒋 ( )=𝟎
( ) ≥ 𝟎 𝒙𝒋 ≥ 𝟎 𝑒𝑖𝑡ℎ𝑒𝑟 𝑜𝑟 ⏟ 𝝏𝒙𝒋 Where 𝒋 = 𝟏, 𝟐, … , 𝒏
𝝏𝒙𝒋 𝑏𝑜𝑡ℎ (=0) 𝑐𝑜𝑚𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑟𝑦
𝝏𝒁
𝝏𝒁 ⇒
⏟ 𝝀𝒊 ( ) = 𝟎
( )≤𝟎 𝝀𝒊 ≥ 𝟎 𝑒𝑖𝑡ℎ𝑒𝑟 𝑜𝑟 ⏟ 𝝏𝝀𝒊 Where 𝒊 = 𝟏, 𝟐, … , 𝒎
𝝏𝝀𝒊 𝑏𝑜𝑡ℎ (=0) 𝑐𝑜𝑚𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑟𝑦
𝝏𝒁
Note: Complementary Slackness is used for the product as the two terms 𝝀𝒊 and (𝝏𝝀 )
𝒊
complement each other in fading to slackness (𝟎).
Numerical Example:
Objective Function 𝑸𝒖𝒂𝒅𝒓𝒂𝒕𝒊𝒄 (𝒏𝒐𝒏𝒍𝒊𝒏𝒆𝒂𝒓) 𝒇𝒖𝒏𝒄𝒕𝒊𝒐𝒏
𝑪= ⏞ (𝒙𝟏 − 𝟒)𝟐 + (𝒙𝟐 − 𝟒)𝟐

𝟐𝒙𝟏 + 𝟑𝒙𝟐 ≥ 𝟔
General Constraints
−𝟑𝒙𝟏 − 𝟐𝒙𝟐 ≥ −𝟏𝟐
Nonnegativity Constraints 𝒙𝟏 , 𝒙𝟐 ≥ 𝟎
STEP – I: Lagrangian Function:

Introduce the two constraints and hence two multipliers.
______________________________________________________________________________
1𝑠𝑡
(𝒙𝟏 − 𝟒)𝟐 + (𝒙𝟐 − 𝟒)𝟐 +
𝒁(𝒙𝟏 , 𝒙𝟐 , 𝝀𝟏 , 𝝀𝟐 ) = ⏟ ⏞𝟏
𝝀 (𝟔 − 𝟐𝒙𝟏 − 𝟑𝒙𝟐 ) +
⏟
𝑜𝑏𝑗𝑒𝑐𝑡𝑖𝑣𝑒 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 1𝑠𝑡 𝑐𝑜𝑛𝑠𝑡𝑟𝑎𝑖𝑛𝑡
2𝑛𝑑
⏞𝟐
𝝀 (−𝟏𝟐 + 𝟑𝒙𝟏 + 𝟐𝒙𝟐 )
⏟
2𝑛𝑑 𝑐𝑜𝑛𝑠𝑡𝑟𝑎𝑖𝑛𝑡
STEP – II: Kuhn-Tucker Conditions – Minimization:

For two variables (𝒙𝟏 , 𝒙𝟐 ) there are two conditions and for two multipliers (𝝀𝟏 , 𝝀𝟐 ) there are also
two conditions.
𝝏𝒁 ⇒
⏟ 𝝏𝒁
( ) ≥ 𝟎 𝒙𝟏 ≥ 𝟎 𝑒𝑖𝑡ℎ𝑒𝑟 𝑜𝑟 𝒙𝟏 ( )=𝟎
𝝏𝒙𝟏 𝑏𝑜𝑡ℎ (=0) 𝝏𝒙𝟏
𝝏𝒁 ⇒
⏟ 𝝏𝒁
( ) ≥ 𝟎 𝒙𝟐 ≥ 𝟎 𝑒𝑖𝑡ℎ𝑒𝑟 𝑜𝑟 𝒙𝟐 ( )=𝟎
𝝏𝒙𝟐 𝑏𝑜𝑡ℎ (=0) 𝝏𝒙𝟐
𝝏𝒁 ⇒
⏟ 𝝏𝒁
( ) ≤ 𝟎 𝝀𝟏 ≥ 𝟎 𝑒𝑖𝑡ℎ𝑒𝑟 𝑜𝑟 𝝀𝟏 ( )=𝟎
𝝏𝝀𝟏 𝑏𝑜𝑡ℎ (=0) 𝝏𝝀𝟏
𝝏𝒁 ⇒
⏟ 𝝏𝒁
( ) ≤ 𝟎 𝝀𝟐 ≥ 𝟎 𝑒𝑖𝑡ℎ𝑒𝑟 𝑜𝑟 𝝀𝟐 ( )=𝟎
𝝏𝝀𝟐 𝑏𝑜𝑡ℎ (=0) 𝝏𝝀𝟐
Developing First Order Conditions:

