000971-0007

International Baccalaureate
World School no. 000971

Extended Essay
MATHEMATICS

From Mice Pursuit Problem to Laplace transform.

SESSION: May 2014

Author: Aleksander Tadeusz Janczewski

Candidate number: 000971-0007
Supervisor: Mr. Krzysztof Michalik

[Word count: 3477]

1
 000971-0007
Abstract

Cyclic pursuit problems’ internal symmetry enables statement of problem

conditions in a very compact form. Laplace transform allows for solving linear
differential equations in a short and ‘elegant’ way. This essay combines both of them by
answering the research question: What are the trajectories of mice pursuing one
another in a regular polygon and how does properties of Laplace transform allow one to
solve this problem?
In the beginning the mice problem is introduced in the simplest regular polygon
(equilateral triangle). Parametrization and initial analysis is performed – this leads to
formulation of a system of two linear differential equations.
The need of method of solving the system of differential equations introduces
the reader to the next part of the essay, where the Laplace transform of real variable
functions is defined and its properties are investigated – linearity, transforms of
derivatives (complete proof included) and Lerch theorem (uniqueness of transforms)
among others. A table of Laplace transforms of a number of useful functions is
presented (some of the results are derived in given examples). Instances of application
of the transform to solving differential equations are also given.
Subsequently, already introduced methods are applied to solving the previously
stated system of linear differential equations – the trajectories of mice in an equilateral
triangle are found explicitly. The generalization of the solution is presented afterwards.
Solution of a general system of equations results in explicit parametric equations of
mice trajectories in a regular n-polygon, examples of solutions are presented in form of
graphs.
Analysis of usefulness of Laplace transform in solving the problem is presented
– the key properties identified are: mapping linear differential equations of order one
into linear algebraic equations and uniqueness of transforms (guaranteeing uniqueness
of solutions). Remarks on potential further research and possible alternative
parametrization are added.

Word count: 295

2
 000971-0007
Table of content

I. Introduction 4

II. Introduction of the mice problem 5

III. Laplace transform and its properties 7

IV. Laplace transform of derivatives 14

V. Laplace transform and its application to solving linear differential equations 18

VI. Laplace transform and Mice problem 21

VII. Conclusion 32

VIII. Bibliography 33

IX. Appendix 34

3
 000971-0007

I. Introduction

Pursuit problems, where pursued moved along a straight line, were studied by
Leonardo da Vinci and the general case was analysed by P. Bouger in 1732. Cyclic
pursuit problems (also known as mice, bugs, missiles, dogs, ships problems) have an
appealing internal symmetry. The mice problem can be traced back to H. Brocard and
was popularized by H. Steinhaus. In the form of the problem adapted in this paper, mice
are placed in the vertices of a regular polygon and each mouse chases its nearest right-
hand neighbour. Analysis of the positions of pursuer and pursued subject leads to
formulation of a system of differential equations.
Laplace transform is an example of the integral transform, which has been
developed by a mathematician and astronomer Pierre-Simon Laplace1 in his work on
probability theory. Laplace transform has various applications – it is used in
probability, signal analysis, solving electrical circuits, solving linear differential
equations and their systems among others - it is recognized as a useful tool in physics
and engineering. See [1].
This essay aim is a short introduction to Laplace transforms method of solving
linear differential equations and its application to solving mice pursuit problem. My
research question is: What are the trajectories of mice pursuing one another in a regular
polygon and how does properties of the Laplace transform allow one to solve this
problem?

1 Pierre-Simon Laplace was a french mathematician, astronomer and physicist, who is one of authors of Probability theory. Laplace
also formulated Laplace equation and developed Laplace transform. He lived from 23rd March 1749 to 5th March 1827.
4
 000971-0007

II. Introduction of the mice problem

For initial analysis let’s chose a particular regular polygon – e.g. an equilateral
triangle (the ‘simplest’ regular polygon). Mice are placed in vertices of the triangle
(facing interior of the polygon). Let’s assume that the mice speeds are equal – we
assume that the mice behave in a ‘symmetrical’ way, i.e. in the same way with respect to
their (symmetrical) position in the polygon and each mouse chases its nearest right-
hand neighbour.

