5 views

Uploaded by memrah2955

- 2013FA-MATH-2420-81001
- Fundamentals of Ordinary Differential Equations
- Gate 2018 Syllabus
- B. Arch. (I - X).pdf
- using ode45
- L6 IA 2011 U1 M1
- Indivi_syllabi_I sem.docx
- 2nd
- Theory of ODE - Hu
- Equations
- 4-1.Differential Equations
- CIV340 Mock Exam
- EE203_04_ode3_jan13_14s
- 02_6243 new B__pp 15-23 new2
- Linear Algebra and Ordinary Differential Equations (Riaz) (1)
- Math 311 Numerical Solutions of Ordinary Differential Equations.doc
- Differential Equations - Ordinary Differential Equations - Integrating Factors
- It Curriculum Book
- Project 309
- Matematici speciale. Probleme

You are on page 1of 174

MATB44H3F

Version September 15, 2011-1949

ii

Contents

1 Introduction 1

1.1 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 Sample Application of Dierential Equations . . . . . . . . . . . 2

2 First Order Ordinary Dierential Equations 5

2.1 Separable Equations . . . . . . . . . . . . . . . . . . . . . . . . . 5

2.2 Exact Dierential Equations . . . . . . . . . . . . . . . . . . . . . 7

2.3 Integrating Factors . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2.4 Linear First Order Equations . . . . . . . . . . . . . . . . . . . . 14

2.5 Substitutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

2.5.1 Bernoulli Equation . . . . . . . . . . . . . . . . . . . . . . 17

2.5.2 Homogeneous Equations . . . . . . . . . . . . . . . . . . . 19

2.5.3 Substitution to Reduce Second Order Equations to First

Order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3 Applications and Examples of First Order odes 25

3.1 Orthogonal Trajectories . . . . . . . . . . . . . . . . . . . . . . . 25

3.2 Exponential Growth and Decay . . . . . . . . . . . . . . . . . . . 27

3.3 Population Growth . . . . . . . . . . . . . . . . . . . . . . . . . . 28

3.4 Predator-Prey Models . . . . . . . . . . . . . . . . . . . . . . . . 29

3.5 Newtons Law of Cooling . . . . . . . . . . . . . . . . . . . . . . 30

3.6 Water Tanks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

3.7 Motion of Objects Falling Under Gravity with Air Resistance . . 34

3.8 Escape Velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

3.9 Planetary Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

3.10 Particle Moving on a Curve . . . . . . . . . . . . . . . . . . . . . 39

iii

iv CONTENTS

4 Linear Dierential Equations 45

4.1 Homogeneous Linear Equations . . . . . . . . . . . . . . . . . . . 47

4.1.1 Linear Dierential Equations with Constant Coecients . 52

4.2 Nonhomogeneous Linear Equations . . . . . . . . . . . . . . . . . 54

5 Second Order Linear Equations 57

5.1 Reduction of Order . . . . . . . . . . . . . . . . . . . . . . . . . . 57

5.2 Undetermined Coecients . . . . . . . . . . . . . . . . . . . . . . 60

5.2.1 Shortcuts for Undetermined Coecients . . . . . . . . . . 64

5.3 Variation of Parameters . . . . . . . . . . . . . . . . . . . . . . . 66

6 Applications of Second Order Dierential Equations 71

6.1 Motion of Object Hanging from a Spring . . . . . . . . . . . . . . 71

6.2 Electrical Circuits . . . . . . . . . . . . . . . . . . . . . . . . . . 75

7 Higher Order Linear Dierential Equations 79

7.1 Undetermined Coecients . . . . . . . . . . . . . . . . . . . . . . 79

7.2 Variation of Parameters . . . . . . . . . . . . . . . . . . . . . . . 80

7.3 Substitutions: Eulers Equation . . . . . . . . . . . . . . . . . . . 82

8 Power Series Solutions to Linear Dierential Equations 85

8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

8.2 Background Knowledge Concerning Power Series . . . . . . . . . 88

8.3 Analytic Equations . . . . . . . . . . . . . . . . . . . . . . . . . . 89

8.4 Power Series Solutions: Levels of Success . . . . . . . . . . . . . . 91

8.5 Level 1: Finding a nite number of coecients . . . . . . . . . . 91

8.6 Level 2: Finding the recursion relation . . . . . . . . . . . . . . . 94

8.7 Solutions Near a Singular Point . . . . . . . . . . . . . . . . . . . 97

8.8 Functions Dened via Dierential Equations . . . . . . . . . . . . 111

8.8.1 Chebyshev Equation . . . . . . . . . . . . . . . . . . . . . 111

8.8.2 Legendre Equation . . . . . . . . . . . . . . . . . . . . . . 113

8.8.3 Airy Equation . . . . . . . . . . . . . . . . . . . . . . . . 115

8.8.4 Laguerres Equation . . . . . . . . . . . . . . . . . . . . . 115

8.8.5 Bessel Equation . . . . . . . . . . . . . . . . . . . . . . . . 116

9 Linear Systems 121

9.1 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

9.2 Computing e

T

. . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

9.3 The 2 2 Case in Detail . . . . . . . . . . . . . . . . . . . . . . . 129

9.4 The Non-Homogeneous Case . . . . . . . . . . . . . . . . . . . . 133

CONTENTS v

9.5 Phase Portraits . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

9.5.1 Real Distinct Eigenvalues . . . . . . . . . . . . . . . . . . 137

9.5.2 Complex Eigenvalues . . . . . . . . . . . . . . . . . . . . . 139

9.5.3 Repeated Real Roots . . . . . . . . . . . . . . . . . . . . . 141

10 Existence and Uniqueness Theorems 145

10.1 Picards Method . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

10.2 Existence and Uniqueness Theorem for First Order ODEs . . . . 150

10.3 Existence and Uniqueness Theorem for Linear First Order ODEs 155

10.4 Existence and Uniqueness Theorem for Linear Systems . . . . . . 156

11 Numerical Approximations 163

11.1 Eulers Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

11.1.1 Error Bounds . . . . . . . . . . . . . . . . . . . . . . . . . 165

11.2 Improved Eulers Method . . . . . . . . . . . . . . . . . . . . . . 166

11.3 Runge-Kutta Methods . . . . . . . . . . . . . . . . . . . . . . . . 167

vi CONTENTS

Chapter 1

Introduction

1.1 Preliminaries

Denition (Dierential equation)

A dierential equation (de) is an equation involving a function and its deriva-

tives.

Dierential equations are called partial dierential equations (pde) or or-

dinary dierential equations (ode) according to whether or not they contain

partial derivatives. The order of a dierential equation is the highest order

derivative occurring. A solution (or particular solution) of a dierential equa-

tion of order n consists of a function dened and n times dierentiable on a

domain D having the property that the functional equation obtained by substi-

tuting the function and its n derivatives into the dierential equation holds for

every point in D.

Example 1.1. An example of a dierential equation of order 4, 2, and 1 is

given respectively by

_

dy

dx

_

3

+

d

4

y

dx

4

+y = 2 sin(x) + cos

3

(x),

2

z

x

2

+

2

z

y

2

= 0,

yy

= 1.

1

2 CHAPTER 1. INTRODUCTION

Example 1.2. The function y = sin(x) is a solution of

_

dy

dx

_

3

+

d

4

y

dx

4

+y = 2 sin(x) + cos

3

(x)

on domain R; the function z = e

x

cos(y) is a solution of

2

z

x

2

+

2

z

y

2

= 0

on domain R

2

; the function y = 2

x is a solution of

yy

= 2

on domain (0, ).

Although it is possible for a de to have a unique solution, e.g., y = 0 is the

solution to (y

)

2

+ y

2

= 0, or no solution at all, e.g., (y

)

2

+ y

2

= 1 has no

solution, most des have innitely many solutions.

Example 1.3. The function y =

4x +C on domain (C/4, ) is a solution

of yy

Note that dierent solutions can have dierent domains. The set of all

solutions to a de is call its general solution.

1.2 Sample Application of Dierential Equations

A typical application of dierential equations proceeds along these lines:

Real World Situation

Mathematical Model

Interpretation of Solution

1.2. SAMPLE APPLICATION OF DIFFERENTIAL EQUATIONS 3

Sometimes in attempting to solve a de, we might perform an irreversible

step. This might introduce extra solutions. If we can get a short list which

contains all solutions, we can then test out each one and throw out the invalid

ones. The ultimate test is this: does it satisfy the equation?

Here is a sample application of dierential equations.

Example 1.4. The half-life of radium is 1600 years, i.e., it takes 1600 years for

half of any quantity to decay. If a sample initially contains 50 g, how long will

it be until it contains 45 g?

Solution. Let x(t) be the amount of radium present at time t in years. The rate

at which the sample decays is proportional to the size of the sample at that

time. Therefore we know that dx/dt = kx. This dierential equation is our

mathematical model. Using techniques we will study in this course (see 3.2,

Chapter 3), we will discover that the general solution of this equation is given

by the equation x = Ae

kt

, for some constant A. We are told that x = 50 when

t = 0 and so substituting gives A = 50. Thus x = 50e

kt

. Solving for t gives

t = ln(x/50) /k. With x(1600) = 25, we have 25 = 50e

1600k

. Therefore,

1600k = ln

_

1

2

_

= ln(2) ,

giving us k = ln(2) /1600. When x = 45, we have

t =

ln(x/50)

k

=

ln(45/50)

ln(2) /1600

= 1600

ln(8/10)

ln(2)

= 1600

ln(10/8)

ln(2)

1600

0.105

0.693

1600 0.152 243.2.

Therefore, it will be approximately 243.2 years until the sample contains 45 g

of radium.

Additional conditions required of the solution (x(0) = 50 in the above ex-

ample) are called boundary conditions and a dierential equation together with

boundary conditions is called a boundary-value problem (BVP). Boundary con-

ditions come in many forms. For example, y(6) = y(22); y

are all examples of boundary conditions. Boundary-value problems, like the one

in the example, where the boundary condition consists of specifying the value

of the solution at some point are also called initial-value problems (IVP).

Example 1.5. An analogy from algebra is the equation

y =

y + 2. (1.1)

4 CHAPTER 1. INTRODUCTION

To solve for y, we proceed as

y 2 =

y,

(y 2)

2

= y, (irreversible step)

y

2

4y + 4 = y,

y

2

5y + 4 = 0,

(y 1) (y 4) = 0.

Thus, the set y 1, 4 contains all the solutions. We quickly see that y = 4

satises Equation (1.1) because

4 =

4 + 2 = 4 = 2 + 2 = 4 = 4,

while y = 1 does not because

1 =

1 + 2 = 1 = 3.

So we accept y = 4 and reject y = 1.

Chapter 2

First Order Ordinary

Dierential Equations

The complexity of solving des increases with the order. We begin with rst

order des.

2.1 Separable Equations

A rst order ode has the form F(x, y, y

of algebra can be used to write it in the form

be factored to give G(x, y) = M(x) N(y),then the equation is called separable.

To solve the separable equation y

f(y)y

_

f(y)y

dx =

_

g(x) dx,

_

f(y) dy =

_

f(y)

dy

dx

dx.

Example 2.1. Solve 2xy + 6x +

_

x

2

4

_

y

= 0.

5

6CHAPTER 2. FIRST ORDER ORDINARY DIFFERENTIAL EQUATIONS

Solution. Rearranging, we have

_

x

2

4

_

y

= 2xy 6x,

= 2xy 6x,

y

y + 3

=

2x

x

2

4

, x ,= 2

ln([y + 3[) = ln

_

x

2

4

_

+C,

ln([y + 3[) + ln

_

x

2

4

_

= C,

where C is an arbitrary constant. Then

(y + 3)

_

x

2

4

_

= A,

(y + 3)

_

x

2

4

_

= A,

y + 3 =

A

x

2

4

,

where A is a constant (equal to e

C

) and x ,= 2. Also y = 3 is a solution

(corresponding to A = 0) and the domain for that solution is R.

Example 2.2. Solve the ivp sin(x) dx +y dy = 0, where y(0) = 1.

Solution. Note: sin(x) dx +y dy = 0 is an alternate notation meaning the same

as sin(x) +y dy/dx = 0.

We have

y dy = sin(x) dx,

_

y dy =

_

sin(x) dx,

y

2

2

= cos(x) +C

1

,

y =

_

2 cos(x) +C

2

,

where C

1

is an arbitrary constant and C

2

= 2C

1

. Considering y(0) = 1, we have

1 =

_

2 +C

2

= 1 = 2 +C

2

= C

2

= 1.

Therefore, y =

_

2 cos(x) 1 on the domain (/3, /3), since we need cos(x)

1/2 and cos(/3) = 1/2.

2.2. EXACT DIFFERENTIAL EQUATIONS 7

An alternate method to solving the problem is

y dy = sin(x) dx,

_

y

1

y dy =

_

x

0

sin(x) dx,

y

2

2

1

2

2

= cos(x) cos(0),

y

2

2

1

2

= cos(x) 1,

y

2

2

= cos(x)

1

2

,

y =

_

2 cos(x) 1,

giving us the same result as with the rst method.

Example 2.3. Solve y

4

y

+y

+x

2

+ 1 = 0.

Solution. We have

_

y

4

+ 1

_

y

= x

2

1,

y

5

5

+y =

x

3

3

x +C,

where C is an arbitrary constant. This is an implicit solution which we cannot

easily solve explicitly for y in terms of x.

2.2 Exact Dierential Equations

Using algebra, any rst order equation can be written in the form F(x, y) dx +

G(x, y) dy = 0 for some functions F(x, y), G(x, y).

Denition

An expression of the form F(x, y) dx +G(x, y) dy is called a (rst-order) dier-

ential form. A dierentical form F(x, y) dx + G(x, y) dy is called exact if there

exists a function g(x, y) such that dg = F dx +Gdy.

If = F dx+Gdy is an exact dierential form, then = 0 is called an exact

dierential equation. Its solution is g = C, where = dg.

Recall the following useful theorem from MATB42:

8CHAPTER 2. FIRST ORDER ORDINARY DIFFERENTIAL EQUATIONS

Theorem 2.4

If F and G are functions that are continuously dierentiable throughout a

simply connected region, then F dx +Gdy is exact if and only if G/x =

F/y.

Proof. Proof is given in MATB42.

Example 2.5. Consider

_

3x

2

y

2

+x

2

_

dx +

_

2x

3

y +y

2

_

dy = 0. Let

=

_

3x

2

y

2

+x

2

_

. .

F

dx +

_

2x

3

y +y

2

_

. .

G

dy

Then note that

G

x

= 6x

2

y =

F

y

.

By Theorem 2.4, = dg for some g. To nd g, we know that

g

x

= 3x

2

y

2

+x

2

, (2.1a)

g

y

= 2x

3

y +y

2

. (2.1b)

Integrating Equation (2.1a) with respect to x gives us

g = x

3

y

2

+

x

3

3

+h(y). (2.2)

So dierentiating that with respect to y gives us

Eq. (2.1b)

..

g

y

= 2x

3

y +

dh

dy

,

2x

3

y +y

2

= 2x

3

y +

dh

dy

,

dh

dy

= y

2

,

h(y) =

y

3

3

+C

2.2. EXACT DIFFERENTIAL EQUATIONS 9

for some arbitrary constant C. Therefore, Equation (2.2) becomes

g = x

3

y

2

+

x

3

3

+

y

3

3

+C.

Note that according to our dierential equation, we have

d

_

x

3

y

2

+

x

3

3

+

y

3

3

+C

_

= 0 which implies x

3

y

2

+

x

3

3

+

y

3

3

+C = C

. Letting D = C

constant, the solution is

x

3

y

2

+

x

3

3

+

y

3

3

= D.

_

3x

2

+ 2xy

2

_

dx +

_

2x

2

y

_

dy = 0, where y(2) = 3.

Solution. We have

_

_

3x

2

+ 2xy

2

_

dx = x

3

+x

2

y

2

+C

for some arbitrary constant C. Since C is arbitrary, we equivalently have x

3

+

x

2

y

2

= C. With the initial condition in mind, we have

8 + 4 9 = C = C = 44.

Therefore, x

3

+x

2

y

2

= 44 and it follows that

y =

44 x

3

x

2

.

But with the restriction that y(2) = 3, the only solution is

y =

44 x

3

x

2

on the domain

_

44,

3

44

_

0.

Let = F dx + Gdy. Let y = s(x) be the solution of the de = 0, i.e.,

F +Gs

(x) = 0. Let y

0

= s(x

0

) and let be the piece of the graph of y = s(x)

from (x

0

, y

0

) to (x, y). Figure 2.1 shows this idea. Since y = s(x) is a solution

to = 0, we must have = 0 along . Therefore,

_

10CHAPTER 2. FIRST ORDER ORDINARY DIFFERENTIAL EQUATIONS

(x

0

,y)

1

(x

0

,y

0

)

(x,y)

y=s(x)

y

x

Figure 2.1: The graph of y = s(x) with connecting (x

0

, y

0

) to (x, y).

by parameterizing by (x) = (x, s(x)), thereby giving us

_

=

_

x

x0

F dx +Gs

(x) dx =

_

x

x0

0 dx = 0.

This much holds for any .

Now suppose that is exact. Then the integral is independent of the path.

Therefore

0 =

_

=

_

1

F dx +Gdy +

_

2

F dx +Gdy

=

_

y

y0

G(x

0

, y) dy +

_

x

x0

F(x, y) dx.

We can now solve Example 2.6 with this new method.

Solution (Alternate solution to Example 2.6). We simply have

0 =

_

4

3

2 2

2

y dy +

_

x

2

_

3x

2

+ 2xy

2

_

dx

= 4y

2

4 (3)

2

+x

3

+x

2

y

2

2

3

2

2

y

2

= 4y

2

36 +x

3

+x

2

y

2

8 4y

2

,

nally giving us x

3

+x

2

y

2

= 44, which agrees with our previous answer.

Remark. Separable equations are actually a special case of exact equations, that

is,

f(y)y

x

f(y) = 0 =

y

(g(x)) .

2.3. INTEGRATING FACTORS 11

So the equation is exact.

2.3 Integrating Factors

Consider the equation = 0. Even if is not exact, there may be a function

I(x, y) such that I is exact. So = 0 can be solved by multiplying both sides

by I. The function I is called an integrating factor for the equation = 0.

Example 2.7. Solve y/x

2

+ 1 +y

/x = 0.

Solution. We have

_

y

x

2

+ 1

_

dx +

1

x

dy = 0.

We see that

_

x

_

1

x

_

=

1

x

2

_

,=

_

1

x

2

=

y

_

y

x

2

+ 1

_

_

.

So the equation is not exact. Multiplying by x

2

gives us

_

y +x

2

_

dx +xdy = 0,

d

_

xy +

x

3

3

_

= 0,

xy +

x

3

3

= C

for some arbitrary constant C. Solving for y nally gives us

y =

C

x

x

3

3

.

There is, in general, no algorithm for nding integrating factors. But the

following may suggest where to look. It is important to be able to recognize

common exact forms:

xdy +y dx = d(xy) ,

xdy y dx

x

2

= d

_

y

x

_

,

xdx +y dy

x

2

+y

2

= d

_

ln

_

x

2

+y

2

_

2

_

,

xdy d dx

x

2

+y

2

= d

_

tan

1

_

y

x

__

,

x

a1

y

b1

(ay dx +bxdy) = d

_

x

a

y

b

_

.

12CHAPTER 2. FIRST ORDER ORDINARY DIFFERENTIAL EQUATIONS

Example 2.8. Solve

_

x

2

y

2

+y

_

dx +

_

2x

3

y x

_

dy = 0.

Solution. Expanding, we have

x

2

y

2

dx + 2x

3

y dy +y dx xdy = 0.

Here, a = 1 and b = 2. Thus, we wish to use

d

_

xy

2

_

= y

2

dx + 2xy dy.

This suggests dividing the original equation by x

2

which gives

y

2

dx + 2xy dy +

y dx xdy

x

2

= 0.

Therefore,

xy

2

+

y

x

= C, x ,= 0,

where C is an arbitrary constant. Additionally, y = 0 on the domain R is a

solution to the original equation.

Example 2.9. Solve y dx xdy

_

x

2

+y

2

_

dx = 0.

Solution. We have

y dx xdy

x

2

+y

2

dx = 0,

unless x = 0 and y = 0. Now, it follows that

tan

1

_

y

x

_

x = C,

tan

1

_

y

x

_

= C x,

tan

1

_

y

x

_

= D x, (D = C)

y

x

= tan(D x) ,

y = xtan(D x) ,

where C is an arbitrary constant and the domain is

D x ,= (2n + 1)

2

, x ,= (2n + 1)

2

for any integer n. Also, since the derivation of the solution is based on the

assumption that x ,= 0, it is unclear whether or not 0 should be in the domain,

i.e., does y = xtan(D x) satisfy the equation when x = 0? We have y xy

_

x

2

+y

2

_

= 0. If x = 0 and y = xtan(D x), then y = 0 and the equation is

satised. Thus, 0 is in the domain.