𝝏𝒁 𝜹
= {(𝒙𝟏 − 𝟒)𝟐 + (𝒙𝟐 − 𝟒)𝟐 + 𝝀𝟏 (𝟔 − 𝟐𝒙𝟏 − 𝟑𝒙𝟐 ) + 𝝀𝟐 (−𝟏𝟐 + 𝟑𝒙𝟏 + 𝟐𝒙𝟐 )} ≥ 𝟎
𝝏𝒙𝟏 𝜹𝒙𝟏
𝝏𝒁
= 𝟐(𝒙𝟏 − 𝟒) − 𝟐𝝀𝟏 + 𝟑𝝀𝟐 ≥ 𝟎
𝝏𝒙𝟏
𝝏𝒁 𝜹
= {(𝒙𝟏 − 𝟒)𝟐 + (𝒙𝟐 − 𝟒)𝟐 + 𝝀𝟏 (𝟔 − 𝟐𝒙𝟏 − 𝟑𝒙𝟐 ) + 𝝀𝟐 (−𝟏𝟐 + 𝟑𝒙𝟏 + 𝟐𝒙𝟐 )} ≥ 𝟎
𝝏𝒙𝟐 𝜹𝒙𝟐
𝝏𝒁
= 𝟐(𝒙𝟐 − 𝟒) − 𝟑𝝀𝟏 + 𝟐𝝀𝟐 ≥ 𝟎
𝝏𝒙𝟐
𝝏𝒁 𝜹
= {(𝒙𝟏 − 𝟒)𝟐 + (𝒙𝟐 − 𝟒)𝟐 + 𝝀𝟏 (𝟔 − 𝟐𝒙𝟏 − 𝟑𝒙𝟐 ) + 𝝀𝟐 (−𝟏𝟐 + 𝟑𝒙𝟏 + 𝟐𝒙𝟐 )}
𝝏𝝀𝟏 𝜹𝝀𝟏
𝝏𝒁
= 𝟔 − 𝟐𝒙𝟏 − 𝟑𝒙𝟐 ≤ 𝟎
𝝏𝝀𝟏
𝝏𝒁 𝜹
= {(𝒙𝟏 − 𝟒)𝟐 + (𝒙𝟐 − 𝟒)𝟐 + 𝝀𝟏 (𝟔 − 𝟐𝒙𝟏 − 𝟑𝒙𝟐 ) + 𝝀𝟐 (−𝟏𝟐 + 𝟑𝒙𝟏 + 𝟐𝒙𝟐 )}
𝝏𝝀𝟐 𝜹𝝀𝟐
𝝏𝒁
= −𝟏𝟐 + 𝟑𝒙𝟏 + 𝟐𝒙𝟐 ≤ 𝟎
𝝏𝝀𝟐
Converting the inequalities into equalities:
𝝏𝒁
= 𝟐(𝒙𝟏 − 𝟒) − 𝟐𝝀𝟏 + 𝟑𝝀𝟐 = 𝟎
𝝏𝒙𝟏
𝝏𝒁
= 𝟐(𝒙𝟐 − 𝟒) − 𝟑𝝀𝟏 + 𝟐𝝀𝟐 = 𝟎
𝝏𝒙𝟐
𝝏𝒁
= 𝟔 − 𝟐𝒙𝟏 − 𝟑𝒙𝟐 = 𝟎
𝝏𝝀𝟏
______________________________________________________________________________
𝝏𝒁
= −𝟏𝟐 + 𝟑𝒙𝟏 + 𝟐𝒙𝟐 = 𝟎
𝝏𝝀𝟐
STEP – III: Trial Solution Cases:

We try two trial solutions:
Trial Solution Case – I: 𝝀𝟏 , 𝝀𝟐 > 𝟎 Trial Solution, Case – II: 𝒙𝟏 , 𝒙𝟐 > 𝟎
𝜹𝒁 𝜹𝒁
(𝝀𝟏 > 𝟎) ∧ 𝝀𝟏 ( ) = 𝟎 (𝒙𝟏 > 𝟎) ∧ 𝒙𝟏 ( ) = 𝟎
⏟𝜹𝝀𝟏 ⏟𝜹𝒙𝟏
𝜹𝒁 𝜹𝒁
{ ( )=𝟎 } { ( )=𝟎 }
𝜹𝝀𝟏 𝜹𝒙𝟏
𝜹𝒁 𝜹𝒁
(𝝀𝟐 > 𝟎) ∧ 𝝀𝟐 ( ) = 𝟎 (𝒙𝟐 > 𝟎) ∧ 𝒙𝟐 ( ) = 𝟎
⏟𝜹𝝀𝟐 ⏟𝜹𝒙𝟐
𝜹𝒁 𝜹𝒁
{ ( } )=𝟎 { ( )=𝟎}
𝜹𝝀𝟐 𝜹𝒙𝟐
𝜹𝒁 𝜹𝒁 𝜹𝒁 𝜹𝒁
𝝀𝟏 , 𝝀𝟐 > 𝟎 ⇒ {( ),( )} = 𝟎 𝒙𝟏 , 𝒙𝟐 > 𝟎 ⇒ {( ),( )} = 𝟎
𝜹𝝀𝟏 𝜹𝝀𝟐 𝜹𝒙𝟏 𝜹𝒙𝟐
𝜹𝒁 𝜹𝒁
Trial Solution Case – I: 𝝀𝟏 , 𝝀𝟐 > 𝟎 ⇒ {(𝜹𝝀 ) , (𝜹𝝀 )} = 𝟎
𝟏 𝟐
𝜹𝒁 𝜹𝒁
{( ),( )} = 𝟎
𝜹𝒁 𝜹𝒁
( ) =𝟎 ( ) =𝟎
(𝟔
⏟ − 𝟐𝒙𝟏 − 𝟑𝒙𝟐 ) = 𝟎 (−𝟏𝟐 + 𝟑 𝒙𝟏 + 𝟐𝒙𝟐 ) = 𝟎
𝟔−𝟑𝒙𝟐
(𝒙𝟏 = )
𝟐
𝟔 − 𝟑𝒙𝟐
−𝟏𝟐 + 𝟑+ 𝟐𝒙𝟐 = 𝟎
𝟐
−𝟐𝟒+𝟑(𝟔−𝟑𝒙𝟐 )+𝟒𝒙𝟐
𝟐
= 𝟎 ⇒ −𝟐𝟒 + 𝟑(𝟔 − 𝟑𝒙𝟐 ) + 𝟒𝒙𝟐 = 𝟎
−𝟐𝟒 + 𝟏𝟖 − 𝟗𝒙𝟐 + 𝟒𝒙𝟐 = 𝟎
𝟔
−𝟔 = 𝟓𝒙𝟐 ⇒ 𝒙𝟐 = −
𝟓
𝟔 𝟏𝟖 𝟑𝟎 + 𝟏𝟖
𝟔 − 𝟑𝒙𝟐 𝟔 − 𝟑 (− 𝟓) 𝟔 + 𝟓 𝟓 𝟐𝟒
𝒙𝟏 = ⇒ ⇒ = ⇒ 𝒙𝟏 =
𝟐 𝟐 𝟐 𝟐 𝟖
𝟐𝟒 𝟔
(𝒙𝟏 , 𝒙𝟐 ) = ( , −
⏟𝟓 ); Nonnegativity constraint is being violated for 𝒙𝟐 .
𝟖
(𝒙𝟏 ≮𝟎)
Trial Solution Case – I violates the nonnegativity conditions, so we try Trial Solution Case – II.
𝜹𝒁 𝜹𝒁
Trial Solution Case – II: 𝒙𝟏 , 𝒙𝟐 > 𝟎 ⇒ {(𝜹𝒙 ) , (𝜹𝒙 )} = 𝟎
𝟏 𝟐
______________________________________________________________________________
𝜹𝒁 𝜹𝒁