Let ( ) ( ( ), ( )) represent the position of mouse no. 1 at time , where

( ) and we choose ( , ). In a similar way ( ) is a position of mouse no. 2
with initial condition ( ) and we choose ( , ). Similarly, ( ) is a position
√
of mouse no. 3 with initial condition ( ) where ( , ) as we choose the

length of triangle’s side as our unit. Mice positions (after some time) are symmetrical.
See the figure below.

Figure 1. Triangle in which mice chase each other.2

2 The figures are obtained by means of free software GeoGebra 4.2.19.0 .

5
 000971-0007

From the symmetry of the system it may be concluded that the relation between the
position of mouse no. 1 and mouse no. 2 may be represented by means of the following
transformations:
( ) rotated around ( , ) and translated by ⃗⃗⃗⃗⃗⃗⃗⃗ gives ( ),

so

( , ) ( )( ,
( , ))

( , )
√
( ,√ ),

where represents rotation3 by angle around origin and ⃗ stands for translation4
by vector .

Vector ⃗⃗⃗⃗ ( ) ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ ( ) (see Figure 1) represents the direction of the movement of
mouse no. 1 (as it chases mouse no. 2) at time , hence it is tangent to the path of
mouse no. 1 at ( ) so its gradient is the gradient of the path in ( ):
√ ( ) ( )
( ) ( )
. ( )
( ) ( ) ( ) √ ( )
Moreover we can notice that

√ ( ) ( )
,
( ) √ ( )

so a5 solution may be obtained by solving the following system of equations,

√
( ) ( ),
( )
√
( ) ( ) ,
{
which is a system of linear differential equations of order one.

3 See Appendix A1 for description of rotations in number-plane.

4 See Appendix A2 for description of translations in number-plane.
5 The ‘a’ means a possible solution but from the statement of the problem we deduce defined behaviour of the mice (at each
moment mouse ‘knows what to do’ which leads to a conclusion that a solution will be the unique one) i.e. we postulate existence of
unique solution to this problem.
6
 000971-0007

In order to solve the problem one may use the Laplace transform, which (among other
applications) allows solving differential equations. Therefore the purpose of this essay
is also to introduce the reader to the Laplace transform and show him or her how the
use of this mathematical tool allows solving the mice problem.

7
 000971-0007

III. The Laplace transform and its properties

Definition 1.
In this paper we will consider the Laplace transform that maps a real6 function ( ) of
real variable t into a corresponding real function ( ) of real variable s defined by

( ( )) ( ) ∫ ( ) st ,

so it is an example of a integral transform; ( ) is appropriately defined if and only if

the integral is convergent. When the integral is divergent and then there is no Laplace
transform defined for a function. See [7].
The notation ( ( )) is used to represent the Laplace transform of function
( ), the resulting function being ( ). The domain of the resulting function’s
independent variable is chosen appropriately so that the integral of Laplace
transform is convergent (if possible). The usual use of t as an independent variable of
function ( ) is due to the transform’s roots from physics – t denoting time. See [8].

Example 1.
Let’s consider a very simple example of Laplace transform of ( ) :

( ( )) ( ) ( ) ∫ st ∫ st .

Now let’s consider following cases.

i) If

∫ st ∫ .

So, for integral is divergent and thus the Laplace transform is undefined.
ii) If

∫ st [ st ]|

( ) ,
[ s ] {
( ) , .

6 We will focus on real functions, however, in general, the Laplace transform maps functions of complex variables.

8
 000971-0007

So, for the integral defining Laplace transform is divergent, so the transform is
undefined there; for Laplace transform of ( ) exists and is equal to . Hence,

Laplace transform of ( ) is well defined for ( , ): ( ) .

Example 2.
Let’s consider Laplace transform of a ( ) .
e begin with

( ) ∫ st ∫ st ,

as ( st) when for any real s,

so Laplace transform of ( ) is undefined.

The two examples presented show that Laplace transform is not always (not for any
( )) well defined. That implies a need of specifying some restrictions to ( ) to
guarantee existence of its Laplace transform. An example of such a set of restrictions is:

f(t) together with it’s ( ) is continuous for [ , ); (O1)

f(t) for ( , ); (O2)
f(t) is a function of exponential order of s i.e.:
, , [ , ) | ( )| (O3)

Function ( ) of real variable t is called original when it satisfies the above conditions.
See [2].

In Example 1 we considered ( ) which satisfies (O1); (O2)- if truncated

appropriately and (O3):
| ( )| , where , , for [ , )
and its transform was well defined for .