Proposition 2.10

Let = dg. Then for any function P : R R, P(g) is exact.

Proof. Let Q =

_

P(t) dy. Then d(Q(g)) = P(g) dg = P(g).

To make use of Proposition 2.10, we can group together some terms of

to get an expression you recognize as having an integrating factor and multiply

the equation by that. The equation will now look like dg + h = 0. If we can

nd an integrating factor for h, it will not necessarily help, since multiplying by

it might mess up the part that is already exact. But if we can nd one of the

form P(g), then it will work.

Example 2.11. Solve

_

x yx

2

_

dy +y dx = 0.

Solution. Expanding, we have

y dx +xdy

. .

d(xy)

yx

2

dy = 0.

Therefore, we can multiply teh equation by any function of xy without disturb-

ing the exactness of its rst two terms. Making the last term into a function of

y alone will make it exact. So we multiply by (xy)

2

, giving us

y dx +xdy

x

2

y

2

1

y

dy = 0 =

1

xy

ln([y[) = C,

where C is an arbitrary constant. Note that y = 0 on the domain R is also a

solution.

Given

M dx +N dy = 0, ()

we want to nd I such that IM dx +IN dy is exact. If so, then

x

(IN)

. .

IxN+INx

=

y

(IM)

. .

IyM+IMy

.

14CHAPTER 2. FIRST ORDER ORDINARY DIFFERENTIAL EQUATIONS

If we can nd any particular solution I(x, y) of the pde

I

x

N +IN

x

= I

y

M +IM

y

, ()

then we can use it as an integrating factor to nd the general solution of ().

Unfortunately, () is usually even harder to solve than (), but as we shall see,

there are times when it is easier.

Example 2.12. We could look for an I having only xs and no ys? For exam-

ple, consider I

y

= 0. Then

I

x

N +IN

x

= IM

y

implies

I

x

I

=

M

y

N

x

N

.

This works if (M

y

N

x

) /N happens to be a function of x alone. Then

I = e

R

MyNx

N

dx

.

Similarly, we can also reverse the role of x and y. If (N

x

M

y

) /M happens to

be a function of y alone, then

e

R

NxMy

M

dy

works.

2.4 Linear First Order Equations

A rst order linear equation (n = 1) looks like

y

+P(x)y = Q(x).

An integrating factor can always be found by the following method. Consider

dy +P(x)y dx = Q(x) dx,

(P(x)y Q(x))

. .

M(x,y)

dx + dy

..

N(x,y)

= 0.

We use the de for the integrating factor I(x, y). The equation IM dx + IN dy

is exact if

I

x

N +IN

x

= I

y

M +IM

y

.

2.4. LINEAR FIRST ORDER EQUATIONS 15

In our case,

I

x

+ 0 = I

y

(P(x)y Q(x)) +IP(x). ()

We need only one solution, so we look for one of the form I(x), i.e., with I

y

= 0.

Then () becomes

dI

dx

= IP(x).

This is separable. So

dI

I

= P(x) dx,

ln([I[) =

_

P(x) dx +C,

[I[ = e

R

P(x) dx

, e

x

> 0

I = e

R

P(x) dx

.

We conclude that e

R

P(x) dx

is an integrating factor for y

+P(x)y = Q(x).

Example 2.13. Solve y

(1/x) y = x

3

, where x > 0.

Solution. Here P(x) = 1/x. Then

I = e

R

P(x) dx

= e

R

1

x

dx

= e

ln(|x|)dx

=

1

[x[

=

1

x

,

where x > 0. Our dierential equation is

xdy y dx

x

= x

3

dx.

Multiplying by the integrating factor 1/x gives us

xdy y dx

x

2

= x

2

dx.

Then

y

x

=

x

3

3

+C,

y =

x

3

3

+Cx

on the domain (0, ), where C is an arbitrary constant (x > 0 is given).

16CHAPTER 2. FIRST ORDER ORDINARY DIFFERENTIAL EQUATIONS

In general, given y

R

P(x) dx

to obtain

e

R

P(x) dx

y

+e

R

P(x) dx

. .

d(ye

R

P(x) dx

)/dx

P(x)y = Q(x)e

R

P(x) dx

.

Therefore,

ye

R

P(x) dx

=

_

Q(x)e

R

P(x) dx

dx +C,

y = e

R

P(x) dx

_

Q(x)e

R

P(x) dx

dx +Ce

R

P(x) dx

,

where C is an arbitrary constant.

Example 2.14. Solve xy

+ 2y = 4x

2

.

Solution. What should P(x) be? To nd it, we put the equation in standard

form, giving us

y

+

2

x

y = 4x.

Therefore, P(x) = 2/x. Immediately, we have

I = e

R

(2/x)dx

= e

ln(x

2

)

= x

2

.

Multiplying the equation by x

2

gives us

x

2

y

+ 2xy = 4x

3

,

x

2

y = x

4

+C,

y = x

2

+

C

x

2

,

where C is an arbitrary constant and x ,= 0.

Example 2.15. Solve e

y

dy +dx + 2xdy = 0.

Solution. This equation is linear with x as a function of y. So what we have is

dx

dy

+ 2x = e

y

,

2.5. SUBSTITUTIONS 17

where I = e

R

2 dy

= e

2y

. Therefore,

e

2y

dx

dy

+ 2xe

2y

= e

y

,

xe

2y

= e

y

+C,

where C is an arbitrary constant. We could solve explicitly for y, but it is messy.

The domain is not easy to determine.

2.5 Substitutions

In many cases, equations can be put into one of the standard forms discussed

above (separable, linear, etc.) by a substitution.

Example 2.16. Solve y

2y

= 5.

Solution. This is a rst order linear equation for y

. Let u = y

. Then the

equation becomes

u

2u = 5.

The integration factor is then I = e

R

2 dx

= e

2x

. Thus,

u

e

2x

2ue

2x

= 5e

2x

,

ue

2x

=

5

2

e

2x

+C,

where C is an arbitrary constant. But u = y

, so

y =

5

2

x +

C

2

e

2x

+C

1

=

5

2

x +C

1

e

2x

+C

2

on the domain R, where C

1

and C

2

are arbitrary constants.

We now look at standard substitutions.

2.5.1 Bernoulli Equation

The Bernoulli equation is given by

dy

dx

+P(x)y = Q(x)y

n

.

Let z = y

1n

. Then

dz

dx

= (1 n) y

n

dy

dx

,

18CHAPTER 2. FIRST ORDER ORDINARY DIFFERENTIAL EQUATIONS

giving us

y

n

dy

dx

+P(x)y

1n

= Q(x),

1

1 n

dz

dx

+P(x)z = Q(x),

dz

dx

+ (1 n) P(x)z = (1 n) Q(x),

which is linear in z.

Example 2.17. Solve y

+xy = xy

3

.

Here, we have n = 3. Let z = y

2

. If y ,= 0, then

dz

dx

= 2y

3

dy

dx

.

Therefore, our equation becomes

y

3

z

2

+xy = xy

3

,

2

+xy

2

= x,

z

2xy = 2x.

We can readily see that I = e

2x dx

= e

x

. Thus,

e

x

z

2xe

x

= 2xe

x

,

e

x

z = e

x

+C,

z = 1 +Ce

x

,

where C is an arbitrary constant. But z = y

2

. So

y =

1

1 +Ce

x

.

The domain is 8

:

R, C > 1,

[x[ >

p

ln(C), C 1.

An additional solution is y = 0 on R.

2.5. SUBSTITUTIONS 21

If it is independent of y, namely, F(y

, y

order equation for y

Consider now the case where it is independent of x, namely, F(y

, y

, y) = 0.

Substitute v = dy/dx for x, i.e., eliminate x between the equations

F

_

d

2

y

dx

2

,

dy

dx

, y

_

= 0

and v = dy/dx. Then

d

2

y

dx

2

=

dv

dx

=

dv

dy

dy

dx

=

dv

dy

v.

Therefore,

F

_

d

2

y

dx

2

,

dy

dx

, y

_

= 0 F

_

dv

dy

v, v, y

_

= 0.

This is a rst order equation in v and y.

Example 2.22. Solve y

= 4 (y

)

3/2

y.

Solution. Let v = dy/dx. Then

d

2

y

dx

2

=

dv

dy

v

and our equation becomes

dv

dy

v = 4v

3/2

y,

dv

v

= 4y dy, v 0,

2

v = 2y

2

+C

1

,

v = y

2

+C

2

,

22CHAPTER 2. FIRST ORDER ORDINARY DIFFERENTIAL EQUATIONS

where C

1

is an arbitrary constant and C

2

= C

1

/2. But v = dy/dx, so we have

dy

dx

=

_

y

2

+C

2

_

2

,

dx =

dy

(y

2

+C

2

)

2

,

x =

_

dy

(y

2

+C

2

)

2

=

_

_

1

2C

3/2

2

_

tan

1

_

y

C2

_

+

C2y

y

2

+C2

_

+C

3

, C

2

> 0,

1

3y

3

+C

3

, C

2

= 0,

1

2(C2)

3/2

y

2

y

2

+C2

+C

3

, C

2

< 0.

P(x)y

+Q(x)y

+R(x)y = 0.

We can eliminate y by letting y = e

v

. Then y

= e

v

v

and y

= e

v

(v

)

2

+e

v

v

.

The equation then becomes

P(x)

_

e

v

(v

)

2

+e

v

v

_

+Q(x)e

v

v

+R(x)e

v

= 0,

which is a rst order equation in v

.

Example 2.23. Solve x

2

y

+

_

x x

2

_

y

e

2x

y = 0.

Solution. Let y = e

v

. Then the equation becomes

x

2

e

v

(v

)

2

+x

2

e

v

v

+

_

x x

2

_

e

v

v

e

2x

e

v

= 0.

Write z = v

. Then

x

2

z

+x

2

z

2

+ (1 x) xz = e

2x

.

Now we are on our ownthere is no standard technique for this case. Suppose

we try u = xy. Then z = u/x and

z

=

u

x

2

+

1

x

u

.

Then it follows that

xu

+u

2

xu = e

2x

.

This is a bit simpler, but it is still nonstandard. We can see that letting u = se

x

2.5. SUBSTITUTIONS 23

will give us some cancellation. Thus, u

= s

e

x

+se

x

and our equation becomes

xs

e

x

+

xse

x

+s

2

e

2x

xse

x

= e

2x

,

xs

+s

2

e

x

= e

x

,

xs

= e

x

_

1 s

2

_

,

s

1 s

2

=

e

x

x

,

1

2

ln

_

1 +s

1 s

_

=

_

e

x

x

dx.

Working our way back through the subsitutions we nd that s = zxe

x

so our

solution becomes

1

2

ln

_

1 +zxe

x

1 zxe

x

_

=

_

e

x

x

dx.

Using algebra, we could solve this equation for z in terms of x and then integrate

the result to get v which then determines y = e

v

as a function of x. The

algebraic solution for z is messy and we will not be able to nd a closed form

expression for the antidervative v, so we will abandon the calculation at this

point. In practical applications one would generally use power series techniques

on equations of this form and calculate as many terms of the Taylor series of

the solution as are needed to give the desired accuracy.

Next consider equations of the form

(a

1

x +b

1

y +c

1

) dx + (a

2

x +b

2

y +c

2

) dy = 0.

If c

1

= c

2

= 0, the equation is homogeneous of degree 1. If not, try letting

x = x h and y = y k. We try to choose h and k to make the equation

homogeneous. Since h and k are constants, we have d x = dx and d y = dy.

Then our equation becomes

(a

1

x +a

1

h +b

1

y +b

1

k +c

1

) d x + (a

2

x +a

2

h +b

2

y +b

2

k +c

2

) d y = 0.

We want a

1

h +b

1

k = c

1

and a

2

h +b

2

k = c

2

. We can always solve for h and

k, unless

a

1

b

1

a

2

b

2

= 0.

24CHAPTER 2. FIRST ORDER ORDINARY DIFFERENTIAL EQUATIONS

So suppose

a

1

b

1

a

2

b

2

= 0.

Then (a

2

, b

2

) = m(a

1

, b

1

). Let z = a

1

x + b

1

y. Then dz = a

1

dx + b

1

dy. If

b

1

,= 0, we have

dy =

dz a

1

dx

b

1

,

(z +c

1

) dx + (mz +c

2

)

dz a

1

dx

b

1

= 0,

_

z +c

1

+

a

1

b

1

_

dx +

_

mz +c

2

b

1

_

dx = 0,

b

1

dx =

mz +c

2

z +c

1

+a

1

/b

1

dz.

This is a separable equation.

If b

1

= 0 but b

2

,= 0, we use z = a

2

x + b

2

y instead. Finally, if both b

1

= 0

and b

2

= 0, then the original equation is separable.

Chapter 3

Applications and Examples

of First Order Ordinary

Dierential Equations

3.1 Orthogonal Trajectories

An equation of the form f(x, y, C) = 0 determines a family of curves, one for

every value of the constant C. Figure 3.1 shows curves with several values of

C. Dierentiating f(x, y, C) = 0 with respect to x gives a second equation

C=1

C=4

C=9

=/4

=0

x

y

Figure 3.1: Graphs of x

2

+y

2

C = 0 with various values of C.

25

26CHAPTER 3. APPLICATIONS ANDEXAMPLES OF FIRST ORDER ODES

g(x, y, C, dy/dx) = 0, which is explicitly given by

g =

f

x

(x, y, C) +

f

y

(x, y, C)

dy

dx

.

Example 3.1. We have

x

2

+y

3

+Cx

4

= 0 implies 2x + 3y

2

dy

dx

+ 4Cx

2

= 0.

We can eliminate C between the equations f = 0 and g = 0, that is,

f = 0 = C =

x

2

+y

2

x

4

.

Therefore,

g = 0 = 2x 3y

2

dy

dx

4

x

2

+y

3

x

= 0.

This yields a rst order de, where the solution includes the initial family of

curves.

So just as a rst order de usually determines a one-parameter family of

curves (its solutions), a one-parameter family of curves determines a de.

A coordinate system (Cartesian coordinates, polar coordinates, etc.) consists

of two families of curves which always intersect at right angles. Two families of

curves which always intersect at right angles are called orthogonal trajectories

of each other. Given a one-parameter family of curves, we wish to nd another

family of curves which always intersects it at right angles.

Example 3.2. Find orthogonal trajectories of y = C sin(x).

Solution. Figure 3.2 shows the plot of y = C sin(x) with several values of C.

We have C = y/ sin(x), so it follows that

dy

dx

= C

..

y/ sin(x)

cos(x) = y

cos(x)

sin(x)

= y cot(x).

Two curves meet at right angles if the product of their slopes is 1, i.e., m

new

=

1/m

old

. So orthogonal trajectories satisfy the equation

dy

dx

=

1

y cot(x)

=

tan(x)

y

.

3.2. EXPONENTIAL GROWTH AND DECAY 27

-3p -2p - p 2p 3p

x

-5

-3

-1

1

3

5

Figure 3.2: A plot of the family y = C sin(x) (in black) and its orthogonal

trajectories (in gray).

This is a separable equation, and we quickly have

y dy = tan(x),

y

2

2

= ln([cos(x)[) +C,

y =

_

2 ln([cos(x)[) +C

1

,

where C

1

= 2C is also an arbitrary constant.

3.2 Exponential Growth and Decay

There are many situations where the rate of change of some quantity x is pro-

portional to the amount of that quantity, i.e., dx/dt = kx for some constant k.

The general solution is x = Ae

kt

.

1. Radium gradually changes into uranium. If x(t) represents the amount of

radium present at time t, then dx/dt = kx. In this case, k < 0.

2. Bank interest. The amount of interest you receive on your bank account is

proportional to your balance. In this case, k > 0. Actually, you do not do

quite this well at the bank because they compound only daily rather than

continuously, e.g., if we measure this time in years, our model predicts

that x = x

0

e

kt

, where x

0

is your initial investment and k is the interest

rate. Actually, you get

x = x

0

_

1 +

k

n

t

_

n

,

28CHAPTER 3. APPLICATIONS ANDEXAMPLES OF FIRST ORDER ODES

where n is the number of interest periods throughout the year. Typically,

n = 365 (daily interest), but the dierence is small when n = 365. Note

that lim

n

(1 +kt/n)

n

= e

kt

.

3. Population growth of rabbits. The number of rabbits born at any time

is roughly proportional to the number of rabbits present. If x(t) is the

number of rabbits at time t, then dx/dt = kx with k > 0. Obviously, this

model is not accurate as it ignores deaths.

Example 3.3. The half-life of radium is 1600 years, i.e., it takes 1600 years for

half of any quantity to decay. If a sample initially contains 50 g, how long will

it be until it contains 45 g?

Solution. Let x(t) be the amount of radium present at time t in years. Then

we know that dx/dt = kx, so x = x

0

e

kt

. With x(0) = 50, we quickly have

x = 50e

kt

. Solving for t gives t = ln(x/50) /k. With x(1600) = 25, we have

25 = 50e

1600k

. Therefore,

1600k = ln

_

1

2

_

= ln(2) ,

giving us k = ln(2) /1600. When x = 45, we have

t =

ln(x/50)

k

=

ln(45/50)

ln(2) /1600

= 1600

ln(8/10)

ln(2)

= 1600

ln(10/8)

ln(2)

1600

0.105

0.693

1600 0.152 243.2.

Therefore, it will be approximately 243.2 years until the sample contains 45 g

of radium.

3.3 Population Growth

Earlier, we discussed population growth where we considered only births. We

now consider deaths also. Take, as our model, the following.

Let N(t) be the number of people at time t. Assume that the land is intrinsi-

cally capable of supporting L people and that the rate of increase is proportional

to both N and L N. Then

dN

dt

= kN (L N) ,

3.4. PREDATOR-PREY MODELS 29

and the solution is

N =

L

1 + (L/N

0

1) e

kLt

,

where k is the proportionality constant.

3.4 Predator-Prey Models

Example 3.4 (Predator-Prey). Consider a land populated by foxes and rab-

bits, where the foxes prey upon the rabbits. Let x(t) and y(t) be the number

of rabbits and foxes, respectively, at time t. In the absence of predators, at any

time, the number of rabbits would grow at a rate proportional to the number

of rabbits at that time. However, the presence of predators also causes the

number of rabbits to decline in proportion to the number of encounters between

a fox and a rabbit, which is proportional to the product x(t)y(t). Therefore,

dx/dt = ax bxy for some positive constants a and b. For the foxes, the

presence of other foxes represents competition for food, so the number declines

proportionally to the number of foxes but grows proportionally to the number

of encounters. Therefore dy/dt = cx + dxy for some positive constants c and

d. The system

dx

dt

= ax bxy,

dy

dt

= cy +dxy

is our mathematical model.

If we want to nd the function y(x), which gives the way that the number

of foxes and rabbits are related, we begin by dividing to get the dierential

equation

dy

dx

=

cy +dxy

ax bxy

with a, b, c, d, x(t), y(t) positive.

This equation is separable and can be rewritten as

(a by) dy

y

=

(c +dx) dx

x

.

Integrating gives

a ln(y) by = c ln(x) +dx +C,

30CHAPTER 3. APPLICATIONS ANDEXAMPLES OF FIRST ORDER ODES

or equivalently

y

a

e

by

= kx

c

e

dx

(3.1)

for some positive constant k = e

C

.

The graph of a typical solution is shown in Figure 3.3.

Figure 3.3: A typical solution of the Predator-Prey model with a = 9.4, b = 1.58,

c = 6.84, d = 1.3, and k = 7.54.

Beginning at a point such as a, where there are few rabbits and few foxes, the

fox population does not initially increase much due to the lack of food, but with

so few predators, the number of rabbits multiplies rapidly. After a while, the

point b is reached, at which time the large food supply causes the rapid increase

in the number of foxes, which in turn curtails the growth of the rabbits. By

the time point c is reached, the large number of predators causes the number of

rabbits to decrease. Eventually, point d is reached, where the number of rabbits

has declined to the point where the lack of food causes the fox population to

decrease, eventually returning the situation to point a.

3.5 Newtons Law of Cooling

Let T and T

s

be the temperature of an object and its surroundings, respectively.

Let T

0

and T

s0

be initial temperatures. Then Newtons Law of Cooling states

3.6. WATER TANKS 31

that

dT

dt

= k (T

s

T) , k > 0.

As T changes, the object gives or takes heat from its surroundings. We now

have two cases.

Case 1: T

s

is constant. This occurs either because the heat given o is

transported elsewhere or because the surroundings are so large that the contri-

bution is negligible. In such a case, we have

dT

dt

+kT = kT

s

,

which is linear, and the solution is

T = T

0

e

kt

+T

s

_

1 e

kt

_

.

Case 2: The system is closed. All heat lost by the object is gained by its

surroundings. We need to nd T

s

as a function of T. We have

change in temperature =

heat gained

weight specic heat capacity

.