{( ),( )} = 𝟎
𝜹𝒙𝟏 𝜹𝒙𝟐
𝜹𝒁 𝜹𝒁
( ) =𝟎 ( ) =𝟎
𝜹𝒙𝟏 𝜹𝒙𝟐
𝟐(𝒙𝟏 − 𝟒) − 𝟐𝝀𝟏 + 𝟑𝝀𝟐 = 𝟎
⏟ 𝟐(𝒙𝟐 − 𝟒) − 𝟑𝝀𝟏 + 𝟐𝝀𝟐 = 𝟎
⏟
2𝜆 −2(𝑥1 −4) 3𝜆1 −2(𝑥2 −4)
{𝜆2 = 1 } {𝜆2 = }
3 2
Equating (𝝀𝟐 ), we get:
𝟐𝝀𝟏 − 𝟐(𝒙𝟏 − 𝟒) 𝟑𝝀𝟏 − 𝟐(𝒙𝟐 − 𝟒)
=
𝟑 𝟐
𝟐𝝀𝟏 − 𝟐𝒙𝟏 + 𝟖 𝟑𝝀𝟏 − 𝟐𝒙𝟐 + 𝟖
=
𝟑 𝟐
𝟐(𝟐𝝀𝟏 − 𝟐𝒙𝟏 + 𝟖) = 𝟑(𝟑𝝀𝟏 − 𝟐𝒙𝟐 + 𝟖)
𝟒𝝀𝟏 − 𝟒𝒙𝟏 + 𝟏𝟔 = 𝟗𝝀𝟏 − 𝟔𝒙𝟐 + 𝟐𝟒
𝟎 = −𝟒𝝀𝟏 + 𝟒𝒙𝟏 − 𝟏𝟔 + 𝟗𝝀𝟏 − 𝟔𝒙𝟐 + 𝟐𝟒
𝟒𝒙𝟏 − 𝟔𝒙𝟐 + 𝟓𝝀𝟏 + 𝟖 = 𝟎
⏟
3 𝑢𝑛𝑘𝑛𝑜𝑤𝑛𝑠⇒𝑙𝑒𝑡 𝝀𝟏 =𝟎
𝟒𝒙𝟏 − 𝟔𝒙𝟐 + 𝟓(𝟎) + 𝟖 = 𝟎
𝟒𝒙𝟏 − 𝟔𝒙𝟐 + 𝟖 = 𝟎
𝟐𝒙𝟏 − 𝟑𝒙𝟐 = −𝟒
𝟐𝒙𝟏 − 𝟑𝒙𝟐 = −𝟒 ⇒ need for another equation in terms of (𝒙𝟏 , 𝒙𝟐 ), to find values of 𝒙𝟏 and 𝒙𝟐 .
⏟
𝒇(𝒙𝟏 ,𝒙𝟐 )
𝜹𝒁
Using condition 𝝀𝟐 > 𝟎 ⇒ {(𝜹𝝀 )} = 𝟎
𝟐
𝜹𝒁
( )=𝟎
𝜹𝝀𝟐
−𝟏𝟐 + 𝟑𝒙𝟏 + 𝟐𝒙𝟐 = 𝟎
𝟑𝒙𝟏 + 𝟐𝒙𝟐 = 𝟏𝟐
⏟
𝒈(𝒙𝟏 ,𝒙𝟐 )
𝟐𝒙
⏟ 𝟏 − 𝟑𝒙𝟐 = −𝟒 𝟑𝒙
⏟ 𝟏 + 𝟐𝒙𝟐 = 𝟏𝟐
𝒇(𝒙𝟏 ,𝒙𝟐 ) 𝒈(𝒙𝟏 ,𝒙𝟐 )
Solving simultaneously for (𝒙𝟏 , 𝒙𝟐 )
𝟐𝒙
⏟𝟏 − 𝟑𝒙𝟐 = −𝟒 𝟑 𝒙𝟏 + 𝟐𝒙𝟐 = 𝟏𝟐
−𝟒+𝟑𝒙𝟐
𝒙𝟏 =
𝟐
−𝟒 + 𝟑𝒙𝟐 𝟗𝒙𝟐
𝟑 + 𝟐𝒙𝟐 = 𝟏𝟐 ⇒ −𝟔 + + 𝟐𝒙𝟐 = 𝟏𝟐
𝟐 𝟐
𝟏𝟑𝒙𝟐 𝟑𝟔
= 𝟏𝟖 ⇒ 𝒙𝟐 =
𝟐 𝟏𝟑
−𝟒+𝟑𝒙𝟐
For 𝒙𝟏 , we substitute 𝒙𝟐 in 𝒙𝟏 = 𝟐
.
𝟑𝟔
−𝟒 + 𝟑 (𝟏𝟑) 𝟐𝟖
𝒙𝟏 = ⇒ 𝒙𝟏 =
𝟐 𝟏𝟑
∗ ∗ 𝟐𝟖 𝟑𝟔
(𝒙𝟏 , 𝒙𝟐 ) = ( , )
𝟏𝟑 𝟏𝟑
𝜹𝒁 𝜹𝒁
For values of 𝝀𝟏 , 𝝀𝟐 substitute 𝒙∗𝟏 and 𝒙∗𝟐 in (𝜹𝒙 ) and (𝜹𝒙 ).