The function considered in Example 2 does not satisfy (O3) as t exceeds for large t,
for any given if t is large enough.
Conditions (O1), (O2) and (O3) sufficiency is guaranteed by the following theorem.

9
 000971-0007

Theorem 1 (Convergence of the Laplace transform of originals).

If ( ) satisfies (O1), (O2) and (O3) then ( ( )) ( ) is well defined (the
improper integral defining the transform is convergent) for .

Proof of Theorem 1.
Let’s consider

∫ | ( )| ∫| ( ) | .

a ing ( ) | ( ) | and using ( )we have

( ) | ( ) | | ( )| ( s) ( ).

Moreover, for s, we have s and we obtain

∫ ( ) ∫ ( s) ∫ ( s) [ ( s) ]|
s

[ ( s) ] [ ] ,
s s

so it is a convergent integral for s.

ence by Comparison est for improper integrals

∫| ( ) | is convergent,

thus

∫ ( ) dt is absolutely convergent ,

so it is also convergent from bsolute Convergence heorem.

10
 000971-0007

Example 3.
Let’s consider one more simple example to demonstrate Theorem 1 in action:
,
( ) { , truncated version of ( ) . This function satisfies (O1), (O2)
,
and (O3) with e.g.: , , so ( ) is well defined for .
Indeed,

( t) ∫ t st ∫ t( s) [ t( s) ]|
s

( s) ( )
s
[ ] s
( ) ,

for .

Example 4.
Let’s consider a more complicated example ( ) ( ), where , truncated
according to (O2). This function also satisfies (O1)-(O3) with e.g.: , , so
( ) is well defined for , although the derivation of its formula is quite
complicated.
st
First, the indefinite integral ( ) ( ) e will be calculated:
st
( ) st | |=
∫ s st

st
st s ∫ st | |
s st

st ( st ∫ st )

st ∫ st

st ( ) ,

so

( ) st ( ) ,

which yields
st ( )
,
( )

and finally
st ( )
∫ ( ) C .

11
 000971-0007

To calculate ( ) we will use a tric

( ) ∫ ( ) st ∑ , where ∫ ( ) st .
( )

We get

∫ ( ) st [ st ( )]|

[ ] ( ) ,

∫ ( ) st [ st ( )]|

[ ] ( ),

and, generally

∫ ( ) st [ st ( )]|
( ) ( )

( )
( )
[ ] ( ).

Having obtained general formula of , we are able to evaluate sum of the following
series:

∑ ∑ ( )

(( ) ( ) ( )

( )
( ))

( ) ( ) ,

for .

Finally we arrive with

( ) ( ) for .

12
 000971-0007

More results of a Laplace transforms of some basic functions are shown in a table
below. See [2]

Function Transform

( )

( )

( )
√
√

Table 1. Functions and their Laplace transforms.

Using the above table one can easily transform functions as it is presented in the next
example.

Example 5.
Let’s consider

( ( )) ( t ) , as stated in the able .

( )

A direct consequence of the definition of the Laplace transform as integral is its

linearity7.
For ( ) and ( ) satisfying O1, O2 and O3, then:
( ( ) ( )) ( ( )) ( ( )) (additivity)
( ( )) ( ( )) (homogeneity)

7 Due to linearity of integrals.

13
 000971-0007

Basing on linearity of the Laplace transform we can more effectively derive the Laplace
transforms of linear combinations of basic functions. This is demonstrated in the
following example.

Example 6.
Let’s consider ( ) ( ) t , so
( ( )) ( ( ) t ) ( ( )) ( t) ( )

We end this chapter with a theorem, which plays a crucial role in applying Laplace
transform techniques in solving differential equation. Theorem resolves the question of
the uniqueness of Laplace transforms.

Theorem 2 (Uniqueness of Laplace transform).

Distinct continuous functions on [ , ) have distinct Laplace transforms. See [6].

he result is nown as Lerch’s theorem, it means

( ( )) ( ( )) ( ) ( ).

14
 000971-0007

IV. The Laplace transform of derivatives.

The Laplace transform of derivative plays a key role, when it comes to differential
equations of any kind. In order to proceed with Laplace transform of derivatives some
crucial lemma is needed.

Lemma 1.
If ( ) satisfies ( ), ( ) and ( ) and then ( ( ) st ) .