Therefore,

T T

0

wc

=

T

s

T

s0

w

s

c

s

,

T

s

= T

s0

+

wc

w

s

c

s

(T

0

T) ,

dT

dt

+k

_

1 +

wc

w

s

c

s

_

T = k

_

T

s0

+

wc

w

s

c

s

T

0

_

.

This is linear with dierent constants than the previous case. The solution is

T = T

0

+

_

T

s0

+

wc

wscs

T

0

1 +

wc

wscs

_

_

1 e

k(1+

wc

wscs

)t

_

.

3.6 Water Tanks

Example 3.5. A tank contains a salt water solution consisting initially of 20 kg

of salt dissolved into 10 of water. Fresh water is being poured into the tank

at a rate of 3

2

/min. Figure 3.4 shows this setup. Find the amount of salt in the tank after

5 minutes.

32CHAPTER 3. APPLICATIONS ANDEXAMPLES OF FIRST ORDER ODES

Figure 3.4: Fresh water is being poured into the tank as the well-mixed solution

is owing out.

Solution. Let Q(t) be the amount (in kilograms) of salt in the tank at time t

(in minutes). The volume of water at time t is

10 + 3t 2t = 10 +t.

The concentration at time t is given by

amount of salt

volume

=

Q

10 +t

kg per litre. Then

dQ

dt

= (rate at which salt is leaving) =

Q

10 +t

2 =

2Q

10 +t

.

Thus, the solution to the problem is the solution of

dQ

dt

=

2Q

10 +t

evaluated at t = 5. We see that it is simply a separable equation.

3.6. WATER TANKS 33

To solve it, we have

dQ

Q

= 2

dt

10 +t

,

_

dQ

Q

= 2

_

dt

10 +t

,

ln([Q[) = 2 ln([10 +t[) ,

ln([Q[) = ln

_

[10 +t[

2

_

+C,

where C is a constant. Thus,

ln([Q[) = ln

_

[10 +t[

2

_

+C,

= ln

_

A[10 +t[

2

_

,

[Q[ = A[10 +t[

2

where A = e

C

. But Q 0 (we cannot have a negative amount of salt) and t 0

(we do not visit the past), so we remove absolute value signs, giving us

Q = A(10 +t)

2

.

Initially, i.e., at t = 0, we know that the tank contains 20 kg of salt. Thus,

the initial condition is Q(0) = 20, and we have

Q(20) = 0 =

A

(10 + 0)

2

= 20 =

A

100

= 20 = A = 2000.

Our equation therefore becomes

Q(t) =

2000

(10 +t)

2

.

Figure 3.5 shows a plot of the solution. Evaluating at t = 5 gives

Q(5) =

2000

15

2

=

80

225

=

80

9

8.89.

Therefore, after 5 minutes, the tank will contain approximately 8.89 kg of salt.

Additional conditions desired of the solution (Q(0) = 20) in the above exam-

ple) are called boundary conditions and a dierential equations together with

boundary conditions is called a boundary-value problem (BVP).

34CHAPTER 3. APPLICATIONS ANDEXAMPLES OF FIRST ORDER ODES

0 100 200 300 400 500 600

t

5

10

15

20

Q

Figure 3.5: The plot of the amount of salt Q(t) in the tank at time t shows that

salt leaves slower as time moves on.

A =

A

(10+t)

2

is the general solution to the DE

dQ

dt

=

2dt

10+t

. Q =

2000

(10+t)

2

is the

solution to the boundary value problem

dQ

dt

=

2dt

10+t

.; Q(0) = 200. Boundary-

value problems like this one where the boundary conditions are initial values of

various quantities are also called initial-value problems (IVP).

3.7 Motion of Objects Falling Under Gravity

with Air Resistance

Let x(t) be the height at time t, measured positively on the downward direction.

If we consider only gravity, then

a =

d

2

x

dt

2

is a constant, denoted g, the acceleration due to gravity. Note that F = ma =

mg. Air resistance encountered depends on the shape of the object and other

things, but under most circumstances, the most signicant eect is a force op-

posing the motion which is proportional to a power of the velocity. So

F

..

ma

= mg kv

n

and

d

2

x

dt

2

= g

k

m

_

dx

dt

_

n

,

3.7. MOTIONOF OBJECTS FALLINGUNDER GRAVITYWITHAIR RESISTANCE35

which is a second order de, but there is no x term. So it is rst order in x

.

Therefore,

dv

dt

= g

k

m

v

n

.

This is not easy to solve, so we will make the simplifying approximation that

n = 1 (if v is small, there is not much dierence between v and v

n

). Therefore,

we have

dv

dt

= g

k

m

v,

dv

dt

+

k

m

v = g,

The integrating factor is

I = e

R

k

m

dt

= e

kt/m

.

Therefore,

_

dv

dt

+

k

m

v

_

e

kt/m

= ge

kt/m

,

e

kt/m

v =

gm

k

e

kt/m

+C,

v =

mg

k

+Ce

kt/m

,

where C is an arbitrary constant. Note that

v(0) = v

0

= v

0

=

mg

k

+C = C = v

0

mb

k

.

So we have

v

..

dx/dt

=

mg

k

+

_

v

0

mg

k

_

e

kt/m

.

Note that

_

x

x0

dx

. .

xx0

=

_

t

0

v dt =

mg

k

t

m

k

_

v

0

mg

k

__

e

kt/m

1

_

.

Thus, we nally have

x = x

0

+

mg

k

t +

m

k

_

v

0

mg

k

__

1 e

kt/m

_

.

36CHAPTER 3. APPLICATIONS ANDEXAMPLES OF FIRST ORDER ODES

3.8 Escape Velocity

Let x(t) be the height of an object at time t measured positively upward (from

the centre of the earth). Newtons Law of Gravitation states that

F

..

ma

=

kmM

x

2

,

where m is the mass of the object, M is the mass of the Earth, and k > 0. Note

that

x

=

kM

x

2

.

We dene the constant g known as the acceleration due to gravity by x

= g

when x = R, where R = radius of the earth. So k = gR

2

/M. Therefore

x

= gR

2

/x

2

. Since t does not appear, letting v = dx/dt so that

d

2

x

dt

2

=

dv

dx

dx

dt

=

dv

dx

v

will reduce it to rst order. Thus,

v

dv

dx

=

gR

2

x

2

,

v dv =

gR

2

x

2

dx,

_

v dv =

_

gR

2

x

2

dx,

v

2

2

=

gR

2

x

+C,

for some arbitrary constant C. Note that v decreases as x increases, and if v is

not suciently large, eventually v = 0 at some height x

max

. So C = gR

2

/x

max

and

v

2

= 2R

2

g

_

1

x

1

x

max

_

.

Suppose we leave the surface of the Earth starting with velocity V . How far

3.9. PLANETARY MOTION 37

will we get? (v = V when x = R). We have

V

2

= 2R

2

g

_

1

R

1

x

max

_

,

V

2

2R

2

g

=

1

R

1

x

max

,

1

x

max

=

1

R

V

2

2R

2

g

=

2Rg V

2

2R

2

g

,

x

max

=

2R

2

g

2Rg V

2

.

That is, if V

2

< 2Rg, we will get as far as

2R

2

g

2Rg V

2

and then fall back. To escape, we need V

2Rg. Thus, the escape velocity

is

2Rg 6.9

mi

/s, or roughly 25000 mph.

3.9 Planetary Motion

Newtons law of gravitation states that

F

..

ma

=

km

r

2

r.

Suppose we let a

r

= k/r

2

and a

0

= 0. Then

a

r

= a

x

cos() +a

y

sin(), a

= a

x

sin() +a

y

cos().

Let x = r cos() and y = r sin(). Then

x

= r

cos() r sin()

,

a

x

= x

= r

cos() r

sin()

sin()

r cos() (

)

2

r sin()

,

y

= r

sin() +r cos()

,

a

y

= y

= r

sin() +r

cos()

+r

cos()

r sin() (

)

2

+r cos()

= r

sin() + 2r

cos() (

)

2

r sin() +

r cos().

38CHAPTER 3. APPLICATIONS ANDEXAMPLES OF FIRST ORDER ODES

Now a

r

becomes

a

r

= r

cos

2

() 2r

sin() cos() (

)

2

r cos

2

()

sin() cos()

+r

sin

2

() + 2r

sin() cos() (

)

2

r sin

2

() +

r sin() cos()

= r

)

2

r

and a

becomes

a

= r

sin() cos() + 2r

sin

2

() + (

)

2

r sin() cos() +

r sin

2

()

+r

sin() cos() + 2r

cos

2

() (

)

2

r sin() cos() +

r cos

2

()

= 2r

r.

Note that

r

)

2

r =

k

r

2

, ()

2r

r = 0. ()

Equation () implies that

2rr

. .

d(r

2

)

r

2

= 0 = r

2

= h.

We want the path of the planet, so want an equation involving r and . There-

fore, we eliminate t between Equation () and r

2

= h. Thus,

r

=

dr

d

=

dr

d

h

r

2

,

r

=

d

2

r

d

2

h

r

2

2

dr

d

h

r

2

r

=

d

2

r

d

2

h

r

2

h

r

2

2

dr

d

h

r

3

dr

d

h

r

2

=

h

2

r

2

_

2

r

3

_

dr

d

_

2

1

r

2

d

2

r

d

2

_

.

Note that

A =

Z

0

Z

r()

0

r dr d =

Z

0

r

()

2

2

d,

so dA/d = r()

2

/2. So in our case, A

Law (area swept out is proportional to time).

3.10. PARTICLE MOVING ON A CURVE 39

Let u = 1/r. Then

du

d

=

1

r

2

dr

d

,

d

2

u

d

2

=

2

r

3

_

dr

d

_

2

1

r

2

d

2

r

d

2

=

r

2

r

h

2

=

r

h

2

u

2

.

Therefore, Equation () implies that

h

2

u

2

d

2

u

d

2

_

h

r

2

_

2

r =

k

r

2

,

h

2

u

2

d

2

u

d

2

u

3

h

2

= ku

2

, u ,= 0,

d

2

u

d

2

+u =

k

h

2

.

Therefore,

u

..

1/r

= Bcos( ) +

k

h

2

= C

1

sin() +C

2

cos() +

k

h

2

,

where

r =

1

k/h

2

+Bcos( )

=

1

k/h

2

+C

1

sin() +C

2

cos()

.

Therefore,

k

h

2

r+C1x+C2y

..

k

h

2

r +C

1

r sin() +C

2

r cos() = 1,

k

h

2

r = 1 C

1

x C

2

x,

k

2

h

4

_

x

2

+y

2

_

= (1 C

1

x C

2

y)

2

,

which is the equation of an ellipse.

3.10 Particle Moving on a Curve

Let x(t) and y(t) denote the x and y coordinates of a point at time t, respectively.

Pick a reference point on the curve and let s(t) be the arc length of the piece

of the curve between the reference point and the particle. Figure 3.6 illustrates

this idea. Then

40CHAPTER 3. APPLICATIONS ANDEXAMPLES OF FIRST ORDER ODES

Reference point

Particle

Figure 3.6: A particle moving on a curve.

v =

ds

dt

, |a| = a =

d

2

s

dt

2

.

The vector v is in the tangent direction. The goal is to nd components a in

the tangential and normal directions. Let be the angle between the tangent

line and the x axis. Then

cos() =

dx

ds

, sin() =

dy

ds

.

In the standard basis, we have a = (a

x

, a

y

), where

a

x

=

d

2

x

dt

2

, a

y

=

d

2

y

dt

2

.

To get components in T, n basis, apply the matrix representing the change of

basis from T, n to the standard basis, i.e., rotation by , i.e.,

_

cos() sin()

sin() cos()

__

a

x

a

y

_

=

_

a

x

cos() +a

y

sin()

a

x

sin() +a

y

cos()

_

,

i.e.,

a

T

= a

x

cos() +a

y

sin() = a

x

dx

ds

+a

y

dy

ds

=

_

a

x

dx

dt

+a

y

dy

dt

_

dt

ds

=

a

x

v

x

+a

y

v

y

v

and

a

n

= a

x

sin() +a

y

cos() =

a

y

v

x

a

x

v

y

v

.

Also, we have v =

_

v

2

x

+v

2

y

, so

d

2

s

dt

2

=

dv

dt

=

2v

x

dvx

dt

+ 2v

y

dvy

dt

2

_

v

2

x

+v

2

y

=

v

x

a

x

+v

y

a

y

v

= a

T

.

3.10. PARTICLE MOVING ON A CURVE 41

But this was clear to start with because it is the component of a in the tangential

direction that aects ds/dt.

We now have

y

=

dy

dx

=

dy/dt

dx/dt

=

v

y

v

x

,

y

=

v

x

dvy

dx

v

y

dvx

dx

v

2

x

=

v

x

dvy

dt

dt

dx

v

y

dvx

dt

dt

dx

v

2

x

=

v

x

a

y

v

y

a

x

v

3

x

=

a

n

v

v

3

x

,

where a

n

= y

v

3

x

/v and

v

x

=

dx

dt

=

dx

ds

ds

dt

=

dx

ds

v,

so

a

n

=

y

_

dx

ds

_

3

v

3

v

=

y

v

2

(ds/dx)

3

=

y

_

1 + (y

)

2

_

3/2

v

2

.

Let

R =

_

1 + (y

)

2

_

3/2

y

(3.2)

so that a

n

= v

2

/R. The quantity R is called the radius of curvature. It is the

radius of the circle which mostly closely ts the curve at that point. This is

shown in Figure 3.7. If the curve is a circle, then R is a constant equal to the

R

Figure 3.7: The radius of curvature is the circle whose edge best matches a

given curve at a particular point.

radius.

Conversely, suppose that R is a constant. Then

Ry

=

_

1 + (y

)

2

_

3/2

.

42CHAPTER 3. APPLICATIONS ANDEXAMPLES OF FIRST ORDER ODES

Let z = y

. Then

Rz

=

_

1 +z

2

_

3/2

and

dx =

Rdz

(1 +z

2

)

3/2

,

so

x =

_

R

(1 +z

2

)

3/2

dz =

Rz

1 +z

2

+A,

x A =

R

_

(1/z)

2

+ 1

,

_

1

z

_

2

+ 1 =

R

2

(x A)

2

,

_

1

z

_

2

=

R

2

(x A)

2

1 =

R

2

(x A)

2

(x A)

2

,

z

2

=

(x A)

2

R

2

(x A)

2

,

z

..

dy/dx

=

x A

_

R

2

(x A)

2

.

Therefore,

y =

_

x A

_

R

2

(x A)

2

dx

=

_

R

2

(x A)

2

+B,

(y B)

2

= R

2

(x A)

2

,

(x A)

2

+ (y B)

2

= R

2

.

Suppose now that the only external force F acting on the particle is gravity,

i.e., F = (0, mg). Then in terms of the T, n basis, the components of F are

_

cos() sin()

sin() cos()

__

0

mg

_

=

_

mg sin()

mg cos()

_

=

_

mg

dy

ds

mg

dx

ds

_

.

3.10. PARTICLE MOVING ON A CURVE 43

Equating the tangential components gives us

m

d

2

s

dt

2

=

mg

dy

ds

,

ds

dt

= v,

d

2

s

dt

2

=

dv

dt

=

ds

dt

=

dv

ds

v,

dv

ds

v = g

dy

ds

,

v dv = g dy,

v

2

2

= g(y y

0

) = g(y

0

y),

ds

dt

= v =

_

2g(y

0

y).

To proceed, we must know y(s), which depends on the curve. Therefore,

dt =

ds

_

2g(y

0

y(s))

,

t =

_

ds

_

2g(y

0

y(s))

.

If the ma

n

ever becomes greater than the normal component of F, the particle

will y o the curve. This will happen when

m

y

v

3

x

v

=

mg

dx

ds

= g

dx

dt

dt

ds

= g

v

x

v

,

y

v

2

x

= g,

v

x

=

dx

dt

=

dx

ds

ds

dt

=

dx

ds

v.

Therefore,

_

_

_

_

_

g = y

_

dx

ds

_

2

v

2

=

y

v

2

(ds/dx)

2

=

y

v

2

1 + (y

)

2

=

y

g(y

0

y)

1 + (y

)

2

_

_

_

_

_

= 2y

(y y

0

) = 1 + (y

)

2

.

Example 3.6. A particle is placed at the point (1, 1) on the curve y = x

3

and

released. It slides down the curve under the inuence of gravity. Determine

whether or not it will ever y o the curve, and if so, where.

44CHAPTER 3. APPLICATIONS ANDEXAMPLES OF FIRST ORDER ODES

Solution. First note that

y = x

3

, y

= 3x

2

, y

= 6x, y

0

= 1.

Therefore,

2y

(y y

0

) = 1 (y

)

2

,

12x

_

x

3

1

_

= 1 + 9x

4

,

12x

4

12x = 1 + 9x

4

,

3x

4

12x 1 = 0.

Solving for the relevant value of x gives us x 0.083, which corresponds to y

0.00057. Therefore, the particle will y o the curve at (0.083, 0.00057).

Figure 3.8 shows the graph.

H1,1L

H-0.083,-0.00057L

-2 -1 1 2

x

-8

-6

-4

-2

2

4

6

8

Figure 3.8: A particle released from (1, 1), under the inuence of gravity, will

y o the curve at approximately (0.083, 0.00057).

Chapter 4

Linear Dierential

Equations

Although we have dealt with a lot of manageable cases, rst order equations

are dicult in general. But second order equations are much worse. The most

manageable case is linear equations. We begin with the general theory of linear

dierential equations. Specic techniques for solving these equations will be

given later.

Denition (nth order linear dierential equation)

An nth order linear dierential equation is an equation of the form

d

n

y

dx

n

+P

1

(x)

d

n1

y

dx

n1

+P

2

(x)

d

n2

y

dx

n2

+ +P

n

(x)y = Q(x).

An nth order linear equation can be written as a linear system (see Chapter

9) as follows. Let

y

1

(x) = y(x), y

2

(x) =

dy

dx

, y

3

(x) =

d

2

y

dx

2

, . . . , y

n

(x) =

d

n1

y

dx

n1

.

Then

dy

1

dx

= y

2

,

dy

2

dx

= y

3

, . . . ,

dy

n1

dx

= y

n

If the coecient of d

n

y/dx

n

is not 1, we can divide the equation through this coecient

to write it in normal form. A linear equation is in normal form when the coecient of y

(n)

is

1.

45

46 CHAPTER 4. LINEAR DIFFERENTIAL EQUATIONS

and we have

dy

n

dx

=

d

n

y

dx

n

= P

1

(x)

d

n1

y

dx

n1

P

n

(x)y +Q(x)

= P

1

(x)y

n

P

n

(x)y

1

+Q(x).

Therefore,

dY

dx

= A(x)Y +B(x),

where

A(x) =

_

_

0 1 0 0

0 0 1 0

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

0 0 0 1

P

n

(x) P

n1

(x) P

n2

(x) P

1

(x)

_

_

and

B(x) =

_

_

0

0

.

.

.

0

Q(x)

_

_

.

Theorem 4.1

If P

1

(x), . . . , P

n

(x), Q(x) are continuous throughout some interval I and x

0

is an interior point of I, then for any numbers W

0

, . . . , W

n1

, there exists

a unique solution

throughout I of

d

n

y

dx

n

+P

1

(x)

d

n1y

dx

n1

+ +P

n

(x)y = Q(x)

satisfying

y(x

0

) = w

0

, y

(x

0

) = w

1

, . . . , y

(n1)

(x

0

) = w

n1

.

Proof. The proof is given in Chapter 10.

4.1. HOMOGENEOUS LINEAR EQUATIONS 47

4.1 Homogeneous Linear Equations

To begin, consider the case where Q(x) 0 called the homogeneous case, where

y

(n)

+P

1

(x)y

(n1)

+P

2

(x)y

(n2)

+ +P

n

(x)y = 0. (4.1)

Proposition 4.2

1. If r(x) is a solution of Equation (4.1), then so is ar(x) for all r R.

2. If r(x) and s(x) are solutions of Equation (4.1), then so is r(x) +s(x).

Proof.

1. Substituting y = ar into the lhs gives

ar

(n)

+P

1

(x)ar

(n1)

+ +P

n

(x)ar = a

_

r

(n)

+ +P

n

(x)r

_

.

But since r is a solution, we have

a

_

r

(n)

+ +P

n

(x)r

_

= a 0.

Therefore, ar is a solution.

2. Substituting y = r +s in the lhs gives

_

_

_

r

(n)

+s

(n)

+P

1

(x)

_

r

(n1)

+ +s

(n1)

_

+

+P

n

(x) (r +s)

_

_

_ =

_

r

(n)

+ +P

n

(x)r

+s

(n)

+ +P

n

(x)s

_

= 0 + 0 = 0.

Therefore, r +s is a solution.

Corollary 4.3

The set of solutions to Equation (4.1) forms a vector space.