𝟏 𝟐
𝟐(𝒙∗𝟏 − 𝟒) − 𝟐𝝀𝟏 + 𝟑𝝀𝟐 = 𝟎 𝟐(𝒙∗𝟐 − 𝟒) − 𝟑𝝀𝟏 + 𝟐𝝀𝟐 = 𝟎
______________________________________________________________________________
𝟐𝟖 𝟑𝟔
𝟐( − 𝟒) − 𝟐𝝀𝟏 + 𝟑𝝀𝟐 = 𝟎 𝟐( − 𝟒) − 𝟑𝝀𝟏 + 𝟐𝝀𝟐 = 𝟎
𝟏𝟑 𝟏𝟑
𝟒𝟖 𝟑𝟐
−𝟐𝝀𝟏 + 𝟑𝝀𝟐 = −𝟑 𝝀𝟏 + 𝟐𝝀𝟐 =
⏟ 𝟏𝟑 𝟏𝟑
𝟏 𝟒𝟖
𝝀𝟏 = (𝟑𝝀𝟐 − )
𝟐 𝟏𝟑
𝟏 𝟒𝟖 𝟑𝟐 𝟗 𝟏𝟒𝟒 𝟑𝟐 𝟏𝟔
−𝟑 (𝟑𝝀𝟐 − ) + 𝟐𝝀𝟐 = ⇒ − 𝝀𝟐 + + 𝟐𝝀𝟐 = ⇒ 𝝀𝟐 =
𝟐 𝟏𝟑 𝟏𝟑 𝟐 𝟐𝟔 𝟏𝟑 𝟏𝟑
𝟏 𝟒𝟖 𝟏 𝟏𝟔 𝟒𝟖
𝝀𝟏 = (𝟑 𝝀𝟐 − ) ⇒ 𝝀𝟏 = (𝟑 − ) ⇒ 𝝀𝟏 = 𝟎
𝟐 𝟏𝟑 𝟐 𝟏𝟑 𝟏𝟑
𝟏𝟔
(𝝀𝟏 , 𝝀𝟐 ) = (𝟎, )
𝟏𝟑
𝟐𝟖 𝟑𝟔 𝟏𝟔
Verification: We found (𝒙∗𝟏 , 𝒙∗𝟐 ) = ( , ) and (𝝀𝟏 , 𝝀𝟐 ) = (𝟎, ). Now we check if the values
𝟏𝟑 𝟏𝟑 𝟏𝟑
satisfy all of the functions:
Objective 𝟐𝟖 𝟐 𝟑𝟔 𝟐 𝟔𝟖
𝑪 = (𝒙𝟏 − 𝟒)𝟐 + (𝒙𝟐 − 𝟒)𝟐 𝑪 = (𝟏𝟑 − 𝟒) + (𝟏𝟑 − 𝟒) = 𝟏𝟑
Function
𝟐𝟖 𝟑𝟔
𝟐𝒙𝟏 + 𝟑𝒙𝟐 ≥ 𝟔 𝟐 (𝟏𝟑) + 𝟑 (𝟏𝟑) = 𝟏𝟐. 𝟔 ≥ 𝟔
General
Constraints 𝟐𝟖 𝟑𝟔
−𝟑𝒙𝟏 − 𝟐𝒙𝟐 ≥ −𝟏𝟐 −𝟑 ( ) − 𝟐 ( ) = −𝟏𝟐 ≥ −𝟏𝟐
𝟏𝟑 𝟏𝟑
Nonnegativity 𝟐𝟖 𝟑𝟔
𝒙𝟏 , 𝒙𝟐 ≥ 𝟎 , ≥𝟎
Constraints 𝟏𝟑 𝟏𝟑
𝟐𝟖 𝟑𝟔
All constraints are satisfied. (𝒙∗𝟏 , 𝒙∗𝟐 ) = ( , ) are the undisputed values at equilibrium that
𝟏𝟑 𝟏𝟑
𝟔𝟖
minimize the cost ( ).
𝟏𝟑
______________________________________________________________________________