Proof of Lemma 1.
From (O3)
| ( )| ,
so
( ) ,
which gives
st ( ) st st ,
and
( s) ( ) st ( s) .

We know that s (since ) so, ( s) .

Therefore ( s) , therefore both boundaries (lower and upper) converge:

( ( s) ) ,

( ( s) ) ,

so, by squee e theorem, ( ) st .

Now let us use Lemma 1 to derive the Laplace transform of first-order derivative of an
original.

Lemma 2 (Laplace transform of first derivative).

Let ( ) be differentiable and satisfy (O1), (O2) and (O3) and ( ) exist, then

( ( )) ( ( )) ( ).

15
 000971-0007

Proof of Lemma 2.
( ) st
( ( )) ∫ ( ) st ∫ ( ) st |
( ) s st |

([ ( ) st ]| ∫ ( ) st )

([ ( ) s ( ) ] ∫ ( ) st )

( ) ∫ ( ) ( ( )) ( ).

Example 7.
Let ( ) and ( ( )) ( ), then
( ( ) ( ) )
( ( )) ( ( )) ( )

( ) ( ) ( )

( ) ( ) ( ) .

Having calculated the Laplace transform of the first derivative we may use it to derive
Laplace transform of the second derivative.

Lemma 3 (Laplace transform of second derivative).

Let ( ) satisfy (O1), (O2) and (O3) and ( ) be differentiable twice, then
( ( )) ( ( )) ( ) ( ).

Proof of Lemma 3.
By Lemma 2 we get

( ( )) (( ( )) ) ( ( )) ( )

( ( ( )) ( )) ( ) ( ( )) ( ) ( ).

Lemmas 2 and 3 reveal a pattern and one could conjecture general formula for the
Laplace transform of nth-order derivative that we will prove by means of mathematical
induction.

16
 000971-0007

Theorem 3 (Laplace transform of the nth-order derivative).

If ( ) is differentiable n times and satisfies (O1), (O2) and (O3), then

( ) ( )
( ( )) ( ( )) ∑ ( ). ( )

Proof of Theorem 3.
Mathematical induction over will be used. Let’s introduce proposition ( )

( ) ( )
( ) ( ( )) ( ( )) ∑ ( ).

I. For
( )
( ) ( ( )) ( ( )) ( ( )) ( ) ( ) ,

by Lemma ( ) is true.

II. We will prove that for any given

( ) ( )
( ) ( ( )) ( ( )) ∑ ( ),

( ) ( )
( ) ( ( )) ( ( )) ∑ ( ) .

roof of

( )
( ) ( ( )) ([ ( )]( ) )

( ( )) ∑ ( )( )
( )

()
( ( ( )) ( )) ∑ ( )

()
( ( )) ( ) ∑ ( )

()
( ( )) ∑ ( ) |changing counter |

( ) ( )
( ( )) ∑ ( ) ( )

17
 000971-0007

so ( ) implies ( ) is true .
From I, II and Mathematical Induction Principle ( ) is true for any given .

18
 000971-0007

V. The Laplace transform and its application to solving

linear differential equations.

Theorem 2 and Theorem 3 open up the possibility to deal with the Laplace transform as
a tool for solving linear differential equations.

General form of a linear differential equation of the -th order is

( ) ( )
( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ),
where ( ), ( ) are functions of t . See [2].

An example of differential equation is

( ) ( ) √ ( ) ( ) ,
which is an example of a linear differential equation of third order.

Example 9.
This example demonstrates how the Laplace transform is used to solve differential
equations.
For instance, let’s consider the following equation
( ) ( ) , nowing that ( ) .
As Laplace transforms of equal functions are equal
( ( ) ( )) ( ).
From linearity we get
( ( )) ( ( )) ( ).
From Table 1
( ( )) ( ( )) ( )
.

Now, by means of Theorem 3

( ) ( ) ( ) ( )
.

We know that f( ) , so
( )( ) ,
( )

hence
( ) ( )
,

19
 000971-0007

and again from Table 1

( ( )) ( ),
by Theorem 2 we get
( ) .
The previously mentioned Theorem 2 guarantees that the solution is unique on [ , ).