48 CHAPTER 4. LINEAR DIFFERENTIAL EQUATIONS

Proof. Let V be the set of solutions to Equation (4.1) and let y

1

, y

2

, . . . , y

n

V .

Further, let x

0

R. Set

W(y

1

, y

2

, . . . , y

n

) (x) =

y

1

(x) y

2

(x) y

n

(x)

y

1

(x) y

2

(x) y

n

(x)

.

.

.

.

.

.

.

.

.

.

.

.

y

(n1)

1

(x) y

(n1)

2

(x) y

(n1)

n

(x)

,

the Wronskian of y

1

, y

2

, . . . , y

n

. Then

dW

dx

=

two equal rows

..

1

(x) y

n

(x)

y

1

(x) y

n

(x)

.

.

.

.

.

.

.

.

.

y

(n1)

1

(x) y

(n1)

n

(x)

+

two equal rows

..

y

1

(x) y

n

(x)

y

1

(x) y

n

(x)

y

1

(x) y

n

(x)

.

.

.

.

.

.

.

.

.

y

(n1)

1

(x) y

(n1)

n

(x)

+

+

y

1

(x) y

n

(x)

.

.

.

.

.

.

.

.

.

y

(n2)

1

(x) y

(n2)

n

(x)

y

(n)

1

(x) y

(n)

n

(x)

y

1

(x) y

n

(x)

.

.

.

.

.

.

.

.

.

y

(n2)

1

(x) y

(n2)

n

(x)

y

(n)

1

(x) y

(n)

n

(x)

y

1

(x) y

n

(x)

.

.

.

.

.

.

.

.

.

y

(n2)

1

(x) y

(n2)

n

(x)

n

k=1

P

k

(x)y

(nk)

1

n

k=1

P

k

(x)y

(nk)

n

=

n

k=1

_

_

_

_

_

_

P

k

(x)

y

1

(x) y

n

(x)

.

.

.

.

.

.

.

.

.

y

(n2)

1

(x) y

(n2)

n

(x)

y

(n)

1

(x) y

(n)

n

(x)

_

_

_

_

_

_

= P

1

(x)W.

Therefore, W = Ae

R

P1(x) dx

. This is Abels formula. There are two cases:

A = 0 = W(x) 0,

A ,= 0 = W(x) ,= x

for all x.

4.1. HOMOGENEOUS LINEAR EQUATIONS 49

Theorem 4.4

Let V be the set of solutions to Equation (4.1) and suppose that

y

1

, y

2

, . . . , y

n

V . Then y

1

, y

2

, . . . , y

n

are linearly independent if and only

if W ,= 0.

Proof. Let V be the set of solutions to Equation (4.1) and suppose that y

1

, y

2

, . . . , y

n

V . Suppose that y

1

, y

2

, . . . , y

n

are linearly dependent. Then there exists con-

stants c

1

, c

2

, . . . , c

n

such that

c

1

y

1

(x) +c

2

y

2

(x) + +c

n

y

n

(x) = 0,

c

1

y

1

(x) +c

2

y

2

(x) + +c

n

y

n

(x) = 0,

.

.

.

c

1

y

(n1)

1

(x) +c

2

y

(n1)

2

(x) + +c

n

y

(n1)

n

(x) = 0.

Since there exists a nonzero solution for (c

1

, c

2

, . . . , c

n

) of the equation, we have

M

_

_

c

1

c

2

.

.

.

c

n

_

_

= 0,

so it follows that [M[ = 0 = W.

Now suppose that W = 0. Select an x

0

R. Since W(x) 0, we have

W(x

0

) = 0. Thus, there exists constants c

1

, c

2

, . . . , c

n

not all zero such that

M

_

_

c

1

c

2

.

.

.

c

n

_

_

,

where

M =

_

_

y

1

(x

0

) y

n

(x

0

)

.

.

.

.

.

.

.

.

.

y

(n1)

1

(x

0

) y

(n1)

n

(x

0

)

_

_

,

50 CHAPTER 4. LINEAR DIFFERENTIAL EQUATIONS

that is,

c

1

y

1

(x

0

) +c

2

y

2

(x

0

) + +c

n

y

n

(x

0

) = 0,

c

1

y

1

(x

0

) +c

2

y

2

(x

0

) + +c

n

y

n

(x

0

) = 0,

.

.

.

c

1

y

(n1)

1

(x

0

) +c

2

y

(n1)

2

(x

0

) + +c

n

y

(n1)

n

(x

0

) = 0.

Let y(x) = c

1

y

1

(x) + c

2

y

2

(x) + + c

n

y

n

(x). Since y

1

, y

2

, . . . , y

n

are solutions

to Equation (4.1), so is f, and f satises

f(x

0

) = 0, f

(x

0

) = 0, f

(x

0

) = 0, . . . , f

(n1)

(x

0

) = 0.

But y = 0 also satises these conditions, and by the uniqueness part of Theorem

4.1, it is the only function satisfying them. Therefore, f 0, that is,

c

1

y

1

(x

0

) +c

2

y

2

(x

0

) + +c

n

y

n

(x

0

) 0.

Therefore, y

1

, y

2

, . . . , y

n

are linearly dependent.

Corollary 4.5

Let V be the set of solutions to Equation (4.1). Then we have dim(V ) = n.

Proof. Let V be the set of solutions to Equation (4.1). Pick a point x

0

R.

Then by Theorem 4.1, given any numbers w

1

, w

2

, . . . , w

n

, there exists a unique

solution of Equation (4.1) satisfying y

(j1)

(x

0

) for j = 1, 2, . . . , n. Let y

1

be

such a solution with (w

1

, w

2

, . . . , w

n

) = (0, 0, . . . , 0), let y

2

be such a solution

with (w

1

, w

2

, . . . , w

n

) = (0, 1, . . . , 0), etc., nally letting y

n

be such a solution

with (w

1

, w

2

, . . . , w

n

) = (0, 0, . . . , 1).

We claim that y

1

, y

2

, . . . , y

n

are linearly independent. Suppose that

c

1

y

1

(x) +c

2

y

2

(x) + +c

n

y

n

(x) 0.

Then dierentiating gives

c

1

y

(j)

1

(x) +c

2

y

(j)

2

(x) + +c

n

y

(j)

n

(x) 0

4.1. HOMOGENEOUS LINEAR EQUATIONS 51

for all j. In particular,

c

1

y

(j)

1

(x

0

) +c

2

y

(j)

2

(x

0

) + +c

n

y

(j)

n

(x

0

) 0

for all j. Therefore, y

1

, y

2

, . . . , y

n

are linearly independent and dim(V )) n.

Conversely, let y

1

, . . . , y

n

be linearly independent. Then W(y

1

, . . . , y

n

) (x) ,=

0 for all x. Suppose that y V and let

M =

_

_

y

1

(x

0

) y

2

(x

0

) y

n

(x

0

)

y

1

(x

0

) y

2

(x

0

) y

n

(x

0

)

.

.

.

.

.

.

.

.

.

.

.

.

y

(n1)

1

(x

0

) y

(n1)

2

(x

0

) y

(n1)

n

(x

0

)

_

_

.

Therefore, [M[ = [W(x

0

)[ ,= 0. So there exists constants (c

1

, c

2

, . . . , c

n

) such

that

M

_

_

c

1

c

2

.

.

.

c

n

_

_

=

_

_

y(x

0

)

y

(x

0

)

.

.

.

y

(n1)

(x

0

)

_

_

,

that is,

y

(j)

(x

0

) = c

1

y

(j)

1

+c

2

y

(j)

2

+ +c

1

y

(j)

1

for 0 i n 1. Let f(x) = c

1

y

1

(x) + c

2

y

2

(x) + + c

n

y

n

(x). Then f is

a solution to Equation (4.1) and y

(j)

(x

0

) = f

(j)

(x

0

) for j = 0, . . . , n 1. By

uniqueness, y = f, i.e., y(x) c

1

y

1

(x) + c

2

y

2

(x) + + c

n

y

n

(x). This means

that y is a linear combination of y

1

, . . . , y

n

. So y

1

, . . . , y

n

forms a basis for V .

Therefore, dim(V ) = n.

We conclude that, to solve Equation (4.1), we need to nd n linearly in-

dependent solutions y

1

, . . . , y

n

of Equation (4.1). Then the general solution to

Equation (4.1) is c

1

y

1

+c

2

y

2

+ +c

n

y

n

.

Example 4.6. Solve y

+ 5y

+ 6y = 0.

Solution. It is easy to check that y = e

3x

and y = e

2x

each satises y

+ 5y

+

6y = 0. Therefore, the general solution is c

1

e

3x

+ c

2

e

2x

, where c

1

and c

2

are

arbitrary constants.

52 CHAPTER 4. LINEAR DIFFERENTIAL EQUATIONS

4.1.1 Linear Dierential Equations with Constant Coe-

cients

The next problem concerns how to nd the n linearly independent solutions. For

n > 1, there is no good method in general. There is, however, a technique which

works in the special cases where all the coecients are constants. Consider the

constant coecient linear equation

a

0

y

(n)

+a

1

y

(n1)

+ +a

n1

y

+a

n

y = 0. (4.2)

We rst look for solutions of the form y = e

x

. If y = e

x

is a solution, then

a

0

n

e

x

+a

1

n1

e

x

+ +a

n

e

x

= 0.

Now since e

x

,= 0, we have

a

0

n

+a

1

n1

+ +a

n

= 0.

Let

p() = a

0

n

+a

1

n1

+ +a

n

. (4.3)

If Equation (4.3) has n distinct real roots r

1

, r

2

, . . . , r

n

, then we have n linearly

independent solutions e

r1x

, e

r2x

, . . . , e

rnx

. We would like one solution for each

root like this.

Suppose that Equation (4.3) has a repeated real root r. Then we do not

have enough solutions of the form e

x

. But consider y = xe

rx

. Then

y

= re

rx

+rxe

rx

= (rx + 1) e

rx

,

y

= re

rx

+r (rx + 1) e

rx

=

_

r

2

x + 2r

_

e

rx

,

y

= r

2

e

rx

+r

_

r

2

x + 2r

_

e

rx

=

_

r

3

x + 3r

2

_

e

rx

,

.

.

.

y

(n)

=

_

r

n

x +nr

n1

_

e

rx

.

4.1. HOMOGENEOUS LINEAR EQUATIONS 53

Substituting y = xe

rx

in Equation (4.2) gives us

n1

k=0

a

k

_

r

nk

x + (n k) r

nk

_

=

_

n

k=0

a

k

r

nk

_

xe

rx

+

_

n1

k=0

a

k

(n k) r

nk1

_

e

rx

= p(r)xe

rx

+p

(r)e

rx

.

Since r is a repeated root of p, both p(r) = 0 and p

p() = ( r)

2

q() = p

() = 2 ( r) q() + ( r)

2

q

()

= p

(r) = 0.

Therefore, xe

rx

is also a solution of Equation (4.2).

What about complex roots? They come in pairs, i.e., r

1

= a + ib and

r

2

= a ib, so e

(a+ib)x

and e

(aib)x

satisfy Equation (4.2). But we want real

solutions, not complex ones. But

e

(a+ib)x

= e

ax

(cos(bx) +i sin(bx)) ,

e

(aib)x

= e

ax

(cos(bx) i sin(bx)) .

Let

z

1

= e

ax

(cos(bx) +i sin(bx)) ,

z

2

= e

ax

(cos(bx) i sin(bx)) .

If z

1

and z

2

are solutions, then

z

1

+z

2

2

= e

ax

cos(bx),

z

1

z

2

2

= e

ax

sin(bx)

are solutions. So the two roots a + ib and a ib give two solutions e

ax

cos(bx)

and e

ax

sin(bx). For repeated complex roots, use xe

ax

cos(bx), xe

ax

sin(bx), etc.

So in all cases, we get n solutions.

The expression c

1

e

ax

cos(bx) +e

ax

c

2

sin(bx) can be rewritten as follows. Let

A =

_

c

2

1

+c

2

2

. Then

= cos

1

_

c

1

_

c

2

1

+c

2

2

_

,

54 CHAPTER 4. LINEAR DIFFERENTIAL EQUATIONS

that is,

cos() =

c

1

_

c

2

1

+c

2

2

, sin() =

_

1 cos

2

() =

c

2

_

c

2

1

+c

2

2

.

Therefore,

c

1

e

ax

cos(bx) +e

ax

c

2

sin(bx) = Ae

ax

_

c

1

A

cos(bx) +

c

2

A

sin(bx)

_

= Ae

ax

(cos() cos(bx) + sin() sin(bx))

= Ae

ax

(cos(bx )) .

Example 4.7. Solve y

(5)

4y

(4)

+ 5

+ 6

36y

+ 40y = 0.

Solution. We have

5

4

4

+ 5

3

+ 6

2

36 + 40 = 0,

( 2)

2

( + 2)

_

2

2 + 5

_

= 0.

Therefore, = 2, 2, 2, 1 2i and

y = c

1

e

2x

+c

2

xe

2x

+c

3

e

2x

+c

4

e

x

cos(2x) +c

5

sin(2x),

where c

1

, c

2

, c

3

, c

4

, c

5

are arbitrary constants.

4.2 Nonhomogeneous Linear Equations

A nonhomogeneous linear equation is an equation of the form

y

(n)

+p

1

(x)y

(n1)

+p

2

(x)y

(n2)

+ +p

n

(x)y = q(x). (4.4)

Let Equation (4.1) (p. 47) be the corresponding homogeneous equation. If u

and v are solutions of Equation (4.4), then uv is a solution of Equation (4.1).

Conversely, if u is a solution of Equation (4.4) and v is a solution of Equation

(4.1), then u +v is a solution of Equation (4.4).

Let y

1

, y

2

, . . . , y

n

be n linearly independent solutions of Equation (4.1) and

let y

p

be a solution of Equation (4.4). If y is any solution to Equation (4.4),

then y y

p

is a solution to Equation (4.1), so

y y

p

= c

1

y

1

+c

2

y

2

+ +c

n

y

n

.

4.2. NONHOMOGENEOUS LINEAR EQUATIONS 55

Therefore,

y = c

1

y

1

+c

2

y

2

+ +c

n

y

n

+y

p

is the general solution to Equation (4.4).

Example 4.8. Solve y

y = x.

Solution. The corresponding homogeneous equation is y

y = 0. It then

follows that

2

1 = 0 = 1. Therefore,

y

1

= e

x

, y

2

= e

x

.

By inspection, y

p

= x solves the given equation. Therefore, the general solution

is

y = c

1

e

x

+c

2

e

x

x,

where c

1

and c

2

are arbitrary constants.

We will consider more systematically how to nd y

p

later.

56 CHAPTER 4. LINEAR DIFFERENTIAL EQUATIONS

Chapter 5

Second Order Linear

Equations

A second order linear equation is an equation of the form

y

+P(x)y

To nd the general solution of Equation (5.1), we need

Two solutions: y

1

and y

2

of the corresponding homogeneous equation

y

+P(x)y

+Q(x) = 0.

One solution: y

p

of Equation (5.1).

Then the general solution is y = c

1

y

1

+c

2

y

2

+y

p

, where c

1

and c

2

are arbitrary

constants.

Except in the case where P(x) and Q(x) are constants (considered earlier

on p. 52), there is no general method of nding y

1

, but

1. given y

1

, there exists a method of nding y

2

.

2. given y

1

and y

2

, there exists a method of nding y

p

.

We will discuss (1) rst.

5.1 Reduction of Order

As mentioned earlier, we can nd a second solution from the rst. We rst

develop the tools we need.

57

58 CHAPTER 5. SECOND ORDER LINEAR EQUATIONS

Lemma 5.1

Let f(x) be a twice dierentiable function on a closed bounded interval

J. If x J : f(x) = 0 is innite, then there exists an x

0

J such that

f(x

0

) = 0 and f

(x

0

) = 0.

Proof. Let S = x J : f(x) = 0. If [S[ is innite, then by the Bolzano-

Weierstrass Theorem, S has an accumulation point x

0

, i.e., there exists an

x

0

S such that every interval about x

0

contains another point of S. Therefore,

by Rolles Theorem, every interval about x

0

contains a point where f

is 0.

Therefore, by continuity, f

(x

0

) = 0.

Corollary 5.2

Let y

1

(x) be a solution of y

+ P(x)y

interval J. If y

1

is not the zero solution, then y

1

(x) = 0 for at most nitely

many x in J.

Proof. Let V be the set of solutions of y

+ P(x)y

+ Q(x)y = 0 on J. Then

V is a two-dimensional vector space. So if y

1

, 0, then there exists a y

2

V

such that y

1

and y

2

are linearly independent. Therefore, W(y

1

, y

2

) ,= 0, i.e.,

W(y

1

, y

2

) is never zero (the Wronskian is either always 0 or never 0). But if

y

1

(x) = 0 for innitely many x in J, by Lemma 5.1, there exists an x

0

J such

that y

1

(x

0

) = 0 and y

1

(x

0

) = 0. Therefore,

W(y

1

, y

2

)(x

0

) =

y

1

(x

0

) y

2

(x

0

)

y

1

(x

0

) y

2

(x

0

)

0 y

2

(x

0

)

0 y

2

(x

0

)

= 0.

This is a contradiction. Therefore, y

1

(x) = 0 for at most nitely many x in J.

Let y

1

a solution to y

+ P(x)y

J. We want to nd a linearly independent second solution y

2

.

We break J into a nite collection of subintervals so that y

1

(x) ,= 0 on the

interior of each subinterval. We will describe how to nd the second solution y

2

Note that f

is dieren-

tiable.

5.1. REDUCTION OF ORDER 59

on the interior of each interval. This will dene y

2

throughout J except for a

few isolated points.

Consider a subinterval I on which y

1

(x) is never zero. So y

1

(x) > 0 or

y

2

(x) < 0 throughout I. We have

W(y

1

, y

2

) =

y

1

y

2

y

1

y

.

If y

2

is also a solution to y

+ P(x)y

+ Q(x)y = 0, then W = Ae

R

P(x) dx

for

some constant A (Albels formula).

y

1

y

2

y

1

y

2

= e

R

P(x) dx

,

where y

1

is known. This is a linear rst order equation for y

2

. In standard form,

we have

y

2

y

1

y

1

y

2

=

e

R

P(x) dx

y

1

,

where y

1

,= 0 throughout I. The integrating factor is

e

R

y

1

y

1

= e

ln(|y1|)+C

=

C

[y

1

[

,

where C

1

. We use 1/y

1

. Therefore,

y

2

y

1

y

2

1

y

2

=

e

R

P(x) dx

y

2

1

,

y

2

y

1

=

_

e

R

P(x) dx

y

2

1

dx,

nally giving us

y

2

= y

1

_

e

R

P(x) dx

y

2

1

dx. (5.2a)

This denes y

2

on J except at a few isolated points where y

1

= 0. At those

points, we have

y

1

y

2

y

1

y

2

= e

R

P(x) dx

,

so

y

2

=

e

R

P(x) dx

y

1

. (5.2b)

This denes y

2

are the boundary points of the subintervals.

2

s.

60 CHAPTER 5. SECOND ORDER LINEAR EQUATIONS

Example 5.3. Observe that y = x is a solution of

y

x + 2

x

y

+

x + 2

x

2

y = 0

and solve

y

x + 2

x

y

+

x + 2

x

2

y = xe

x

.

1

= x, we have

y

2

= x

_

e

R

(

x+2

x

)dx

x

2

dx = x

_

e

R

(1+

2

x

)dx

x

2

dx = x

_

e

x+2 ln(|x|)

x

2

dx

= x

_

e

x

e

2 ln(|x|)

x

2

dx = x

_

e

x

[x[

2

x

2

dx = x

_

e

x

dx = xe

x

.

Therefore,

W =

x xe

x

1 e

x

+xe

x

= xe

x

+x

2

e

x

xe

x

= x

2

e

x

.

From this, we have

v

1

=

y

2

R

W

=

xe

x

xe

x

x

2

e

x

= e

x

= v

1

= e

x

,

v

2

=

y

1

R

W

=

xxe

x

x

2

e

x

= 1 = v

2

= x.

Therefore,

y

p

= v

1

y

1

+v

2

y

2

= xe

x

+x

2

e

x

,

and the general solution is

y = C

1

x +C

2

xe

x

xe

x

+x

2

e

x

= C

1

x +C

2

xe

x

+x

2

e

x

, x ,= 0,

where C

1

, C

2

, and C

2

are arbitrary constants.

5.2 Undetermined Coecients

Suppose that y

1

and y

2

are solutions of y

+ P(x)y

method for nding y

p

is called variation of parameters. We will rst consider

another method, the method of undetermined coecients, which does not always

work, but it is simpler than variation of parameters in cases where it does work.

5.2. UNDETERMINED COEFFICIENTS 61

The method of undetermined coecients has two requirements, that

1. we have constant coecients, i.e., P(x) = p and Q(x) = q for some con-

stants p and q.

2. R(x) is nice.

Example 5.4. Solve y

y = e

2x

.