Complete Handouts Eco607 (1-44)

Uploaded by

Copyright:

Available Formats

Complete Handouts Eco607 (1-44)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Complete Handouts Eco607 (1-44)

Uploaded by

Copyright:

Available Formats

Mathematical Economics II (ECO607) VU

TOPIC 001: NEED AND IMPORTANCE OF DYNAMIC ANALYSIS IN ECONOMICS

Pre-requisites: Basic Mathematical Economics

TOPIC 002: DYNAMIC ANALYSIS: EXAMPLE FROM POPULATION

TOPIC 003: INTRODUCTION TO INTEGRAL CALCULUS: INDEFINITE INTEGRALS

RATIONALE FOR INTEGRATION:

𝒚 = ∫ 𝒇(𝒙) 𝒅𝒙 = ∫(𝟏) 𝒅𝒙 = (𝟏)𝒙 + 𝒄

∫ 𝒚 𝒅𝒚 = 𝒂 ∙ {𝑭(𝒙)} + 𝒄 [∵ ∫ 𝒇(𝒙) 𝒅𝒙 = 𝑭(𝒙)]

∫ 𝒚 𝒅𝒚 = 𝟓 ∙ {𝑭(𝒙)} + 𝒄 [∵ ∫ 𝒇(𝒙) 𝒅𝒙 = 𝑭(𝒙)]

TOPIC 005: POWER RULE OF INTEGRATION

Power Rule and Constant Function Rule

TOPIC 006: DERIVATIVE OVER FUNCTION RULE OF INTEGRATION

Recalling and applying the derivative over function rule:

TOPIC 007: EXPONENTIAL FUNCTION RULE OF INTEGRATION

TOPIC 008: NATURAL EXPONENTIAL FUNCTION RULE OF INTEGRATION

TOPIC 009: SUM-DIFFERENCE RULE OF INTEGRATION

∫{𝒈(𝒙) − 𝒉(𝒙) + 𝒊(𝒙)} 𝒅𝒙 = ∫ 𝒈(𝒙) 𝒅𝒙 − ∫ 𝒉(𝒙) 𝒅𝒙 + ∫ 𝒊(𝒙) 𝒅𝒙 + 𝒄

TOPIC 010: MORE COMPLEX RULES: INTEGRATION BY SUBSTITUTION

Example 5: ∫ 𝟏𝟐𝒙𝟐 (𝒙𝟑 + 𝟐) 𝒅𝒙

TOPIC 011: MORE COMPLEX RULES: INTEGRATION BY PARTS

Rationale for integration by parts:

Formula for Integration by Parts:

Derivation of the Rule: Reversing Product Theorem

𝒈(𝒙)𝒉(𝒙) − ∫{𝒈(𝒙)𝒉′ (𝒙)} 𝒅𝒙 = ∫{𝒉(𝒙)𝒈′ (𝒙)} 𝒅𝒙

∫{𝒉(𝒙)𝒈′ (𝒙)} 𝒅𝒙 = 𝒈(𝒙)𝒉(𝒙) − ∫{𝒈(𝒙)𝒉′ (𝒙)} 𝒅𝒙

∫{𝒉(𝒙)𝒈′ (𝒙)} 𝒅𝒙 = 𝒈(𝒙)𝒉(𝒙) − ∫{𝒈(𝒙)𝒉′ (𝒙)} 𝒅𝒙

Example: ∫ 𝟒𝒙(𝒙 + 𝟏)𝟑 𝒅𝒙

Given extracted terms; 𝒉(𝒙) = 𝟒𝒙 , 𝒈′ (𝒙) = (𝒙 + 𝟏)𝟑 .