Example 10
Let’s consider following equation
( ) ( ) ( ) ,
with initial conditions ( ) , and ( ) .
We have
( ( ) ( ) ( )) ( ),
so
( ( )) ( ( )) ( ( )) ( ) ( ),
then
( ( )) ( ( )) ( ( )) ,

and we get
( ( )) ( ) ( ) ( ( )) ( ) ( ( )) .

We know that ( ) , and ( ) , so

( ( )) ( ( )) ( ( )) ,

then
( ( )) [ ] ,

which yields
( ( )) ( )
.

The result needs to be simplified into partial fractions in order to identify the original
functions. So

( ) ( ) ( ) ( )

( )( ) ( )
( ) ( )

( ) ( ) ( )
( )
,

20
 000971-0007

which leads to
,
,
{
,
.
Solving the above we get
A=3, B=1, C=–3, D=2 .
Therefore the fraction can be written as follows

,
( ) ( )

hence

( ( )) ( )
( ( )
)

( ) ( ) ( ) (( )=
)

( ) ,
so, by Theorem 2
( ) .

21
 000971-0007

VI. Laplace transform and mice problem

Getting back to the concept of pursuit curve and three-mouse problem, now, thanks
to Laplace transform we may solve the system of differential equations and derive
formulas depicting position and movement of mice on a xy-plane in time. See Figure 1.

Getting back to (2)

√
( ) ( ),

√
( ) ( ).
{
These conditions reflect instantaneous chec ing ‘targets’ positions and adjusting mice’s
own positions instantly while running.
√
For simplicity and legibility of notation let’s put , and ( ), ( )for

( ), ( ). This gives

( ) ( ), ( )
{
( ) ( ), ( )

with initial conditions ( ) and ( ) .

After taking the Laplace transform of both sides of the equations we arrive with
( ) ( ( )) ( ( )) ( ) ( ),

( ) ( ) ( ( )) ( ( )) ( ) ( ).

Also since they are interrelated they create a system of equations:

( ) ( ) ( ) ( ) ,
{
( ) ( ) ( ) ( ) ,

and also
( ) ( ) ( ) ( ) ,
{
( ) ( ) ( ) ( ) .

Adding the equations eliminates ( ) and we get

( ) ( ) ( ) ( ) ,

22
 000971-0007

which gives

( ) .
( ( ) )

sing ( ) ( ) ( ) ( ) and the above solution we can derive

( ) ( )
( ) .
( ( ) )
Now

() ,
( ( ) )
{
() .
( ( ) )

Partial fractions

( )
( ( ) ) ( )
(( ) )
( ( ) )
( ) ( ) ( )
.
( ( ) )
It leads to
,
{ ,
( ) ,
and

,
( )
,

{ .

Hence
( )
( )
( ) ( )
( )
( ) ( ) ( )
( )
( ) ( ) ( )
( )
( ) ( ) ( )
( )

23
 000971-0007

( ) ( ) ( )

( cos sin ),

so

( ) cos sin . ( )

For given A and B

√ √ √
( ) .

The solution for ( ) may be derived from the previous equation

( ) ( ),

which is equivalent to

( )
( ) .

Derivative of ( ) equals as follows

( ),

hence
( cos sin ) ( )
( )

cos sin ( )

( ) cos ( ) sin
,

so

( ) . ( )

For given A and B

√ √ √ √
( ) .

The trajectory of the movement of the first mouse starting from ( , ) of a equilateral
triangle with a side length 1 at time is determined by the given formula

24
 000971-0007

√ √ √ √ √ √ √
( ) ( ( ), ( )) ( ( ), ( )).

Figure 2 – The trajectory of movement of mouse no. 1.

The trajectory of movement of the other two mice can be derived by rotating the
trajectory of movement of first mouse ( , ) as follows

( ) rotated around ( , ) and translated by ⃗⃗⃗⃗⃗⃗⃗⃗ gives ( ),

( ) rotated around ( , ) and translated by ⃗⃗⃗⃗⃗⃗⃗⃗ gives ( ),

this leads to

⃗⃗⃗⃗⃗⃗⃗⃗⃗ ( ) ( )( ) ( )

√ √ √ √
( )
( )( ) ( )
√ √ √ √ √
( )

√ √ √
( )
( )
√ √ √ √
( ) .

We get
√ √ √ √ √ √ √
( ) ( ( ), ( )) ( ( ), ( )).

25
 000971-0007

Figure 3 – The trajectory of movement of mouse no. 2.

And the third mouse

( ) ( )( ) ( )

√ √ √ √
( )
( )( ) ( )
√ √ √ √ √
( )

√ √
( )
( ),
√
(√ √
)

and finally
√ √ √ √ √
( ) ( ( ), ( )) ( ( ), ( )).

26
 000971-0007

Figure 4 – The trajectory of movement of mouse no. 3.

We have derived the parametric formulas that determine position of each mice in time:
√ √ √ √ √ √ √
( ) ( ( ), ( )) ( ( ), ( )) ,

√ √ √ √ √ √ √
( ) ( ( ), ( )) ( ( ), ( )) ,

√ √ √ √ √
( ) ( ( ), ( )) ( ( ), ( )) .

27
 000971-0007

Figure 5 – The trajectories of all three mice.

To generalize solution, let’s consider regular n-polygon where ( , ) and

( , ) and let be the centre of a circumscribed circle (see Appendix B). Using
notation introduced previously we can observe that
( ) rotated around ( , ) and translated by ⃗⃗⃗⃗⃗⃗⃗⃗ gives ( ),

which leads to

⃗⃗⃗⃗⃗⃗⃗⃗⃗ ( ) ( )( ) ⃗⃗⃗⃗⃗⃗⃗⃗

( ) ( )

( )
.
so
( ) ( ( ) ( ) , ( ) ( ) ).

herefore
( ) ( ) ( ) ( ) ( )
.
( ) ( ) ( ) ( ) ( )

28
 000971-0007

Moreover we can notice that

( ) ( )( )
,
( )( ) ( )

so, we get

( ) ( )( ), ( )

( )( ) ( ) . ( )
{

For simplicity and legibility of notation let’s put , and

( ), ( ) for ( ), ( ). This gives

( ) ( ),
{
( ) ( ),

with initial conditions ( ) and ( ) , which is the same as ( ), ( ). So we

have ( ) and ( ) again:

( ) cos sin ,

( ) .

But for given A and B these simplify now to

( )
( ) ( (( ) ) (( ) )) ,

( )
( ) ( (( ) ) (( ) )),

in a more compact form

( )
( ) ( ),

( )
( ) ( ).

Reversing our notation simplification we get parametric equations of the position of

mouse no. 1 for in regular n-polygon

( )
( ) ( ),

( )
( ) ( ).

29
 000971-0007

Now, we will derive equations of the position of the remaining mice. Let’s focus on
mouse no. k, ( ). We know that
( ) rotated around ( , ) and translated by ⃗⃗⃗⃗⃗⃗⃗⃗ gives ( ),

( ) rotated around ( , ) and translated by ⃗⃗⃗⃗⃗⃗⃗⃗ gives ( ),

and so on, generally

( ) rotated ( ) around ( , ) and translated by ⃗⃗⃗⃗⃗⃗⃗⃗⃗ gives ( ).

Therefore (in matrix-vector notation - see Appendix)

( ) ( )
⃗⃗⃗⃗⃗⃗⃗⃗⃗ ( ) ( ( ) ( ) )( ) ⃗⃗⃗⃗⃗⃗⃗⃗⃗

( ) ( )

( ( ) ( ) ) ⃗⃗⃗⃗⃗⃗⃗⃗⃗ .

Since ( , ) is and ⃗⃗⃗⃗⃗⃗⃗⃗⃗ ⃗⃗⃗⃗⃗⃗⃗ , we know that (see Appendix part 2)

( ) ( )
⃗⃗⃗⃗⃗⃗⃗⃗⃗ ( ( ) ( ) ),

so we get
( ) ( ) ( ) ( )
⃗⃗⃗⃗⃗⃗⃗⃗⃗ ( ).
( ) ( ) ( ) ( )

Substitution of ( ) and ( ) (and simplification8) leads to

( ) ( )
( ) ( )
( ( ) ( ) ) ,

( ) ( )
( )
( ) ( ( ) ( ) ) ,

which may be expressed in more compact form

( ) ( )
( ) ( ),

( ) ( )
( ) ( ).
The above equations describe position of mouse no. k in a regular n-polygon at time
, if we place the n-polygon in such a way that ( , ) and ( , ).

Graphs of solution of other cases n=5 and n=10 are presented below.

8 Simplified with a use of CAS (Computer Algebra System) software (Derive 5.04).

30
 000971-0007

Figure 6 – Five mice pursuit problem trajectories.