Solution. We have

2

1 = 0. So y

1

= e

x

and y

2

= e

x

. Since R(x) = e

2x

, we

look for a solution of the form y = Ae

2x

. Then

y

= 2Ae

2x

, y

= 4Ae

2x

.

Then according to the dierential equation, we have

4Ae

2x

Ae

2x

. .

3Ae

2x

= e

2x

= A =

1

3

.

Therefore, the general solution is

y = c

1

e

x

+c

2

e

x

+

1

3

e

2x

,

where c

1

and c

2

are arbitrary constants.

Undetermined coecients works when P(x) and Q(x) are constants and

R(x) is a sum of the terms of the forms shown in Table 5.1.

Table 5.1: The various forms of y

p

ought to be given a particular R(x).

R(x) Form of y

p

Cx

n

A

0

x

n

+A

1

x

n1

+ +A

n1

x +A

n

Ce

rx

Ae

rx

Ce

rx

cos(kx)

Ce

rx

sin(kx)

Ae

rx

cos(kx) +Be

rx

sin(kx)

The success of undetermined coecients depends on the derivative of R(x)

having some basic form as R(x). There is an easy way to deal with a complica-

tion: whenever a term in the trial y

p

is already part of the general solution to the

homogeneous equation, multiply it by x. For example, if we have y

y = e

x

,

as before, we have y

1

= e

x

and y

2

= e

2x

. Thus, we try y

p

= Axe

2x

and nd A.

62 CHAPTER 5. SECOND ORDER LINEAR EQUATIONS

Example 5.5. Solve y

+ 3y

+ 2y = e

x

3.

Solution. We have

2

+ 3 + 2 = 0. Factoring gives ( + 2) ( + 1) = 0, so

1, 2, and we have y

1

= e

x

and y

2

= e

2x

. Let y = Ae

x

+ B. Then

y

= Ae

x

and y

= Ae

x

, giving us

y

+ 3y

+ 2y = Ae

x

+ 3Ae

x

+ 2Ae

x

+ 2B = 6Ae

x

+ 2B.

Comparing coecients with e

x

3, we have

6A = 1 = A =

1

6

,

2B = 3 = B =

3

2

.

Therefore, y

p

= e

x

/6 3/2, and we have

y = c

1

e

x

+c

2

e

2x

+

e

x

6

3

2

,

where c

1

and c

2

are arbitrary constants.

Example 5.6. Solve y

+y = sin(x).

Solution. We have

2

+ 1 = 0, so = i, and it follows that y

1

= cos(x) and

y

2

= sin(x). Let y = Axsin(x) +Bxcos(x). Then

y

= (ABx) sin(x) + (Ax +B) cos(x),

y

= (Ax 2B) sin(x) + (2ABx) cos(x).

Therefore,

y

Comparing coecients with sin(x), we see that

2B = 1 = B =

1

2

,

2A = 0 = A = 0.

5.2. UNDETERMINED COEFFICIENTS 63

Therefore, y

p

= xcos(x)/2, and the general solution is

y = c

1

sin(x) +c

2

cos(x)

xcos(x)

2

,

where c

1

and c

2

are arbitrary constants.

Example 5.7. Solve y

4y = x

2

.

Solution. We have

2

4 = 0, so = 2, and it follows that y

1

= e

2x

and

y

2

= e

2x

. Let y = Ax

2

+Bx+C. Then y

= 2Ax+B and y

= 2A. Therefore,

y

4y = 2A 4Ax

2

4B 4C = 4Ax

2

+ 0x + (2A 4B 4C) .

Comparing coecients with x

2

, we immediately see that

4A = 1 = A =

1

4

,

4B = 0 = B = 0,

2A 4C = 0 = C =

A

2

=

1/4

2

=

1

8

.

Therefore, y

p

= x

2

/4 1/8, and the general solution is

y = c

1

e

2x

+c

2

e

2x

x

2

4

1

8

,

where c

1

and c

2

are arbitrary constants.

Example 5.8. Solve y

= x.

Solution. We have

2

= 0, so 1, 0, and it follows that y

1

= e

x

and

y

2

= 1. Let y = Ax

2

+Bx. Then y

= 2Ax +B and y

= 2A. Therefore,

y

y = 2A 2Ax B.

Comparing coecients with x, we immediately see that

2A = 1 = A =

1

2

,

2AB = 0 = B = 1.

Therefore, y

p

= x

2

/2 x, and the general solution is

y = c

1

e

x

+c

2

x

2

2

x,

64 CHAPTER 5. SECOND ORDER LINEAR EQUATIONS

where c

1

and c

2

are arbitrary constants.

5.2.1 Shortcuts for Undetermined Coecients

Suppose we have

y

+ay

+by = R(x).

Let p() =

2

+a +b. We now consider two cases for R(x).

Case I: R(x) = ce

x

. It follows that y

p

= Ae

x

, so y

= Ae

x

and

y

= Ae

x

. Then

y

+ay

+by = ce

x

_

2

+a +b

_

Ae

x

= p()Ae

x

.

Therefore, A = c/p(), unless p() = c.

What if p() = 0? Then e

x

is a solution to the homogeneous equation. Let

y = Axe

x

. Then

y

= Ae

x

+Axe

x

,

y

= Ae

x

+Ae

x

+A

2

xe

x

= 2Ae

x

+A

2

xe

x

.

Thus,

y

+y

+y = A

__

2

x +ax +b

_

+ 2 +a

e

x

= A(p()xe

x

+p

()e

x

)

= Ap

()e

x

.

Therefore, A = c/p

(), unless p

() = 0 too.

If p() = 0 and p

() = 0, then e

x

and xe

x

both solve the homogeneous

equation. Letting y = Ax

2

e

x

, we have A = c/p

().

Case II: R(x) = ce

rx

cos(bx) or ce

rx

sin(bx). We use complex numbers. Set

= r +ib. Then

ce

x

= ce

rx

cos(bx) +ice

rx

sin(bx).

Let z = y +iw. Consider the equation

z

+az

+bz = ce

x

. (5.3)

5.2. UNDETERMINED COEFFICIENTS 65

The real part and imaginary parts are

y

+ay

+by = ce

rx

cos(bx), (5.4a)

w

+aw

+bw = ce

rx

sin(bx), (5.4b)

respectively. The solutions of Equation (5.3,5.4a,5.4b) are then

z =

ce

x

p()

, y = Re

_

ce

x

p()

_

, w = Im

_

ce

x

p()

_

,

respectively.

Example 5.9. Solve y

+ 8y

+ 12y = 3 cos(2x).

Solution. We have p() =

2

+ 8 + 12 = ( + 2) ( + 6). Therefore,

2, 6 and it follows that y

1

= e

2x

and y

2

= e

6x

. So

z =

3e

2ix

p(2i)

=

3e

2ix

(2i)

2

+ 16i + 12

=

3e

2ix

4 + 16i + 12

=

3e

2ix

8 + 16i

=

3

8

1

1 + 2i

e

2ix

=

3

8

1 2i

(1 2i) (1 + 2i)

e

2ix

=

3

8

1 2i

1 + 2i

e

2ix

=

3

40

(1 2i) e

2ix

=

3

40

(1 2i) (cos(2x) +i sin(2x))

=

3

40

[cos(2x) + 2 sin(2x) +i (2 cos(2x) + sin(2x))] .

Therefore,

Re(z) =

3

40

(cos(2x) + 2 sin(2x)) , Im(z) =

3

40

(2 cos(x) + sin(2x)) .

Hence, the solution is

y = c

1

e

2x

+c

2

e

6x

+

3

40

(cos(2x) + 2 sin(2x)) ,

where c

1

and c

2

are arbitrary constants.

We conclude that there is no shortcut when R(x) is a polynomial or any

multiple of a polynomial.

66 CHAPTER 5. SECOND ORDER LINEAR EQUATIONS

5.3 Variation of Parameters

Consider the equation

y

+P(x)y

where P(x) and Q(x) are not necessarily constant. Let y

1

and y

2

be linearly

independent solutions of the auxiliary equation

y

+P(x)y

+Q(x)y = 0. (H)

So for any constants C

1

and C

2

, C

1

y

1

+C

2

y

2

satises Equation (H).

The idea of variation of parameters is to look for functions v

1

(x) and v

2

(x)

such that v

1

y

1

+ v

2

y

2

satises Equation (I). Let y

1

and y

2

. If we want y to

be a solution to Equation (I), then substitution into Equation (I) will give one

condition on v

1

and v

2

.

Since there is only one condition to be satised and two functions to be

found, there is lots of choice, i.e., there are many pairs v

1

and v

2

which will

do. So we will arbitrarily decide only to look for v

1

and v

2

which also satisfy

v

1

y

1

+ v

2

y

2

= 0. We will see that this satisfaction condition simplies things.

So consider

y = v

1

y

1

+v

2

y

2

,

y

= v

1

y

1

+v

2

y

2

+v

1

y

1

+v

2

y

2

= v

1

y

1

+v

2

y

2

,

y

= v

1

y

1

+v

2

y

2

+v

1

y

1

+v

2

y

2

.

To solve Equation (I) means that

_

v

1

y

1

+v

2

y

2

+v

1

y

1

+v

2

y

2

+P(x)v

1

y

1

+P(x)v

2

y

2

+Q(x)v

1

y

1

+Q(x)v

2

y

2

_

= R(x),

_

_

_

v

1

(y

1

+P(x)y

1

+Q(x)y

1

)

+v

2

(y

2

+P(x)y

2

+Q(x)y

2

)

+v

1

y

1

+v

2

y

2

_

_

_ = R(x).

Therefore,

v

1

y

1

+v

2

y

2

= R(x),

v

1

y

1

+v

2

y

2

= 0,

5.3. VARIATION OF PARAMETERS 67

and it follows that

v

1

=

y

2

R

W

, v

2

=

y

1

R

W

,

where

W =

y

1

y

2

y

1

y

,= 0,

i.e., it is never zero regardless of x, since y

1

and y

2

are linearly independent.

Example 5.10. Solve y

5y

+ 6y = x

2

e

3x

.

Solution. The auxiliary equation is

y

5y

+ 6y = 0

2

5 + 6 = 0 = = 3, 2.

Therefore, the solutions are y

1

= e

3x

and y

2

= e

2x

. Now,

W =

y

1

y

2

y

1

y

e

3x

e

2x

3e

3x

2e

2x

= 2e

5x

3e

5x

= e

5x

.

Therefore,

v

1

=

y

2

R

W

=

e

2x

x

2

e

3x

e

5x

= x

2

= v

1

=

x

3

3

.

We can use any function v

1

with the right v

1

, so choose the constant of integra-

tion to be 0.

For v

2

, we have

v

2

=

y

1

R

W

=

e

3x

x

2

e

3x

e

5x

= x

2

e

x

,

so

v

2

=

_

x

2

e

x

dx

. .

by parts

= x

2

e

x

+ 2

_

xe

x

dx

. .

by parts

= x

2

e

x

+ 2xe

x

2

_

e

x

dx

= x

2

e

x

+ 2xe

x

2e

x

=

_

x

2

2x + 2

_

e

x

.

Finally,

y

p

=

x

3

3

e

3x

_

x

2

2x + 2

_

e

x

e

2x

= e

3x

_

x

3

3

x

2

+ 2x 2

_

,

68 CHAPTER 5. SECOND ORDER LINEAR EQUATIONS

and the general solution is

y = C

1

e

3x

+C

2

e

2x

+e

2x

_

x

3

3

x

2

+ 2x 2

_

= e

3x

_

x

3

3

x

2

+ 2x 2 +C

1

_

+C

2

e

2x

=

_

x

3

3

x

2

+ 2x +C

1

_

e

3x

+C

2

e

2x

,

where C

1

and C

2

are arbitrary constants.

Example 5.11. Solve y

+ 3y

+ 2y = sin(e

x

).

Solution. We immediately have

2

+ 3 + 2 = 0, which gives 1, 2.

Therefore, y

1

= e

x

and y

2

= e

2x

. Now,

W =

e

x

e

2x

e

x

2e

2x

= 2e

3x

+e

3x

= e

3x

.

Therefore,

v

1

=

e

2x

sin(e

x

)

e

3x

= e

x

sin(e

x

) = v

1

= cos(e

x

) .

For v

2

, we have

v

2

=

e

x

sin(e

x

)

e

3x

= e

2x

sin(e

x

) .

Therefore,

v

2

=

_

e

2x

sin(e

x

) dx

. .

substitute t = e

x

=

_

t sin(t) dt

. .

by parts

=

_

t cos(t) +

_

cos(t) dt

_

= (t cos(t) + sin(t))

= e

x

cos(e

x

) sin(e

x

) .

5.3. VARIATION OF PARAMETERS 69

Finally,

y

p

= v

1

y

1

+v

2

y

2

= e

x

cos(e

x

) +e

2x

(e

x

cos(e

x

) sin(e

x

))

= e

x

cos(e

x

) +e

x

cos(e

x

) e

2x

sin(e

x

)

= e

2x

sin(e

x

) ,

and the general solution is

y = C

1

e

x

+C

2

e

2x

e

2x

sin(e

x

) ,

where C

1

and C

2

are arbitrary constants.

70 CHAPTER 5. SECOND ORDER LINEAR EQUATIONS

Chapter 6

Applications of Second

Order Dierential

Equations

We now look specically at two applications of rst order des. We will see that

they turn out to be analogs to each other.

6.1 Motion of Object Hanging from a Spring

Figure 6.1 shows an object hanging from a spring with displacement d.

d

Figure 6.1: An object hanging from a spring with displacement d.

The force acting is gravity, spring force, air resistance, and any other external

71

72CHAPTER 6. APPLICATIONS OF SECONDORDER DIFFERENTIAL EQUATIONS

forces. Hookes Law states that

F

spring

= kd, (6.1)

where F is the spring force, d is the displacement from the spring to the natural

length, and k > 0 is the spring constant. As we lower the object, the spring

force increases. There is an equilibrium position at which the spring and grav-

ity cancel. Let x be the distance of the object from this equilibrium position

measured positively downwards. Let s be the displacement of the spring from

the natural length at equilibrium. Then it follows that ks = mg.

If the object is located at x, then the forces are

1. gravity.

2. spring.

3. air resistance.

4. external.

Considering just gravity and the spring force, we have

F

gravity

+F

spring

= mg k (x +s) = kx.

Suppose air resistance is proportional to velocity. Then

F

air

= rv, r > 0.

In other words, the resistance opposes motion, so the force has the opposite sign

of velocity. We will suppose that any external forces present depend only on

time but not on position, i.e., F

external

= F(t). Then we have

F

total

. .

ma

= kx rv +F(t),

so it follows that

m

d

2

x

dt

2

+r

dx

dt

+kx = F(t). (6.2)

Therefore, m

2

+r +k = 0 and we have

=

r

r

2

4mk

2m

=

r

2m

_

_

r

2m

_

2

k

m

.

Intuitively, when x = 0, they balance. Changing x creates a force in the opposite direction

attempting to move the object towards the equilibrium position.

6.1. MOTION OF OBJECT HANGING FROM A SPRING 73

We now have three cases we need to consider.

Case 1: 4mk > r

2

. Let = r/2m, let =

_

k/m, and let

=

_

k

m

_

r

2m

_

2

=

_

2

.

Then = i and

x = e

t

(C

1

cos(t) +C

2

sin(t)) = Ae

t

cos(t ), (6.3)

where C

1

and C

2

are arbitrary constants. Figure 6.2a shows how this solution

qualitatively looks like.

Damped vibrations is the common case. The oscillations die out in time,

where the period is 2/. If r = 0, then = 0. In such a case, there is no

resistance, and it oscillates forever.

Case 2: 4mk = r

2

. In such a case, we simply have

x = e

r

2m

t

(C

1

+C

2

t) , (6.4)

where C

1

and C

2

are arbitrary constants. Figure 6.2b shows how this solution

qualitatively looks like.

Case 3: 4mk < r

2

. Let a = r/2m and

b =

_

_

r

2m

_

k

m

.

Then the solution is

x = C

1

e

(ab)t

+C

2

e

(a+b)t

, (6.5)

where C

1

and C

2

are arbitrary constants. This is the overdamped case. The

resistance is so great that the object does not oscillate (imagine everything

immersed in molasses). It will just gradually return to the equilibrium position.

Figure 6.2c shows how this solution qualitatively looks like.

Consider Equation (6.2) again. We now consider special cases of F(t), where

F(t) =

_

_

_

F

0

cos(t),

F

0

sin(t).

74CHAPTER 6. APPLICATIONS OF SECONDORDER DIFFERENTIAL EQUATIONS

t

(a) 4mk > r

2

t

(b) 4mk = r

2

t

(c) 4mk < r

2

Figure 6.2: The various cases for a spring-mass system.

Use undetermined coecients to get the solution of the form

x

p

= Bcos(t) +C sin(t).

For F(t) = F

0

cos(t), we get

x

p

=

F

0

(k m

2

)

2

+ (r)

2

[(k m) cos(t) +r sin(t)] = Acos(t ) ,

where

A =

F

0

_

(k m

2

)

2

+ (r)

2

,

cos() =

k m

2

_

(k m

2

)

2

+ (r)

2

,

sin() =

r

_

(k m

2

)

2

+ (r)

2

.

For F(t) = F

0

sin(t), we get x

p

= Asin(t ), where A and is as above.

6.2. ELECTRICAL CIRCUITS 75

In any solution

e

t

() +Asin(t ) ,

the rst term eventually dies out, leaving Asin(t ) as the steady state so-

lution.

Here we consider a phenomenon known as resonance. The above assumes

that cos(t) and sin(t) are not part of the solution to the homogeneous equa-

tion. Consider now

m

d

2

x

dt

2

+kx = F

0

cos(t),

where =

_

k/m (the external force frequency coincides with the natural

frequency). Then the solution is

x

p

=

F

0

2

km

t sin(t).

Another phenomenon we now consider is beats. Consider r = 0 with near

but not equal to

_

k/m. Let =

_

k/m. Then the solution, with x(0) = 0 =

x

(0), is

x =

F

0

m(

2

2

)

cos(t t)

=

2F

0

( +) ( )

sin

_

( ) t

2

_

sin

_

( + ) t

2

_

F

0

2m

sin(t) sin(t),

where = ( ) /2.

6.2 Electrical Circuits

Consider the electrical circuit shown in Figure 6.3.

Let Q(t) be the charge in the capacitor at time t (Coulombs). Then dT/dt

is called the current, denoted I. The battery produces a voltage (potential

dierence) resulting in current I when the switch is closed. The resistance R

results in a voltage drop of RI. The coil of wire (inductor) produces a magnetic

eld resisting change in the current. The voltage drop created is L(dI/dt). The

F

0

2

km

t,

which increases with time.

76CHAPTER 6. APPLICATIONS OF SECONDORDER DIFFERENTIAL EQUATIONS

battery

resistor

switch

capacitor

inductor I amperes

+

Figure 6.3: An electrical circuit, where resistance is measured in Ohms, capac-

itance is measured in Farads, and the inductance is measured in Henrys.

capacitor produces a voltage drop of Q/C. Unless R is too large, the capacitor

will create sine and cosine solutions and, thus, an alternating ow of current.

Kirkhos Law states that the sum of the voltage changes around a circuit is

zero, so

E(t) +RI +L

dI

dt

+

Q

C

,

so we have

L

d

2

Q

dt

2

+R

dQ

dt

+

Q

C

= E(t). (6.6)

The solution is the same as in the case of a spring. The analogs are shown in

Table 6.1.

Table 6.1: Analogous terms between a spring-mass system and an electrical

circuit.

Circuit Spring

Charge Q Displacement x

Inductance I Mass m

Resistance R Friction r

Capacitance inverse 1/C Spring constant k

Voltage generated by battery E(t) External force F(t)

Amperes I Velocity v

An example of this application is choosing a station on an old-fashioned

radio. The radio has a capacitor with two charged metal bars. When the

user turns the tuning dial, it changes the distance between the bars, which in

6.2. ELECTRICAL CIRCUITS 77

turn changes the capacitance C. This changes the frequency of the solution

of the homogeneous equation. When it agrees with the frequency E(t) of some

incoming signal, resonance results so that the amplitude of the current produced

from this signal is much greater than any other.

78CHAPTER 6. APPLICATIONS OF SECONDORDER DIFFERENTIAL EQUATIONS

Chapter 7

Higher Order Linear

Dierential Equations

In this section we generalize some of the techniques for solving second order

linear equations discussed in Chapter 5 so that they apply to linear equations

of higher order. Recall from Chapter 4 that an nth order linear dierential

equation has the form

y

(n)

+P

1

(x)y

(n1)

+P

2

(x)y

(n2)

+ +P

n

(x)y = Q(x). (7.1)

The general solution is

y = C

1

y

1

+C

2

y

2

+ +C

n

y

n

+y

p

,

where y

1

, y

2

, . . . , y

n

are linearly independent solutions to the auxiliary homoge-

neous equation and y

p

is a solution of Equation (7.1).

So given y

1

, y

2

, . . . , y

n

, how do we nd y

p

? We again consider the two meth-

ods we have looked at in 5.2 and 5.3 (p. 60 and p. 66, respectively).

7.1 Undetermined Coecients

If P

1

(x), P

2

(x), . . . , P

n

(x) are constants and R(x) is nice, we can use undeter-

mined coecients.

Example 7.1. Solve y

(4)

y = 2e

2e

+ 3e

x

.

79

80CHAPTER 7. HIGHER ORDER LINEAR DIFFERENTIAL EQUATIONS

Solution. We have

p() =

4

1 = ( 1) ( + 1)

_

2

+ 1

_

.

Therefore,

y

1

= e

x

, y

2

= e

x

, y

3

= cos(x), y

4

= sin(x).

Let y

p

= Ae

3x

+Bxe

3x

. Then by the short cut method (5.2.1, p. 64), we have

A =

2

p(2)

=

2

16 1

=

2

15

,

B =

3

p

(1)

=

3

4 1

3

=

3

4

.

Therefore, the general solution is

y = C

1

e

x

+C

2

e

x

+C

3

cos(x) +C

4

sin(x) +

2

15

e

2x

+

3

4

xe

x

.

Suppose that y

1

, y

2

, . . . , y

n

are solutions of the auxiliary homogeneous equation.

We look for solutions of Equation (7.1) of the form

y = v

1

y

1

+v

2

y

2

+ +v

n

y

n

,

where v

1

, v

2

, . . . , v

n

are functions of x. One condition on v

1

, v

2

, . . . , v

n

is given

by substituting into Equation (7.1). We choose n 1 conditions to be

v

1

y

1

+v

2

y

2

+ +v

n

y

n

= 0, (1)

v

1

y

1

+v

1

y

2

+ +v

n

y

n

= 0, (2)

.

.

.

.

.

.

v

1

y

(n2)

1

+v

2

y

(n2)

2

+ +v

n

y

(n2)

n

= 0. (n 1)

7.2. VARIATION OF PARAMETERS 81

If conditions (1), . . . , (n 1) are satised, then

y

=

n

i=1

(v

i

y

i

)

=

n

i=1

v

i

y

i

+

n

i=1

v

i

y

i

=

n

i=1

v

i

y

i

,

y

=

n

i=1

(v

i

y

i

)

=

n

i=1

v

i

y

i

+

n

i=1

v

i

y

i

=

n

i=1

v

i

y

i

,

.

.

.

y

(n1)

=

n

i=1

v

i

y

(n1)

i

+

n

i=1

v

i

y

(n2)

=

n

i=1

v

i

y

(n1)

i

,

y

(n)

=

n

i=1

v

i

y

(n)

i

+

n

i=1

v

i

y

(n1)

i

.

Therefore,

y

(n)

+P

1

(x)y

(n1)

+ +P

n

(x)y =

n

i=1

v

i

y

(n)

i

+

n

i=1

v

i

y

(n1)

i

+P

1

(x)

n

i=1

v

i

y

(n1)

i

+ +P

n

(x)

n

i=1

v

i

y

i

=

n

i=1

v

i

y

(n1)

i

+

n

i=1

v

i

_

y

(n)

i

+P

1

(x)y

(n1)

i

+

+P

n

(x)y

i

_

=

n

i=1

v

i

y

(n1)

i

+

i=1

v

i

0

=

n

i=1

v

i

y

(n1)

i

.

Therefore, Equation (7.1) becomes the rst condition

n

i=1

v

i

y

(n1)

i

y = R(x). (n)

82CHAPTER 7. HIGHER ORDER LINEAR DIFFERENTIAL EQUATIONS

Conditions (1), . . . , (n 1) can be written as

M

_

_

v

1

v

2

.

.

.

v

n

_

_

=

_

_

0

0

.

.

.

R(x)

_

_

,

where

M =

_

_

y

1

y

n

y

1

y

n

.

.

.

.

.

.

.

.

.

y

(n1)

1

y

(n1)

n

_

_

.

Therefore, [M[ = W ,= 0.

7.3 Substitutions: Eulers Equation

Consider the equation

x

n

y

(n)

+a

1

x

n1

y

(n1)

+ +a

n1

y

+a

n

y = R(x), (7.2)

where a

1

, a

2

, . . . , a

n

are constants. Consider x > 0. Let x = e

u

so that u = ln(x)

(for x < 0, we must use u = ln(x)). Then

dy

dx

=

dy

du

du

dx

=

dy

du

1

x

,

d

2

y

dx

2

=

d

2

y

du

2

1

x

2

dy

du

1

x

2

=

_

d

2

y

du

2

dy

du

_

1

x

2

,

d

3

y

dx

3

=

_

d

3

ydx

3

d

2

y

du

2

_

1

x

3

2

_

d

2

y

du

2

dy

du

_

1

x

3

=

_

d

3

y

du

3

3

d

2

y

du

2

+ 2

dy

du

_

1

x

3

,

.

.

.

d

n

y

dx

n

=

_

d

n

y

du

n

+C

1

d

n1

y

du

n1

+ +C

n

dy

du

_

1

x

n

7.3. SUBSTITUTIONS: EULERS EQUATION 83

for some constants C

1

, C

2

, . . . , C

n

. Therefore, Equation (7.2) becomes

d

n

y

du

n

+b

1

d

n1

y

du

n1

+ +b

n1

dy

du

+b

n

y = R(e

u

)

for some constants b

1

, b

2

, . . . , b

n

.

Example 7.2. Solve x

2

y

+xy

y = x

3

for x > 0.

Solution. Let x = e

u

so that u = ln(x). Then

dy

dx

=

dy

du

du

dx

=

dy

du

1

x

,

d

2

y

dx

2

=

d

2

y

du

2

1

x

2

dy

du

1

x

2

=

_

d

2

y

du

2

dy

du

_

1

x

2

.

Therefore,

_

d

2

y

du

2

dy

du

_

+

dy

du

y = e

3u

,

d

2

u

du

2

y = e

3u

.

Considering the homogeneous case, we have

2

1 = 0 = 1. Therefore,

y

1

= e

u

and y

2

= e

u

. Using undetermined coecients, let y

p

= Ae

3u

. Then

y

= 3Ae

3u

and y

= 9Ae

3u

. Substitution gives

9Ae

3u

Ae

3u

. .

8Ae

3u

= e

3u

.

We can immediately see that A = 1/8. Therefore, the general solution is

y = C

1

e

u

+C

2

e

u

+

1

8

e

3u

= C

1

e

ln(x)

+C

2

e

ln(x)

+

1

8

e

3 ln(x)

= C

1

x +

C

2

x

+

x

3

8

,

where C

1

and C

2

are arbitrary constants.

We can also consider a direct method. Solving

d

2

y

dx

2

+p

dy

du

+qy = 0

84CHAPTER 7. HIGHER ORDER LINEAR DIFFERENTIAL EQUATIONS

involves looking for solutions of the form y = e

u

= (e

u

)

x = e

u

, we could directly look for solutions of the form x

With this in mind and considering Example 7.2 again, we have

y = x

,

y

= x

1

,

y

= ( 1) x

2

.

Therefore,

x

2

y

+xy

y = 0 = ( 1) x

+x

= 0

= ( 1) + 1 = 0

=

2

1 = 0

= = 1,

as before.

Chapter 8

Power Series Solutions to

Linear Dierential

Equations

8.1 Introduction

Consider the equation

y

(n)

+P

1

(x)y

(n1)

+ +P

n

(x)y = R(x). (8.1)

Given solutions y

1

, . . . , y

n

to the homogeneous equation

y

(n)

+P

1

(x)y + +P

n

y = 0,

we know how to solve Equation (8.1), but so far there is no method of nding

y

1

, . . . , y

n

unless P

1

, . . . , P

n

are constants. In general, there is no method of

nding y

1

, . . . , y

n

exactly, but we can nd the power series approximation.

Example 8.1. Solve y

xy

2y = 0.

Solution. Let

y = a

0

+a

1

x +a

2

x

2

+ +a

n

x

n

+ =

n=0

a

n

x

n

.

85

86CHAPTER 8. POWER SERIES SOLUTIONS TOLINEAR DIFFERENTIAL EQUATIONS

Then it follows that

y

= a

1

+ 2a

2

x + +na

n

x

n1

+ (n + 1) a

n+1

x

n

+

=

n=1

na

n

x

n1

=

k=0

(k + 1) a

k+1

x

k

,

where k = n 1. Relabeling, we have

y

n=0

(n + 1) a

n+1

x

n

.

By the same procedure, we have

y

n=0

n(n + 1) a

n+1

x

n1

=

n=0

(n + 1) (n + 2) a

n+2

x

n

.

Now considering y

xy

2y = 0, we have

y

..

n=0

(n + 1) (n + 2) a

n+2

x

n

x

y

..

n=0

na

n

x

n1

2

y

..

n=0

a

n

x

n

= 0,

n=0

[(n + 1) (n + 2) a

n+2

na

n

2a

n

] x

n

= 0.

This gives us the recurrence relation

(n + 1) (n + 2) a

n+2

na

n

2a

n

= 0,

(n + 1) (n + 2) a

n+2

(n + 2) a

n

= 0,

(n + 1) a

n+2

= a

n

,

a

n+2

=

a

n

n + 1

for all n 0. Therefore,

a

2

=

a

0

1

, a

4

=

a

2

3

=

a

0

1 3

, a

6

=

a

4

5

=

a

0

1 3 5

, . . . , a

2n

=

a

0

n

k=1

(2k 1)

,

a

3

=

a

1

2

, a

5

=

a

3

4

=

a

1

2 4

, a

7

=

a

5

6

=

a

1

2 4 6

, . . . , a

2n+1

=

a

1

n

k=1

2k

.

8.1. INTRODUCTION 87

Therefore,

y = a

0

+a

1

x +a

2

x

2

+a

3

x

3

+ +a

n

x

n

+

= a

0

+a

1

x +a

0

x

2

+

a

1

2

x

3

+ +

a

0

n

k=1

(2k 1)

x

2n

+

a

1

n

k=1

2k

x

2n+1

+

= a

0

_

1 +x

2

+ +

x

2n

n

k=1

(2k 1)

+

_

+a

1

_

x +

x

3

2

+ +

x

2n+1

n

k=1

2k

+

_

= a

0

n=0

x

2n

n

k=1

(2k 1)

+a

1

n=0

x

2n+1

2

n

n!

.

That is, we have y = C

1

y

1

+C

2

y

2

, where C

1

= a

0

and C

2

= a

1

such that

y

1

=

n=0

x

2n

n

k=1

(2k 1)

, y

2

=

n=0

x

2n+1

2

n

n!

.

Can we recognize y

1

and y

2

as elementary functions? In general, no. But in

this case, we have

y

2

=

n=0

x

2n+1

2

n

n!

= x

n=0

x

2n

2

n

n!

= x

n=0

_

x

2

_

n

2

n

n!

= x

n=0

_

x

2

/2

_

n

n!

= xe

x

2

/2

.

Having one solution in closed form, we can try to nd another (see 5.1, p. 57)

by

y

1

= y

2

_

e

R

P(x) dx

y

2

2

= xe

x

2

/2

_

e

R

P(x) dx

x

2

e

x

2

dx

= xe

x

2

/2

_

e

x

2

/2

x

2

e

x

2

dx = xe

x

2

/2

_

1

x

2

e

x

2

/2

dx,

which is not an elementary function.

88CHAPTER 8. POWER SERIES SOLUTIONS TOLINEAR DIFFERENTIAL EQUATIONS

8.2 Background Knowledge Concerning Power

Series

Theorem 8.2

Consider the power series

n=0

a

n

(z p)

n

.

1. There exists an R such that 0 R < called the radius of con-

vergence such that

n=0

a

n

(z p)

n

converges for [z p[ < R and

diverges for [z p[ > R. The radius of convergence is given by

R = lim

n

a

n

a

n+1

(8.2)

if the limit exists (ratio test).

2. The function f(z) =

n=0

a

n

z

n

dened for [z p[ < R can be dier-

entiated and integrated term-wise within its radius of convergence.

Proof. Proof is given in MATA37 or MATB43.

Remark. The series may or may not converge when [z p[ = R.

Using (2) of Theorem 8.2, f

(z) =

n=0

na

n

z

n1

.

Similarly,

f(z) =

n=0

a

n

(z p)

n

, z = p = a

0

= f(p),

f

(z) =

n=0

na

n

(z p)

n1

, z = p = a

1

= f

(p),

f

(z) =

n=0

n(n 1) a

n

(z p)

n2

, z = p = a

2

=

f

(p)

2

,

.

.

.

f

(n)

(z) =

n=0

n! a

n

, z = p = a

n

=

f

(n)

(p)

n!

.

8.3. ANALYTIC EQUATIONS 89

Conversely, if f is innitely dierentiable at p, we dene a

n

= f

(n)

(p)/n!. Then

n=0

a

n

(z p)

n

is called the Taylor series of f about p.

From the above, if any power series converges to f near p, it must be its

Taylor series. But

1. the Taylor series might not converge.

2. if it does converge, it does not have to converge to f.

Denition (Analytic function, ordinary/singular point)

A function f is called analytic at p if

n=0

a

n

(z p)

n

converges to f(z) in

some neighbourhood of p (i.e., if there exists an r > 0 such that it converges for

[z p[ < r). If f is analytic at p, then p is called an ordinary point of f. If f

is not analytic at p, then p is called a singular point of f.

Theorem 8.3 (Theorem from complex analysis)

Let f(z) be analytic at p. Then the radius of convergence of the Taylor

series of f at p equals the distance from p to the closest singularity of f in

the complex plane.

Proof. Proof is given in MATC34.

Example 8.4. Consider f(x) = 1/

_

x

2

+ 1

_

. The singularities are i. There-

fore, the radius of convergence of the Taylor series about the origin is 1, with

the series given by

f(x) =

n=0

(1)

n

x

2n

.

The radius of convergence of the Taylor series about x = 1 is

2.

8.3 Analytic Equations

Consider y

+ P(x)y

then p is called an ordinary point of the de. Otherwise, p is called a singular

90CHAPTER 8. POWER SERIES SOLUTIONS TOLINEAR DIFFERENTIAL EQUATIONS

point or a singularity of the de. When looking at a de, do not forget to write

it in the standard form. For example,

x

2

y

+ 5y

+xy = 0

has a singularity at x = 0, since the standard form is

y

+

_

5

x

2

_

y

+

_

1

x

_

y = 0.

Theorem 8.5

Let x = p be an ordinary point of y

+ P(x)y

distance from p to the closest singular point of the de in the complex plane.

Then the de has two series y

1

(x) =

n=0

a

n

x

n

and y

2

(x) =

n=0

b

n

x

n

which converge to linearly independent solutions to the de on the interval

[x p[ < R.

Remark. The content of the theorem is that the solutions are analytic within

the specied interval. The fact that solutions exist on domains containing at

least this interval is a consequence of theorems in Chapter 10.

Proof. The proof is given in Chapter 10.

Example 8.6. Consider

_

x

2

+ 1

_

y

+ 2xy

+ 2y = 0. Dividing by x

2

+ 1, we

have

y

+

3x

x

2

+ 1

y

+

2

x

2

+ 1

y = 0.

The singularities are i. If p = 0, then r = [i 0[ = 1. Therefore, analytic

solutions about x = 0 exist for [x[ < 1. If p = 1, then r = [i 1[ =

2. There-

fore, there exist two linearly independent series of the form

n=0

a

n

(x 1)

n

converging to the solutions for at least [x 2[ <

2.

Note that the existence and uniqueness theorem (see Chapter 10) guarantees

solutions from to (since x

2

+ 1 has no real zeros), but only in a nite

subinterval can we guarantee that the solutions are expressible as convergent

power series.

Example 8.7. Consider y

xy

a series solution converging on R exists about any point.

8.4. POWER SERIES SOLUTIONS: LEVELS OF SUCCESS 91

8.4 Power Series Solutions: Levels of Success

When using power series techniques to solve linear dierential equations, ideally

one would like to have a closed form solution for the nal answer, but most often

one has to settle for much less. There are various levels of success, listed below,

and one would like to get as far down the list as one can.

For simplicity, we will assume that we have chosen to expand about the

point x = 0, although the same considerations apply to x = c for any c.

Suppose 0 is an ordinary point of the equation L(x)y

+M(x)y

+N(x)y = 0

and let y =

n=0

a

n

x

n

be a solution. One might hope to:

1. nd the coecients a

n

for a few small values of n; (For example, nd the

coecients as far as a

5

which thereby determines the 5th degree Taylor

approximation to y(x));

2. nd a recursion formula which for any n gives a

n+2

in terms of a

n1

and a

n

;

3. nd a general formula a

n

in terms of n;

4. recognize the resulting series to get a closed form formula for y(x).

In general, level (1) can always be achieved. If L(x), M(x), and N(x)

are polynomials, then level (2) can be achieved. Having achieved level (2), it

is sometimes possible to get to level (3) but not always. Cases like the one

in Example 8.1 where one can achieve level (4) are rare and usually beyond

expectation.

8.5 Level 1: Finding a nite number of coe-

cients

Suppose c is an ordinary point of Equation (8.3). One approach is to nding

power series solutions is as follows. Expand P(x) and Q(x) into Taylor series

P(x) =

n=0

p

n

(x c)

n

, Q(x) =

n=0

q

n

(x c)

n

.

and set y =

n=0

a

n

(x c)

n

solves the equation. Then

y

n=0

na

n

(x c)

n1

, y

n=0

n(n 1) a

n

x

n2

,

92CHAPTER 8. POWER SERIES SOLUTIONS TOLINEAR DIFFERENTIAL EQUATIONS

and Equation (8.3) becomes

y

..

n=0

n(n 1) a

n

x

n2

+

P(x)

..

_

n=0

p

n

(x c)

n

_

y

..

_

n=0

na

n

(x c)

n1

_

+

_

n=0

q

n

(x c)

n

_

. .

Q(x)

_

n=0

a

n

(x c)

n

_

. .

y

= 0.

Since p

n

s and q

n

s are known (as many as desired can be computed from deriva-

tives of P(x) and Q(x)), in theory we can inductively nd the a

n

s one after

another.

In the special case where P(x) and Q(x) are polynomials, we can get a

recursion formula by writing a

n+2

in terms of a

n

and a

n1

(level 2) and, under

favourable circumstances, we can nd the general formula for a

n

(level 3) as in

Example 8.1. But in general, computation gets impossibly messy after a while

and, although a few a

n

s can be found, the general formula cannot.

But if you only want a few a

n

s, there is an easier way. We have a

n

=

y

(n)

(c)/n!, so

a

2

=

y

(c)

2!

=

P(x)y

(c) +Q(c)y(c)

2

=

a

1

P(c) a

0

Q(c)

2

,

a

3

=

y

(c)

3!

=

1

6

(P

(c)y

(c) P(c)y

(c)y(c) Q(c)y

(c))

=

1

6

_

P

(c)a

1

+

a

1

P(c)

2

+a

0

P(c)Q(c)

2

a

0

Q

(c) a

1

Q(c)

_

,

a

4

= ,

and so on.

Example 8.8. Compute the Taylor expansion about 0 as far as degree 4 for

the solution of y

e

7x

y

(0) = 1.

Solution. Rearranging the equation, we have y

= e

5x

y

xy. Therefore

y

= e

5x

y

+ 5e

5x

y

xy

y

y

(4)

= e

5x

y

+ 5e

5x

y

+ 5e

5x

y

xy

= e

5x

y

+ 10e

5x

y

xy

2y

.

8.5. LEVEL 1: FINDING A FINITE NUMBER OF COEFFICIENTS 93

Substituting gives

y

(0) = e

0

y

(0) 0 = 1

y

(0) = e

0

y

(0) + 5e

0

y

(0) 0 y(0) = 1 + 5 2 = 4

y

(4)

(0) = e

0

y

(0) + 10e

0

y

(0) 0 2y

(0) = 4 + 10 2 = 12

Therefore,

y = 2 +x +

x

2

2!

+

4x

3

3!

+

12x

4

4!

+

= 2 +x +

1

2

x

2

+

2

3

x

3

1

2

x

4

+ .

Example 8.9. Find the series solution for y

+ (x + 3) y

2y = 0, where

y(0) = 1 and y

(0) = 2.

Solution. Rearranging the equation, we have y

= (x + 3) y

+ 2y. Then

y

(0) = (0 + 3) y

(0) + 2y(0) = 3 2 + 2 1 = 4,

y

= y

(x + 3) y

+ 2y

= (x + 3) y

+y

,

y

(0) = 3y

(0) +y

y

(4)

= = 34.

Therefore,

y = 1 + 2x

4x

2

2!

+

14x

3

3!

34x

4

4!

+

= 1 + 2x 2x

2

+

7

3

x

3

17

12

x

4

+ .

In this case, since the coecients were polynomials, it would have been possible

to achieve level 2, (although the question asked only for a level 1 solution).

However even had we worked out the recursion formula (level 2) it would have

been too messy to use it to nd a general formula for a

n

.

94CHAPTER 8. POWER SERIES SOLUTIONS TOLINEAR DIFFERENTIAL EQUATIONS

8.6 Level 2: Finding the recursion relation

Example 8.10. Consider

_

x

2

5x + 6

_

y

5y

y

3

and determine the lower bound

on its radius of convergence from the recursion formula.

Solution. We have

2y =

n=0

2a

n

x

n

,

5y

n=0

5na

n

x

n1

=

n=0

5 (n + 1) a

n+1

x

n

=

n=0

5 (n + 1) a

n+1

x

n

,

6y

n=0

6n(n 1) a

n

x

n2

=

n=0

6 (n + 2) (n + 1) a

n+2

x

n

,

5xy

n=0

5n(n 1) a

n

x

n1

=

n=0

5 (n + 1) na

n+1

x

n

,

x

2

y

n=0

n(n 1) a

n

x

n

.

Therefore,

n=0

_

_

_

_

_

n(n 1) a

n

5 (n + 1) na

n+1

+6 (n + 2) (n + 1) a

n+2

5 (n + 1) a

n+1

2a

n

_

_

_

_

_

x

n

=

n=0

_

_

_

6 (n + 2) (n + 1) a

n+2

5 (n + 1) (n + 1) a

n

+

_

n

2

n 2

_

a

n

_

_

_x

n

,

so it follows that

6 (n + 2) (n + 1) a

n+2

5 (n + 1)

2

a

n+1

+ (n + 1) (n 2) a

n

= 0,

a

n+2

=

5

6

n + 1

n + 2

a

n+1

1

6

n 2

n + 2

a

n

.

With a

0

= y(0) = 1 and a

1

= y

(0) = 1, we have

a

2

=

5

6

1

2

a

1

1

6

2

2

a

0

=

5

12

+

2

12

=

7

12

,

a

3

=

5

6

2

3

1

6

1

3

a

1

=

5

9

7

12

+

1

6

1

3

=

35

108

+

6

108

=

41

108

.

Therefore,

y = 1 +x +

7

12

x

2

+

41

108

x

3

+ .

8.6. LEVEL 2: FINDING THE RECURSION RELATION 95

Notice that although we need to write our equation in standard form (in

this case dividing by x

2

5x + 6 to determine that 0 is an ordinary point of

the equation), it is not necessary to use standard form when computing the

recursion relation. Instead we want to use a form in which our coecients are

polynomials. Although our recursion relation is too messy for us to achieve

nd a general formula for a

n

(level 3) there are nevertheless some advantages

in computing it. In particular, we can now proceed to use it to nd the radius

of convergence.

To determine the radius of convergence, we use the ratio test R = 1/ [L[,

where L = lim

n

(a

n+1

/a

n

), provided that the limit exists. It is hard to prove

that the limit exists, but assuming that

a

n+2

a

n+1

=

5

6

n + 1

n + 2

1

6

n 2

n + 2

a

n

a

n+1

,

taking the limit gives

L =

5

6

1

1

6

1

L

=

5

6

1

6

1

L

.

So 6L

2

= 5L 1, and solving for L gives L = 1/2 or L = 1/3. So R = 2 or

R = 3. The worst case is when R = 2. Therefore, the radius of convergence is

at least 2. Theorem 8.3 also gives this immediately.

Example 8.11. Consider

_

x

2

5x + 6

_

y

5y

independent series solutions y

1

and y

2

as far as x

3

.

Solution. Let y

1

(x) be the solution satisfying y

1

(0) = 1; y

1

(0) = 0 and let y

2

(x)

be the solution satisfying y

2

(0) = 0; y

2

(0) = 1. Then y

1

(x) and y

2

(x) will be

linearly independent solutions. We already computed the recursion relation in

Example 8.10. For y

1

(x), we have a

0

= y(0) = 1 and a

1

= y

(0) = 0. Thus,

a

2

=

5

6

1

2

a

1

1

6

_

2

2

_

a

0

=

1

6

,

a

3

=

5

6

2

3

a

2

1

6

1

3

a

1

=

5

54

.

Therefore,

y

1

= 1 +

x

2

6

+

5

54

x

3

+ .

96CHAPTER 8. POWER SERIES SOLUTIONS TOLINEAR DIFFERENTIAL EQUATIONS

For y

2

, we have a

0

= y(0) = 0 and a

1

= y

(0) = 1, so

a

2

=

5

6

1

2

a

1

2

2

a

0

=

5

12

,

a

3

=

5

6

2

3

a

2

_

1

3

_

a

1

=

5

6

2

3

5

12

+

1

6 3

=

25

108

+

6

108

=

31

108

.

Therefore,

y

2

= x +

5

12

x

2

+

61

108

x

3

+ .

With

y = 1 +x +

7

12

x

2

+

41

108

x

3

+ ,

Looking at the initial values, the solution y(x) of Example 8.10 ought to be

given by y = y

1

+y

2

, and comparing are answers we see that they are consistent

with this equation.

Summarizing: Let p be a regular point of the equation

y

+P(x)y

+Q(x)y = 0, (8.3)

Let D be the distance from p to the closest singularities of P or Q in the complex

plane. Then Equation (8.3) has two linearly independent solutions expressible

as a convergent power series

n=0

a

n

(x p)

n

with [x p[ < D. Theorem 8.5

guarantees that the radius of convergence is at least D, but it might be larger.

Given a

0

= y(p) and a

1

= y

(p), to nd a

n

, we have a

n

= y

(n)

(p)/n!. We

can nd as many a

n

s as desired, but in general we cannot nd the formula.

Since we have y

= P(x)y

(p)

and higher derivatives can be found by further dierentiation. Substituting the

values at p gives formulas for the a

n

s.

In the special case where P(x) and Q(x) are polynomials, we can instead

nd a recurrence relation. From the recurrence relation there is a chance to get

the formula for a

n

(if so, then there is a chance to actually recognize the series).

Even without the formula, we can determine the radius of convergence from the

recurrence relation (even if P(x) and Q(x) are not polynomials, we could at

least try to get a recurrence relation by expanding them into Taylor series,but

in practice this is not likely to be successful).

8.7. SOLUTIONS NEAR A SINGULAR POINT 97

8.7 Solutions Near a Singular Point

A singularity a of y

+ P(x)y

(x a) P(x) and (x a)

2

Q(x) are analytic at x = a. To simplify notation, we

assume that a = 0 (we can always change the variable by v = x a to move

the singularity to the origin). We assume that x > 0. For x < 0, we substitute

t = x and apply below to t.

Our primary goal in this section is to consider the Method of Frobenius. We

look for solutions of the form

y = x

(a

0

+a

1

x + +a

n

x

n

+ ) = a

0

x

+a

1

x

+1

+ +a

n

x

+n

+

for some , where a

0

,= 0. Then

y

= a

1

x

1

+a

1

( + 1) x

+ +a

n

( +n) x

+n1

+ ,

y

= a

1

( 1) x

2

+a

1

( + 1) x

1

+

+a

n

( +n) ( +n 1) x

+n2

+ ,

xP(x) = p

0

+p

1

x + +p

n

x

n

+ ,

x

2

Q(x) = q

0

+q

1

x + +q

n

x

n

+ .

So if x ,= 0, then

a

0

( 1) x

2

+a

1

( + 1) x

1

+ +a

n

( +n) ( +n 1) x

+n2

+

+x

2

(a

0

+a

1

( + 1) x + ) (p

0

+p

1

x + +p

n

x

n

+ )

+x

2

(a

0

+a

1

x + +a

n

x

n

) (q

0

+q

1

x + +q

n

x

n

+ ) = 0.

We have

a

0

( 1) +a

0

p

0

+a

0

q

0

= 0, (0)

a

1

( + 1) +a

1

( + 1) p

0

+a

0

p

1

+a

0

q

1

+a

1

q

0

= 0, (1)

.

.

.

.

.

.

(n)

Note that Equation (0) implies

( 1) +p

0

+q

0

= 0,

2

+ (p

0

1) +q

0

= 0. (8.4)

98CHAPTER 8. POWER SERIES SOLUTIONS TOLINEAR DIFFERENTIAL EQUATIONS

Equation (8.4) is known as an indicial equation. Solving for gives us two

solutions. Then Equation (1) gives a

1

in terms of a

0

, Equation (2) gives a

2

in

terms of a

1

, etc. We will get the solution y = a

0

x

solution y = b

0

x

n

from Equation (n) for n 1.

Equation (n) looks like

() a

n

+ () a

n1

+ + () a

0

= 0,

where the coecients are in terms of , p

0

, . . . , p

n

, q

0

, . . . , q

n

. There is one pos-

sible problem: what happens if the coecient of a

n

is zero? Note that

y

+P(x)y

+Q(x)y = y

+xP(x)

y

x

+x

2

Q(x)

y

x

2

,

so we have

y

+xP(x)

y

x

+x

2

Q(x)

y

x

2

=

n=0

a

n

( +n) ( +n 1) x

+n2

+

_

p

0

+p

1

x +p

2

x

2

+

_

n=0

a

n

( +n) x

+n2

+

_

q

0

+q

1

x +q

2

x

2

+

_

n=0

a

n

x

+n2

.

In Equation (n), the coecient of a

n

is

( +n) ( +n 1) +p

0

( +n) +q

0

.

Suppose that

( +n) ( +n 1) +p

0

( +n) +q

0

= 0.

Let = +n. Then

( 1) +p

0

+q

0

= 0

is the indicial equation again, i.e., is the other solution to the indicial equation.

We conclude that if two solutions of the indicial equation do not dier by an

integer, the method of Frobenius provides two solutions.

do dier by an integer, then the larger one will give a solution by this method,

but the smaller one may fail.

Let F() = ( 1) + p

0

+ q

0

. Suppose the solutions to F() = 0 are r

This includes the case where the solutions are complex, i.e., in this case, we get a complex

solution z = x

a+ib

(c

0

+ c

1

x + ), so y

1

= Re(x) and y

2

= Im(z) give two real solutions.

8.7. SOLUTIONS NEAR A SINGULAR POINT 99

and r +k, where k Z

+

. Given f(x), set

L(f)(x) = f

(x) +P(x)f

(x) +Q(x)f(x).

Let w =

n=0

a

n

x

n

, where a

n

is to be chosen. For any given , we have

L(x

w) =

n=0

c

n

x

2+n

,

where c

n

= F( + n)a

n

+ () a

n1

+ + () a

0

. Chose a

0

= 1. Then for

n 1, we can always solve c

n

= 0 for a

n

in terms of its predecessors unless

F(+n) = 0. Provided that ,= s 1, s 2, . . ., we have F(+n) ,= 0 for any

n, so we can solve it. Suppose that we call the solution a

n

(), i.e., a

0

() = 1 for

any and a

n

is chosen to make c

n

= 0 for this . Set W(, x)

n=0

a

n

()x

n

.

Then

L(x

w(, x)) =

n=0

c

n

x

2+n

= c

0

x

2

.

But since c

n

= 0 for n 0 by the choice of a

n

(), we have

L(x

2

.

Since F(s) = 0, y

1

(x) = x

s

w(s, x) is the solution to the de. We now need

a second solution. Unfortunately, we cannot set = r since our denition of

a

n

() depends on ,= s (positive integer). Instead, set

L(x

2

and dierentiated with respect to . Note that for g(, x), we have

x

(Lg) = L

_

g

x

_

since dierentiation with respect to x commutes with dierentiation with respect

to (cross derivatives equal). Therefore,

L

_

x

ln(x)w(, x) +x

_

= F

()x

2

+F()x

2

ln(x).

100CHAPTER 8. POWER SERIES SOLUTIONS TOLINEAR DIFFERENTIAL EQUATIONS

But note that

= s = L

_

ln(x)x

s

w(s, x) +x

s

_

w

=s

__

=

k

..

F

(s) x

s2

+

0

..

F(s) x

s2

ln(x)

= L

_

ln(x)y

1

+

_

w

=s

_

x

s

_

= kx

s2

.

If k = 0, then

y

2

= ln(x)y

1

+

_

w

=s

_

x

s

is the second solution to our de.

If k > 0, consider

L

_

x

r

n=0

a

n

x

n

_

= L

_

x

r

n=0

c

n

x

r+n2

_

,

where a

n

s are to be chosen and c

n

= F(r +n)a

n

+ . We have c

0

= F(r) = 0,

where r is the root of the indicial equation. Select a

0

= 1 and inductively choose

a

1

, . . . , a

n

such that c

1

, . . . , c

n

= 0. With this choice, we have

L

_

x

r

n=0

a

n

x

n

_

=

n=k

c

n

x

r+n2

= c

k

x

s2

+

n=k+1

c

n

(r)x

r+n2

,

where c

k

is independent of the choice of a

k

. Set A = c

k

/k. Then

L

_

x

r

n=0

a

n

x

n

A

_

ln(x)y

1

(x) +x

s

_

w

=s

__

_

=

n=k+1

c

n

x

r+n2

,

L

_

_

_

_

_

_

_

_

x

r

_

_

a

0

+a

1

x + +a

k1

x

k1

+

_

a

k

A

w

=s

_

x

k

_

_

+

n=k+1

a

n

x

n

Aln(x)y

1

(x)

_

_

_

_

_

_

_

_

=

n=k+1

c

n

x

r+n2

.

The choice of a

k

does not matterwe can choose any. Then inductively solve

for a

k+1

, a

k+2

, . . . in terms of predecessor a

s to make c

n

= 0 for n > k. This

gives a solution to the de of the form

y

2

= x

r

n=0

a

n

x

n

Aln(x)y

1

(x).

8.7. SOLUTIONS NEAR A SINGULAR POINT 101

We conclude that if 0 is a regular singularity, then there are always two

linearly independent solutions (converging near 0)

y

1

= x

s

n=0

a

n

x

n

, y

2

= x

r

n=0

b

n

x

n

+B(ln(x))y

1

,

where r and s are the two solutions to the indicial equation. We have B = 0,

unless s = r +k for some k 0, 1, 2, . . ..

Example 8.12. Solve 4xy

+ 2y

y = 0.

Solution. Let y =

n=0

a

n

x

+n

. Then

y

n=0

( +n) a

n

x

+n1

, y

n=0

( +n) ( +n 1) a

n

x

+n2

.

Our dierential equation now becomes

n=0

4 ( +n) ( +n 1) a

n

x

+n1

+

n=0

2 ( +n) a

n

x

+n1

n=0

a

n

x

+n

= 0,

n=0

[4 ( +n) ( +n 1) + 2 ( +n)] a

n

x

+n1

n=1

a

n1

x

+n1

= 0.

Note that

4 ( +n) ( +n 1) + 2 ( +n) = 2 ( +n) (2 + 2n 2 + 1)

= 2 ( +n) (2 + 2n 1) .

Therefore,

2(2 1) a

0

x

1

+

n=1

[2 ( +n) (2 + 2n 1) a

n1

] x

+n1

.

Therefore, the indicial equation is 2( 1) 0, 1/2 and the recursion

relation is

2 ( +n) (2 + 2n 1) a

n

a

n1

= 0

for n 1.

For = 0, we have

2n(2n 1) a

n

a

n1

= 0 = a

n

=

1

2n(2n 1)

a

n1

.

102CHAPTER 8. POWER SERIES SOLUTIONS TOLINEAR DIFFERENTIAL EQUATIONS

Thus,

a

1

=

1

2 1

a

0

,

a

2

=

1

4 3

a

1

=

1

4!

a

0

,

.

.

.

a

n

=

1

(2n) (2n 1)

a

n1

=

1

2n(2n 1) (2n 2)

=

1

(2n)!

a

0

.

Therefore,

y = x

0

a

0

_

1 +

x

2!

+

x

2

4!

+ +

x

n

(2n)

! +

_

= a

0

_

1 +

x

2!

+

x

2

4!

+ +

x

n

(2n)

! +

_

= a

0

cosh

_

x

_

.

For = 1/2, we have

2

_

1

2

+n

_

(1 + 2n 1) b

n

= b

n1

= b

n

=

1

(2n + 1) (2n)

b

n1

.

Thus,

b

1

=

1

3 2

b

0

,

b

2

=

1

5 4

b

1

=

1

5!

b

0

,

.

.

.

b

n

=

1

(2n + 1)!

b

0

.

8.7. SOLUTIONS NEAR A SINGULAR POINT 103

Therefore,

y = x

1/2

b

0

_

1 +

x

3!

+

x

2

5!

+ +

x

n

(2n + 1)!

+

_

= b

0

_

x

1/2

+

x

3/2

3!

+

x

5/2

5!

+ +

x

(2n+1)/2

(2n + 1)!

+

_

= b

0

sinh

_

x

_

.

In general, however, it is possible to only calculate a few low terms. We cannot

even get a general formula for a

n

, let alone recognize the series.

As mentioned, the above methods work near a by using the above technique

after the substitution v = x a. We can also get solutions near innity, i.e.,

for large x, we substitute t = 1/x and look at t = 0. Then

dy

dx

=

dy

dt

dt

dx

=

dy

dx

_

1

x

2

_

=

1

x

2

dy

dt

= t

2

dy

dt

,

d

2

y

dx

2

=

2

x

3

dy

dt

=

1

x

2

d

2

y

dt

2

dt

dx

=

2

x

3

dy

dt

1

x

2

d

2

y

dt

2

_

1

x

2

_

= 2t

3

dy

dt

+t

4

d

2

y

dt

2

.

Then

d

2

y

dx

2

+P(x)

dy

dx

+Q(x)y = 0 = t

4

d

2

y

dt

2

+ 2t

3

dy

dt

P

_

1

t

_

t

2

dy

dt

+Q

_

1

t

_

y = 0.

Example 8.13 (Legendres equation). Find the behaviour of the solutions

for large x of

_

1 x

2

_

y

2xy

Solution. Let t = 1/x so that x = 1/t. Then our dierential equation becomes

_

1

1

t

2

__

2t

3

dy

dt

+t

4

d

2

y

dt

2

_

2

t

_

t

2

dy

dt

_

+n(n + 1) y = 0,

_

t

4

t

2

_

d

2

y

dt

2

+

_

2t

3

2t + 2y

_

dy

dt

+n(n + 1) y = 0,

d

2

y

dt

2

+

2t

t

2

1

dy

dt

+

n(n + 1)

t

4

t

2

y = 0.

We see that we have a regular singularity at t = 0. Therefore, there exists a

solution y = t

n=0

a

n

t

n

near t = 0, or equivalently, since t = 1/x,

y =

1

x

n=0

1

x

n

104CHAPTER 8. POWER SERIES SOLUTIONS TOLINEAR DIFFERENTIAL EQUATIONS

for large x.

Example 8.14. Find the solutions of y

+xy

Solution. Let

y = x

n=0

a

n

x

n

=

n=0

a

n

x

n+

.

Then

y

n=0

a

n

(n +) x

n+1

,

y

n=0

a

n

(n +) (n + 1) x

n+2

,

xy

n=0

a

n

x

n+

=

n=2

a

n2

x

n+2

,

y

x

=

n=0

a

n

x

n+1

=

n=1

a

n1

x

n+2

.

Note that

Indicial equation

..

a

0

( 1) = 0, (0)

a

1

( + 1) +a

0

= 0, (1)

.

.

.

.

.

.

a

n

(n +) (n + 1) +a

n1

+a

n2

= 0. (n)

Note that Equation (0) implies that = 0, 1. The roots dier by an integer.

The largest always gives the solution. For = 1, Equation (1) implies that

2a

1

+a

0

= 0, so a

1

= a

0

/2. Equation (n) then becomes

a

n

(n + 1) n +a

n1

+a

n2

= 0 = a

n

=

a

n1

a

n2

n(n + 1)

.

Equation (2) implies that

6a

2

= a

0

a

1

= a

0

_

1 +

1

2

_

=

a

0

2

= a

2

=

a

0

12

;

8.7. SOLUTIONS NEAR A SINGULAR POINT 105

Equation (3) implies that

12a

3

= a

0

_

1

2

+

1

12

_

= a

0

7

12

= a

3

= a

0

7

144

.

Therefore,

y

1

= a

0

x

_

1

1

2

x

1

12

x

2

+

7

144

x

3

+

_

.

As for the second solution, we have

y = x

0

_

b

0

+b

1

x +b

2

x

2

+

_

+C ln(x)y

1

,

y

= b

1

+ 2b

2

x + +

C

x

y

1

+C ln(x)y

1

,

y

= 2b

2

+ 6b

3

x +

C

x

2

y

1

+

C

x

y

1

+

C

x

y

1

+C ln(x)y

1

.