Plugging values in the formula:

= {𝒙(𝒙 + 𝟏)𝟒 + 𝟒𝒄𝟏 𝒙} − ∫{(𝒙 + 𝟏)𝟒 }𝒅𝒙 − ∫ 𝟒𝒄𝟏 𝒅𝒙

TOPIC 012: DEMAND ANALYSIS USING INTEGRATION BY SUBSTITUTION

𝒑(𝒙) = −𝟏𝟑𝟓 √𝟗 + 𝒙𝟐 + 𝒄 is a general solution as 𝒄 is still arbitrary in nature.

Graph of 𝒑(𝒙) = −𝟏𝟑𝟓 √𝟗 + 𝒙𝟐 + 𝟕𝟎𝟓

c. How many units are demanded at a price of PKR 20 per unit?

TOPIC 013: SUPPLY ANALYSIS USING INTEGRATION BY SUBSTITUTION

TOPIC 014: REVENUE ANALYSIS USING INTEGRATION BY SUBSTITUTION

Given the rate of change of Revenue function;

Applying integration by parts to solve underlined integrand:

If 𝟏, 𝟎𝟎𝟎 units are being produced then revenue shall be:

TOPIC 015: INVESTMENT PORTFOLIO ANALYSIS USING INTEGRATION BY

An investment portfolio changes value at the rate:

Given the rate of change of investment portfolio function;

𝑽(𝒖) = 𝟏𝟐 ∫ {𝒆𝒖 (𝒆 −𝟔𝒖 − 𝟑)} (−𝟐𝟎𝒅𝒖)

𝑽(𝒖) = 𝟏𝟐 × (−𝟐𝟎) ∫{𝒆𝒖 (𝒆−𝟔𝒖 − 𝟑)} 𝒅𝒖

𝑽(𝒖) = −𝟐𝟒𝟎 ∫{𝒆𝒖 (𝒆−𝟔𝒖 − 𝟑)} 𝒅𝒖

𝑽(𝒖) = −𝟐𝟒𝟎 ∫(𝒆−𝟓𝒖 − 𝟑𝒆𝒖 ) 𝒅𝒖 ⇒ 𝟐𝟒𝟎 ∫(𝟑𝒆𝒖 − 𝒆−𝟓𝒖 ) 𝒅𝒖

𝑽(𝒖) = 𝟐𝟒𝟎 {∫(𝟑𝒆𝒖 ) 𝒅𝒖 − ∫(𝒆−𝟓𝒖) 𝒅𝒖}

𝑽(𝒖) = 𝟐𝟒𝟎 {𝟑 ∫(𝒆𝒖 ) 𝒅𝒖 − ∫(𝒆−𝟓𝒖 ) 𝒅𝒖}

TOPIC 016: DEFINITE INTEGRALS

By the virtue of indefinite integral, we know:

TOPIC 017: DEFINITE INTEGRAL: AREA UNDER CURVE

𝑨 = 𝒇(𝒙𝒐 ) × ( 𝒙𝟏 − 𝒙𝒐 ) + 𝒇(𝒙𝟏 ) × (𝒙𝟐 − 𝒙𝟏 ) + 𝒇(𝒙𝟐 ) × (𝒙𝟑 − 𝒙𝟐 ) + 𝒇(𝒙𝟑 ) × (𝒙𝟒 − 𝒙𝟑 ) + 𝒇(𝒙𝟒 )

𝑨 = 𝒇(𝒙𝟏 ) × (𝒙𝟐 − 𝒙𝟏 ) + 𝒇(𝒙𝟐 ) × (𝒙𝟑 − 𝒙𝟐 ) + 𝒇(𝒙𝟑 ) × (𝒙𝟒 − 𝒙𝟑 ) + 𝒇(𝒙𝟒 ) × (𝒙𝟓 − 𝒙𝟒 ) + 𝒇(𝒙𝟓 )

TOPIC 018: PROPERTIES OF DEFINITE INTEGRALS

Property – 3: Definite integral can be expressed as a sum of a finite number of definite

Property – 4: A coefficient in definite integral appears a coefficient outside the integral:

∫{𝒈(𝒙) ± 𝒉(𝒙)} 𝒅𝒙 = ∫{𝒈(𝒙) 𝒅𝒙} ± ∫{𝒉(𝒙)} 𝒅𝒙 + 𝒄

TOPIC 019: IMPROPER INTEGRALS

Improper implies ‘unfitting’ – something not in accordance with accepted standards

1. INFINITE LIMITS OF INTEGRATION: