20

Nonhomogeneous Equations in General

Now that we know how to solve a couple of rather broad classes of homogeneous equations, it
is time to start looking at nonhomogeneous equations.

20.1 Basic Theory

Recollections about Linearity∗
Let us go back to our generic, N th -order, linear differential equation,

dN y d N −1 y d2 y dy
a0 N
+ a 1 N −1
+ · · · + a N −2 2
+ a N −1 + aN y = g .
dx dx dx dx
Remember, g and the ak ’s denote known functions of x over some interval of interest, I . As
usual, we will assume these functions are continuous and that a0 is never zero on this interval.
As before, it is convenient to let L denote the corresponding differential operator from the
left side of the equation,

dN d N −1 d2 d
L = a0 N
+ a 1 N −1
+ · · · + a N −2 2
+ a N −1 + aN .
dx dx dx dx
That is, given any sufficiently differentiable function φ(x) on I ,

dNφ d N −1 φ d 2φ dφ
L[φ] = a0 N
+ a 1 N −1
+ · · · + a N −2 2
+ a N −1 + aN φ .
dx dx dx dx
Using this operator, we can write our generic differential equation as

L[y] = g .

Remember, this equation is said to be homogeneous if g is always zero on our interval, and
nonhomogeneous otherwise. Since we have already discussed the homogeneous case, let us now
assume g(x) is a function that is nonzero over at least a portion of our interval of interest.
Don’t forget, however, that, for each nonhomogeneous equation

L[y(x)] = g(x) ,
∗ You may want to briefly review the material in chapter 12.

409
410 Nonhomogeneous Equations in General

we still have the corresponding homogeneous equation

L[y(x)] = 0

where we simply replace g(x) with 0 . This equation will play a significant role in solving the
nonhomogeneous equation.

General Solutions to Nonhomogeneous Equations

Recall that, if we have a bunch of sufficiently differentiable functions — φ1 (x) , φ2 (x) , …,
φ K (x) (where K is some positive integer) — and a corresponding set of constants — c1 , c2 ,
…, c K — then

L[c1 φ1 (x) + c2 φ2 (x) + · · · + c K φ K (x)]

= c1 L[φ1 (x)] + c2 L[φ2 (x)] + · · · + c K L[φ K (x)] .

We used this to construct general solutions to homogeneous equations as linear combinations of

different solutions. With nonhomogeneous equations we must be a little more careful. After all,
if y p (x) and yq (x) are two particular solutions to a nonhomogeneous equation

L[y(x)] = g(x)

(i.e., L[y p (x)] = g(x) and L[yq (x)] = g(x) ),

and c1 and c2 are any two constants that do not add up to 1 , then, since g is a nonzero function,

L[c1 y p (x) + c2 yq (x)] = c1 L[y p (x)] + c2 L[yq ]

= c1 g(x) + c2 g(x)
= [c1 + c2 ]g(x)
6= g(x) .

Thus, an arbitrary linear combination of solutions to a NONhomogeneous linear differential

equation is a normally NOT a solution to the differential equation.
Notice, however, what happens when we consider the difference

yq (x) − y p (x)

between these two particular solutions to our nonhomogeneous equation. Plugging this into L
gives us

L[yq (x) − y p (x)] = L[yq (x)] − L[y p (x)] = g(x) − g(x) = 0 .

So
yq (x) − y p (x) = a solution to the corresponding homogeneous equation .
Let me rephrase this:
If y p and yq are any two solutions to a given nonhomogeneous linear differential
equation, then

yq (x) = y p (x) + a solution to the corresponding homogeneous equation .

Basic Theory 411

On the other hand, if we start with

yq (x) = y p (x) + y0 (x)

where y p is any particular solution to the nonhomogeneous equation and y0 is any solution to
the corresponding homogeneous equation


i.e., L[y p (x)] = g(x) and L[y0 (x)] = 0 ,

then

L[yq (x)] = L[y p (x) + y0 (x)] = L[y p (x)] + L[y0 ] = g(x) + 0 = g(x) .

Thus:

If y p is a particular solution to a given nonhomogeneous linear differential equation,

and

yq (x) = y p (x) + any solution to the corresponding homogeneous equation ,

then yq is also a solution to the nonhomogeneous differential equation.

If you think about it, you will realize that we’ve just derived the form for a general solution
to a given nonhomogeneous linear differential equation. We only need one particular solution
to that nonhomogeneous differential equation and the general formula describing all solutions
to the corresponding homogeneous linear differential equation. To be precise, the two results
derived above, combined, yield the following theorem.

Theorem 20.1 (general solutions to nonhomogeneous equations)

A general solution to a given nonhomogeneous linear differential equation is given by

y(x) = y p (x) + yh (x)

where y p is any particular solution to the nonhomogeneous equation, and yh is a general solution
to the corresponding homogeneous differential equation.1

If we limit ourselves to second-order equations and recall how we can construct general
solutions to the corresponding homogeneous equations, then we get the following corollary of
the above theorem:

Corollary 20.2 (general solutions to nonhomogeneous second-order equations)

A general solution to a second-order, nonhomogeneous linear differential equation

ay ′′ + by ′ + cy = g

is given by
y(x) = y p (x) + c1 y1 (x) + c2 y2 (x) (20.1)
1 Many texts refer to the general solution of the corresponding homogeneous differential equation as “the comple-
mentary solution” and denote it by yc instead of yh . We are using yh to help remind us that this is the general
solution to the corresponding homogeneous differential equation.
412 Nonhomogeneous Equations in General

where y p is any particular solution to the nonhomogeneous equation, and {y1 , y2 } is any fun-
damental set of solutions for the corresponding homogeneous equation
ay ′′ + by ′ + cy = 0 .

Do note that there are only two arbitrary constants c1 and c2 in formula (20.1), and that
they are multiplying only particular solutions to the corresponding homogeneous equation. The
particular solution to the nonhomogeneous equation, y p , is NOT multiplied by an arbitrary
constant!

!◮Example 20.1: Consider the nonhomogeneous differential equation

y ′′ − 4y = 5e3x . (20.2)
Observe that  3x ′′
− 4 e3x = 32 e3x − 4e3x = 5e3x
 
e .
So one particular solution to our nonhomogeneous equation is
y p (x) = e3x .
The corresponding homogeneous equation is
y ′′ − 4y = 0 ,
a linear equation with constant coefficients. Its characteristic equation,
r2 − 4 = 0 ,
has solutions r = 2 and r = −2 . So
y1 (x) , y2 (x) = e2x , e−2x
 

is a fundamental set of solutions to this homogeneous equation. According to our work above
(summarized in theorem 20.1 and corollary 20.2), the general solution to the nonhomogeneous
differential equation (20.2) is
y(x) = y p (x) + c1 y1 (x) + c2 y2 (x)
= e3x + c1 e2x + c2 e−2x .

Of course, if we don’t limit ourselves to second-order equations, but still recall how to
construct general solutions to homogeneous equations from a fundamental set of solutions to
that homogeneous equation, then we get the N th -order analog of the last corollary:

Corollary 20.3 (general solutions to nonhomogeneous N th -order equations)

A general solution to an N th -order, nonhomogeneous linear differential equation
a0 y (N ) + a1 y (N −1) + · · · + a N −2 y ′′ + a N −1 y ′ + a N y = g
is given by
y(x) = y p (x) + c1 y1 (x) + c2 y2 (x) + · · · + c N y N (x) (20.3)
where y p is any particular solution to the nonhomogeneous equation, and {y1 , y2 , . . . , y N } is
any fundamental set of solutions for the corresponding homogeneous equation
a0 y (N ) + a1 y (N −1) + · · · + a N −2 y ′′ + a N −1 y ′ + a N y = 0 .
Additional Exercises 413

20.2 Superposition for Nonhomogeneous Equations

Before discussing methods for finding particular solutions, we should use linearity to make one
final observation: If g1 (x) and g2 (x) are two functions, and

L[y1 (x)] = g1 (x) and L[y2 (x)] = g2 (x) ,

then
L[y1 (x) + y2 (x)] = L[y1 (x)] + L[y2 (x)] = g1 (x) + g2 (x) .
This fact is sometimes called the principle of superposition for nonhomogeneous equations.2 It
tells us that a particular solution to

L[y(x)] = g1 (x) + g2 (x)

can be constructed by just adding a particular solution of

L[y(x)] = g1 (x)
to a particular solution of
L[y(x)] = g2 (x) .

More generally, of course, if we have K functions y p,1 , y p,2 , . . . , y p,K satisfying

L[y p,1 ] = g1 , L[y p,2 ] = g2 , ... and L[y p,K ] = g K

for known functions g1 , g2 , . . . , gk , then

y p = y p,1 + y p,2 + · · · + y p,K

will be a particular solution to

L[y] = g1 + g2 + · · · + gk .

Thus, if we have a problem involving an equation whose right side is the sum of several
functions, we can convert this probem to a collection of somewhat simpler problems, and then
add up the solutions to those simpler problems to get a solution to the original problem.

20.3 Reduction of Order

In practice, finding a particular solution to a nonhomogeneous linear differential equation can be a
challenge. One method, the basic reduction of order method for second-order, nonhomogeneous
linear differential equations, was briefly discussed in section 13.3. If you haven’t already looked
at that section, or don’t remember the basic ideas discussed there, you can go back and skim that
section. Or not. Truth is, better methods will be developed in the next few sections.
2 You might want to compare this principle of superposition to the principle of superposition for homogeneous
equations was described in theorem 12.2 on page 266.
414 Nonhomogeneous Equations in General

Additional Exercises

20.1. What should g(x) be so that y(x) = e3x is a solution to

a. y ′′ + y = g(x) ?
b. x 2 y ′′ − 4y = g(x) ?
c. y (3) − 4y ′ + 5y = g(x) ?

20.2. What should g(x) be so that y(x) = x 3 is a solution to

a. y ′′ + 4y ′ + 4y = g(x) ?
b. x 2 y ′′ + 4x y ′ + 4y = g(x) ?
3 ′
c. y (4) + x y (3) + 4y ′′ − y = g(x) ?
x
20.3 a. Can y(x) = sin(x) be a solution to

y ′′ + y = g(x)

for some nonzero function g ? (Give a reason for your answer.)

b. What should g(x) be so that y(x) = x sin(x) is a solution to

y ′′ + y = g(x) ?

20.4. Consider the nonhomogeneous linear differential equation

y ′′ + 4y = 24e2x .

a. Verify that one particular solution to this nonhomogeneous differential equation is

y p (x) = 3e2x .

b. What is yh , the general solution to the corresponding homogeneous equation?

c. What is the general solution to the above nonhomogeneous equation?
d. Find the solution to the above nonhomogeneous equation that also satisfies each of
the following sets of initial conditions:
i. y(0) = 6 and y ′ (0) = 6 ii. y(0) = −2 and y ′ (0) = 2

20.5. Consider the nonhomogeneous linear differential equation

y ′′ + 2y ′ − 8y = 8x 2 − 3 .

a. Verify that one particular solution to this equation is

1
y p (x) = −x 2 − x .
2
Additional Exercises 415

b. What is yh , the general solution to the corresponding homogeneous equation?

c. What is the general solution to the above nonhomogeneous equation?
d. Find the solution to the above nonhomogeneous equation that also satisfies each of
the following sets of initial conditions:
i. y(0) = 0 and y ′ (0) = 0 ii. y(0) = 1 and y ′ (0) = −3

20.6. Consider the nonhomogeneous linear differential equation

y ′′ − 9y = 36 .

a. Verify that one particular solution to this equation is

y p (x) = −4 .

b. Find the general solution to this nonhomogeneous equation.

c. Find the solution to the above nonhomogeneous equation that also satisfies

y(0) = 8 and y ′ (0) = 6 .

20.7. Consider the nonhomogeneous linear differential equation

y ′′ + 6y ′ + 9y = 169 sin(2x) .

a. Verify that one particular solution to this equation is

y p (x) = 5 sin(2x) − 12 cos(2x) .

b. Find the general solution to this nonhomogeneous equation.

c. Find the solution to the above nonhomogeneous equation that also satisfies

y(0) = −10 and y ′ (0) = 9 .

20.8. Consider the nonhomogeneous linear differential equation

y ′′ − 3y ′ − 10y = −6e4x .

a. Verify that one particular solution to this equation is

y p (x) = e4x .

b. Find the general solution to this nonhomogeneous equation.

c. Find the solution to the above nonhomogeneous equation that also satisfies

y(0) = 6 and y ′ (0) = 8 .

20.9. Consider the nonhomogeneous linear differential equation

y ′′ − 3y ′ − 10y = 7e5x .
416 Nonhomogeneous Equations in General

a. Verify that one particular solution to this equation is

y p (x) = xe5x .

b. Find the general solution to this nonhomogeneous equation.

c. Find the solution to the above nonhomogeneous equation that also satisfies

y(0) = 12 and y ′ (0) = −2 .

20.10. Consider the nonhomogeneous linear differential equation

x 2 y ′′ − 4x y ′ + 6y = 10x + 12 for x >0 .

a. Verify that one particular solution to this equation is

y p (x) = 5x + 2 .

b. Find the general solution to this nonhomogeneous equation.

c. Find the solution to the above nonhomogeneous equation that also satisfies

y(1) = 6 and y ′ (1) = 8 .

20.11. Consider the nonhomogeneous linear differential equation

y (4) + y ′′ = 1 .

a. Verify that one particular solution to this equation is

1 2
y p (x) = x .
2
b. Find the general solution to this nonhomogeneous equation.
c. Find the solution to the above nonhomogeneous equation that also satisfies

y(0) = 4 , y ′ (0) = 3 , y ′′ (0) = 0 and y (3) (0) = 2 .

20.12. In exercises 20.8 and 20.9 you saw that y p (x) = e4x is a particular solution to

y ′′ − 3y ′ − 10y = −6e4x ,

and that y p (x) = xe5x is a particular solution to

y ′′ − 3y ′ − 10y = 7e5x .

Using this and superposition, find a particular solution to

a. y ′′ − 3y ′ − 10y = 7e5x − 6e4x
b. y ′′ − 3y ′ − 10y = 14e5x