Figure 7 – Ten mice pursuit problem trajectories.

31
 000971-0007

VII. Conclusion

Mice pursuing problem, considered firstly in a regular triangle (as a simplest

case of regular polygon), led us to a system of simultaneous differential equations.
Thanks to a symmetry of the mice behaviour these were first order linear differential
equations for ( )and ( ) – parametric equations of the mouse no. 1 trajectory.
An introduction to Laplace transform and some of its properties allowed us to
solve the system of equations in this case. The key property was presented and proved
in a Theorem 3 which, when combined with linearity of Laplace transform guarantees
that Laplace transform maps linear differential equations (systems of linear differential
equations) into algebraic equations (systems of algebraic linear equations,
respectively). Moreover equations of first order are mapped into linear algebraic
equations, so in our case we obtained a system of two linear algebraic equations,
solution of which gave us Laplace transforms of our unknown functions ( ) and ( ).
Partial fraction decomposition of the results confronted with the table of Laplace
transforms (Table 1) enabled us to derive explicit formulas for ( ) and ( ). The
(Theorem 2) Lerch’s theorem (another key property for us) guaranteed uniqueness of
the solutions. The symmetry of the problem allowed us to produce formulas for ( ),
( ) and ( ), ( ).
Restating the problem in a general form (regular n-polygon case) did not change
the general pattern of solving process (than s to “elegance” of Laplace transform)
although the calculations involved were more complicated.
The mice trajectories turned out to be logarithmic spirals beginning at the
polygon’s vertices and converging to its centre9. A number of interesting questions
arises now (for example what are the lengths of the trajectories, what happens when n
goes to infinity) and some possible improvements are apparent - the centre of polygon
could be placed in O and one could use constant size polygon (e.g. R=1=radius of a
circumscribed circle) and complex numbers for parameterization (vertices would be
roots of 1 and rotation would be just multiplication).

9 A centre of an inscribed and circumscribed circle, mass centre.

32
 000971-0007
VIII. Bibliography

1. Beerends, R. 2003. Fourier and Laplace transforms. Cambridge: Cambridge university

press.

2. Gewert, M. and S oc ylas Z. ( ). ównanie różnic owe wyc ajne eoria, pr y łady,
zadania, rocław ficyna ydawnic a GiS

3. Larson, R. and Edwards, B. 2014. Calculus. Boston, MA: Brooks/Cole, Cengage Learning.

4. Liu, V. 2005. The four mice pursuit curve solution using Laplace Transforms. [e-book]
http://www.stanford.edu/~vkl/documents/4mice.pdf [Accessed: 30 September 2013].

5. Kąc i, E. and Siewiers i L. ( ). ybrane d iały matematy i wyżs ej ćwic enimi,

Warszawa: Wydawnictwi Naukowe PWN

6. Orloff, J. 2013. 18.03 Uniqueness of Laplace Transform. [e-book] B: Available through:

http://web.mit.edu/jorloff/www/18.03-esg/notes/laplaceuniqueness.pdf [Accessed: 20
October 2013].

7. Stewart, J., (2012). Calculus. 7th ed. Belmonts: Brooks/Cole Pub Co.

8. Shiff, L. J. (1999). The Laplace Transform: Theory and Applications. [e-book] New York:
Springer-Verlag Available at: Google Books <http://books.google.com>[Accessed: 20
September 2013].

9. Strang, G. n.d. Calculus. [e-book] Available through:

http://ocw.mit.edu/ans7870/resources/Strang/Edited/Calculus/Calculus.pdf [Accessed:
10 September 2013].

33
 000971-0007
IX. Appendix

Analytical forms of geometrical transforms used (in this paper):

A1
- rotation about ( , ) by directed angle :
( , ) ( , ) ( , )
in position rotation:

( ) ( ) ( )( ) ( )

where is the rotation matrix.

A2

- translation by vector ( )

⃗( , ) ( , ) ( , )
in position vector rotation:

( ) ( ) ( )

B
Derivation of the coordinates of – centre of circumscribed circle of the regular
n-polygon. Let’s consider following regular polygon

From the above figure it may be concluded that

, which leads to .

Therefore, ( , ) is a centre of an inscribed and circumscribed circle in every

regular n-polygon, if we place the n-polygon in such a way that ( , ) and
( , ).

34