Therefore, our dierential equation becomes

y

+xy

+

y

x

= 0,

_

_

_

_

_

_

_

2b

2

+ 6b

3

x + 24b

4

x

2

+

_

C

x

2

y

1

+

2C

x

y

1

+C ln(x)y

1

+

_

b

1

x + 2b

2

x

2

+

_

+Cy

1

+C ln(x)xy

1

+

_

b

0

x

+b

1

+b

2

x +

_

+C ln(x)

y

1

x

_

_

_

_

_

_

= 0,

_

_

_

_

_

_

_

_

_

_

2b

2

+ 6b

3

x + 24b

4

x

2

+

_

C

x

2

y

1

+

2C

x

y

1

+

_

b

1

x + 2b

2

x

2

+

_

+Cy

1

+

_

b

0

x

+b

1

+b

2

x +

_

+C ln(x)y

1

+C ln(x)xy

1

+C ln(x)

y

1

x

. .

0

_

_

_

_

_

_

_

_

_

= 0,

_

_

_

_

_

_

_

_

_

_

_

_

_

2b

2

+ 6b

3

x + 24b

4

x

2

+

_

C

x

_

1

1

2

x

1

12

x

2

+

7

144

x

3

+

_

+

2C

x

_

1 x +

1

4

x

2

+

7

36

x

3

+

_

+

_

b

1

x + 2b

2

x

2

+

_

+C

_

x

1

2

x

2

1

12

x

3

+

7

144

x

4

+

_

+

_

b

0

x

+b

1

+b

2

x +

_

_

_

_

_

_

_

_

_

_

_

_

_

= 0.

Looking at the coecient of x

1

, we have b

0

C + 2C = 0 C = b

0

.

Note that b

1

is arbitrary, so choose b

0

= 0. Then the coecient of x

0

implies

that 2b

2

+ C/2 2C = 0 b

2

= (3/4) b

0

; the coecient of x

1

implies that

106CHAPTER 8. POWER SERIES SOLUTIONS TOLINEAR DIFFERENTIAL EQUATIONS

6b

3

+C/12 +C/2 +C +b

2

= 0 b

3

= (7/18) b

0

. Therefore,

y

2

= b

0

__

1

3

4

x

2

+

7

18

x

2

+

_

ln(x)

_

x

1

2

x

2

1

12

x

3

+

7

144

x

4

+

__

.

2

y

xy

+ 10e

x

y = 0.

Solution. Dividing out by x

2

, we have

y

x

+

10e

x

x

2

y = 0.

We have a singularity at x = 0. Note that xP(x) = 1 and x

2

Q(x) = 10e

x

=

10

n=0

x

n

/n!. Therefore,

y =

n=0

a

n

x

+n

,

y

x

2

=

n=0

a

n

x

+n2

,

y

n=0

( +n) a

n

x

+n1

,

y

x

=

n=0

( +n) a

n

x

+n2

,

y

n=0

( +n) ( +n 1) a

n

x

+n2

.

With these components in place, our dierential equation becomes

n=0

( +n) ( +n 1) a

n

x

+n2

n=0

( +n) a

n

x

+n2

+10

_

n=0

x

n

n!

__

n=0

a

n

x

+n2

_

= 0.

We now have

( 1) a

0

a

0

+ 10a

0

= 0.

Therefore, F() = ( 1) + 10 = 0 = 1 3i. For 1 + 3i, we have

( + 1) a

1

( + 1) a

1

+ 10 (a

0

+a

1

) = 0,

F( + 1)a

1

+ 10a

1

= 0.

Therefore,

a

1

=

10

F( + 1)

a

0

=

10

F(2 + 3i)

a

0

8.7. SOLUTIONS NEAR A SINGULAR POINT 107

and

F(2 + 3i) = (2 + 3i)

2

2 (2 + 3i) + 10

= 4 + 12i 9 4 6i + 10

= 1 + 6i.

Therefore,

a

1

=

10

1 + 6i

a

0

=

10 (1 6i)

(1 + 6i) (1 6i)

a

0

=

10 60i

1 + 36

a

0

=

10 60i

37

a

0

.

Also note that

( + 2) ( + 1) a

2

( + 2) a

2

+ 10

_

a

0

2!

+a

1

+a

2

_

= 0,

and we have

F( + 2)a

2

+ 10a

1

+ 5a

0

= F( + 2)a

2

100 600i

37

a

0

+

175

37

a

0

= F( + 2)a

2

+

75 + 600i

37

a

0

.

Therefore,

a

2

=

75 + 600i

37F( + 2)

a

0

and we have

z = x

1+3i

a

0

_

1

_

10 60i

37

_

x +

_

.

But

x

1+3i

= xx

3i

= xe

3i ln(x)

= x[cos (3 ln(x)) +i sin(3 ln(x))] ,

so

z = a

0

xcos(3 ln(x)) +a

0

i sin(3 ln(x))

+a

0

_

10

37

x

2

cos(3 ln(x))

60

37

x

2

sin(3 ln(x)) +

60

37

sin(3 ln(x))

_

+

60

37

ix

2

cos(3 ln(x))

10

37

ix

2

sin(3 ln(x)) + .

108CHAPTER 8. POWER SERIES SOLUTIONS TOLINEAR DIFFERENTIAL EQUATIONS

Therefore,

Re(z) = a

0

_

_

_

_

xcos(3 ln(x))

+x

2

_

10

37

cos(3 ln(x))

60

37

sin(3 ln(x))

_

+x

3

[ cos(3 ln(x)) + sin(3 ln(x))]

_

_

_

_

+

+ (1 x) y

y = 0.

Solution. Note that

p

0

= lim

x0

xP(x) = lim

x0

x

1 x

x

= 1

and

q

0

= lim

x0

x

2

Q(x) = lim

x0

x

2

_

1

x

_

= 0.

Therefore, 0 is a regular singular point. The indicial equation is

( 1) +p

0

+q

0

= 0,

( 1) + = 0,

2

+ = 0,

2

= 0.

Therefore, = 0 is a double root and

y

1

n=0

x

n

, y

2

=

n=0

b

n

x

n

+C ln(x)y

1

, x > 0.

To nd y

1

, we have

y =

n=0

a

n

x

n

,

y

n=0

na

n

x

n1

=

n=0

(n + 1) a

n+1

x

n

,

y

n=0

(n + 1) nx

n1

.

8.7. SOLUTIONS NEAR A SINGULAR POINT 109

Therefore, our dierential equation becomes

xy

+ (1 x) y

y = 0,

n=0

[(n + 1) na

n+1

+ (n + 1) a

n+1

na

n

a

n

] x

n

= 0.

From this, we have

(n + 1) na

n+1

+ (n + 1) a

n+1

na

n

a

n

= 0,

(n + 1)

2

a

n+1

= (n + 1) a

n

,

a

n+1

=

a

n

n + 1

.

Therefore,

a

1

=

a

0

1

, a

2

=

a

1

2

=

a

0

2

, a

3

=

a

2

2

=

a

0

3!

, . . . , a

n

=

a

0

n!

.

So we have

y = a

0

n=0

x

n

n!

= a

0

e

x

.

Immediately, y

1

= e

x

. To nd y

1

(see 5.1, p. 57), we have

y

2

= y

1

_

e

R

P(x) dx

y

2

1

, dx = e

x

_

e

R

1x

x

dx

e

2x

dx

= e

x

_

e

R

(

1

x

1)dx

e

2x

dx = e

x

_

e

(ln(x)x)

e

2x

dx

= e

x

_

e

x

/x

e

2x

dx = e

x

_

e

x

x

dx,

which is not an elementary function.

Example 8.17. Solve

_

x

2

+ 2x

_

y

2

_

x

2

+ 2x 1

_

y

+

_

x

2

+ 2x 2

_

y = 0.

Solution. Let y =

n=0

a

n

x

+n

. Then

y

n=0

( +n) a

n

x

+n1

, y

n=0

( +n) ( +n 1) a

n

x

+n2

.

110CHAPTER 8. POWER SERIES SOLUTIONS TOLINEAR DIFFERENTIAL EQUATIONS

The de then becomes

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

n=0

( +n) ( +n 1) a

n

x

+n

+ 2

n=0

( +n) ( +n 1) a

n

x

+n1

2

n=0

( +n) a

n

x

+n+1

4

n=0

( +n) a

n

x

+n

+2

n=0

( +n) a

n

x

+n1

+

n=0

a

n

x

+n+2

+ 2

n=0

a

n

x

+n+1

2

n=0

a

n

x

+n

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

= 0,

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

n=1

( +n 1) ( +n 2) a

n1

x

+n1

+2

n=0

( +n) ( +n 1) a

n

x

+n1

2

n=2

( +n 2) a

n2

x

+n1

4

n=1

( +n 1) a

n1

x

+n1

+ 2

n=0

( +n) a

n

x

+n1

+

n=3

a

n3

x

+n1

+ 2

n=2

a

n2

x

+n1

2

n=1

a

n1

x

+n1

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

_

= 0,

_

n=0

_

_

_

[2 ( +n) ( +n 1) + 2 ( +n)] a

n

+[( +n 1) ( +n 2) 4 ( +n 1) 2] a

n1

+[2 ( +n 2) + 2] a

n2

+a

n3

_

_

_

_

_x

+n1

= 0,

n=0

_

_

2 ( +n)

2

a

n

+

_

( +n)

2

7 ( +n) + 4

_

a

n1

+[2 ( +n) + 6] a

n2

+a

n3

_

_

= 0.

Note that 2

2

= 0 = 0, giving us a double root. Therefore we get the

recursion relation

2n

2

a

n

+

_

n

2

7n + 4

_

a

n1

+ (2n + 6) a

n2

+a

n3

= 0.

Substituting gives

2a

1

2a

0

= 0 = a

1

= a

0

,

8a

2

6a

1

+ 2a

0

= 0 = 8a

2

= 6a

1

2a

0

= (6 2) a

0

= 4a

0

= a

0

=

1

2

,

18a

3

+ (8) a

2

+ (0) a

1

+a

2

= 0 = 18a

3

= 8

1

2

1 = 3 = a

3

=

1

6

,

32a

4

8a

3

(2) a

2

+ (0) a

1

+a

2

= 0 = 32a

3

= 8

1

6

2

1

2

1 =

4

3

= a

3

=

1

24

.

8.8. FUNCTIONS DEFINED VIA DIFFERENTIAL EQUATIONS 111

If we our intuition leads us to guess that perhaps a

n

= 1/n! be can use induction

to check that this guess is in fact correct. Explicitly, suppose by induction that

a

k

= 1/k! for all k < n. Then the recursion relation yields

2n

2

a

n

=

n

2

7n + 4

(n 1)!

+

2n 6

(n 2)!

1

(n 3)!

=

n

2

+ 7n 4 + (2n 6)(n 1) (n 1)(n 2)

(n 1)!

=

n

2

+ 7n 4 + 2n

2

8n + 6 n

2

+ 3n 2

(n 1)!

=

2n

(n 1)!

and so a

n

= 1/n!, completing the induction.

This gives the solution y

1

=

n=0

x

n

/n! = e

x

. (If we had guessed initially

that this might be the solution, we could easily have checked it by direct subsi-

tution into the dierential equation.) Using reduction of order, one then nds

that the solutions are

y

1

= e

x

, y

2

= xe

x

+ 2 ln(x) y

1

.

8.8 Functions Dened via Dierential Equations

8.8.1 Chebyshev Equation

The Chebyshev equation is given by

_

1 x

2

_

y

xy

+

2

y = 0 (8.5)

112CHAPTER 8. POWER SERIES SOLUTIONS TOLINEAR DIFFERENTIAL EQUATIONS

of which 0 is an ordinary point. The series solution is guaranteed to converge

for [x[ < 1. We have

y =

n=0

a

n

x

n

,

y

n=0

na

n

x

n1

,

y

n=0

n(n 1) a

n

x

n2

=

n=0

(n + 2) (n + 1) a

n+2

x

n

,

and our dierential equation becomes

n=0

_

(n + 2) (n + 1) a

n+2

n(n 1) a

n

na

n

+

2

a

n

x

n

= 0.

From this, we see that

a

n+2

=

n(n 1) +n

2

(n + 2) (n + 1)

a

n

=

n

2

2

(n + 2) (n + 1)

a

n

.

Therefore,

a

0

= 1, a

1

= 0, a

2

=

2

2

, a

3

= 0,

a

4

=

2

4 3

=

_

4

2

_ _

2

_

4!

, a

5

= 0,

a

6

=

_

16

2

_ _

4

2

_ _

2

_

6!

,

.

.

.

a

2n

=

_

4 (n 1)

2

2

__

4 (n 2)

2

2

_

_

2

_

(2n)!

a

1

.

8.8. FUNCTIONS DEFINED VIA DIFFERENTIAL EQUATIONS 113

Also,

a

0

= 0, a

1

= 1, a

2

= 0, a

3

=

_

1

2

_

3 2

a

1

,

a

4

= 0, a

5

=

_

9

2

_ _

1

2

_

5!

a

1

,

.

.

.

a

2n+1

=

_

(2n 1)

2

2

_

_

1

2

_

(2n + 1)!

a

1

.

Note that if is an integer, then one of these solutions terminates, i.e.,

a

k

= 0 beyond some point. In this case, one solution is a polynomial known as

a Chebyshev polynomial.

8.8.2 Legendre Equation

The Legendre equation is given by

_

1 x

2

_

y

2xy

+( + 1) y = 0 (8.6)

of which 0 is an ordinary point. The series solution is guaranteed to converge

for [x[ < 1. Expressed as a power series, the equation becomes

n=0

= [(n + 2) (n + 1) a

n+2

n(n 1) a

n

2na

n

+( + 1) a

n

] x

n

= 0.

From this, we have

a

n+2

=

n(n 1) + 2n ( + 1)

(n + 2) (n + 1)

a

n

=

n(n + 1) ( + 1)

(n + 2) (n + 1)

a

n

.

If is an integer, then a

k

= 0 for k gives the th Legendre polynomial.

Instead, expanding around x = 1 gives

p(x)

x

1 x

2

.

114CHAPTER 8. POWER SERIES SOLUTIONS TOLINEAR DIFFERENTIAL EQUATIONS

Then

p

0

= lim

x1

(x 1) P(x) = lim

x1

x

x + 1

=

1

2

,

q

0

= lim

x1

(x 1)

2

Q(x) = lim

x1

(x 1)

2

2

x

2

1

= 0.

Therefore, x = 1 is a regular singular point. The indicial equation is

( 1) +p

0

+q

0

= 0,

( 1) +

1

2

+q

0

= 0,

2

+

1

2

= 0,

1

2

= 0.

Therefore, = 0 or = 1/2, so the solutions look like

y

1

=

n=0

a

n

(x 1)

n

, y

2

= (x 1)

1/2

n=0

b

n

(x 1)

n

for x > 1.

For a Legendre equation, around x = 1, we have

p

0

= lim

x1

(x 1)

_

2x

1 x

2

_

= 1,

q

0

= lim

x1

(x 1)

2

( + 1)

x

2

1

= 0.

Therefore, x = 1 is a regular singular point. The indicial equation is

( 1) +p

0

+q

0

= 0,

( 1) + = 0,

2

= 0.

Therefore, = 0 is a double root and the solutions look like

y

1

=

n=0

a

n

(x 1)

n

, y

2

=

n=0

b

n

(x 1)

n

+C ln(x 1) y

1

for x > 1.

8.8. FUNCTIONS DEFINED VIA DIFFERENTIAL EQUATIONS 115

8.8.3 Airy Equation

The Airy equation is given by

y

xy = 0 (8.7)

of which 0 is an ordinary point. The solution converges for all x. Expressed as

a power series, we have

n=0

[(n + 2) (n + 1) a

n+2

a

n1

] x

n

= 0.

From this, we see that

a

n+2

=

a

n1

(n + 2) (n + 1)

,

therefore,

a

3

=

a

0

3 2

, a

6

=

a

0

6 5

, . . . , a

3n

=

a

0

2 3 5 6 8 9 (3n 1) (3n)

,

a

4

=

a

1

4 3

, . . . , a

3n+1

=

a

1

3 4 6 7 (3n) (3n + 1)

,

a

2

=

a

1

2

= 0, 0 = a

2

= a

5

= a

8

= .

8.8.4 Laguerres Equation

Laguerres equation

is given by

xy

(1 x) y

+y = 0 (8.8)

of which 0 is a regular singular point. Note that

p

0

= lim

x0

x

_

(1 x)

x

_

= 1,

q

0

= lim

x0

x

2

x

= 0.

116CHAPTER 8. POWER SERIES SOLUTIONS TOLINEAR DIFFERENTIAL EQUATIONS

Therefore, x = 0 is a regular singular point. The indicial equation is

( 1) = 0,

2

= 0,

2

2 = 0,

( 2) = 0.

Therefore, = 0 or = 2, so the solutions look like

y

1

= x

2

n=0

a

n

x

n

, y

2

=

n=0

b

n

x

n

+C ln(x)y

1

for x > 0.

8.8.5 Bessel Equation

The Bessel equation is given by

x

2

y

+xy

+

_

x

2

2

_

y = 0. (8.9)

We have

p

0

= lim

x0

x

x

x

2

= 1,

q

0

= lim

x0

x

2

x

2

2

x

2

= .

The indicial equation is then

( 1) +

2

= 0,

2

= 0,

( +) ( ) = 0.

Therefore, = . If ,= k/2 for some integer k, then we get

y

1

= x

n=0

a

n

x

n

, y

2

= x

n=0

a

n

x

n

8.8. FUNCTIONS DEFINED VIA DIFFERENTIAL EQUATIONS 117

for x > 0. We encounter trouble if = k/2. Consider = 1/2. Then

y

1

= x

1/2

n=0

a

n

x

n

, y

2

= x

1/2

n=0

b

n

x

n

+C ln(x)y

1

.

For y

1

, we have

y =

n=0

a

n

x

n+1/2

,

y

n=0

_

n +

1

2

_

a

n

x

n1/2

,

y

n=0

_

n +

1

2

__

n

1

2

_

. .

n

2

1/4

a

n

x

n3/2

.

Therefore, the equation x

2

y

+xy

+

_

x

2

1/4

_

y = 0 becomes

n=0

__

n

2

1

4

_

a

n

+

_

n +

1

2

_

a

n

+a

n2

1

4

a

n

_

x

n+1/2

= 0.

From this, we see that

_

n

2

1

4

+n +

1

2

1

4

_

a

n

+a

n2

= 0,

a

n

=

a

n2

n

2

+n

=

a

n2

n(n + 1)

.

Therefore,

a

2

=

a

0

3 2

, a

4

=

a

2

5 4

=

a

0

5!

, a

6

=

a

0

7!

, . . . , a

2n

= (1)

n

a

0

(2n + 1)!

.

Therefore

y

1

=

n=0

(1)

n

x

2n+1/2

(2n + 1)!

=

1

n=0

(1)

n

x

2n+1

(2n + 1)!

=

sin(x)

x

.

118CHAPTER 8. POWER SERIES SOLUTIONS TOLINEAR DIFFERENTIAL EQUATIONS

Example 8.18. Identify

y =

n=0

x

3n

(3n)!

= 1 +

x

3

3!

+

x

6

6!

+

x

9

9!

+ .

3

1 = 0,

( 1)

_

2

+ + 1

_

= 0.

Therefore, = 1 or

=

1

1 4

2

=

1

2

3

2

i.

Therefore, the general solution is

y = C

1

e

x

+C

2

e

x/2

sin

_

3

2

x

_

+C

3

e

x/2

cos

_

3

2

x

_

,

where C

1

, C

2

, C

3

are arbitrary constants.

To nd C

1

, C

2

, C

3

, we note that

y(0) = 1, y

(0) = 0, y

(0) = 0.

Therefore,

C

1

+ 0C

2

+C

3

= 1,

C

1

+

3

2

C

2

1

2

C

3

= 0,

C

1

3

2

C

2

1

2

C

3

= 0.

8.8. FUNCTIONS DEFINED VIA DIFFERENTIAL EQUATIONS 119

Solving this system, we have

_

_

1 0 1 1

1

3

2

1

2

0

1

3

2

1

2

0

_

_

_

_

1 0 1 1

0

3

2

3

2

1

0

3

2

3

2

1

_

_

_

_

1 0 1 1

0

3

2

3

2

1

0 0 3 2

_

_

_

_

1 0 1 1

0

1

3

1

2

3

0 0 1

2

3

_

_

_

_

1 0 0

1

3

0

1

3

0 0

0 0 1

2

3

_

_

_

_

1 0 0

1

3

0 1 0 0

0 0 1

2

3

_

_.

Therefore, C

1

= 1/3, C

2

= 0, and C

3

= 2/3, so

y =

1

3

e

x

+

2

3

e

x/2

cos

_

3

2

x

_

.

Chapter 9

Linear Systems

9.1 Preliminaries

Consider

dY

dx

= AY +B,

where

A =

_

a b

c d

_

, Y =

_

y

1

y

2

_

such that

dy

1

dx

= ay

1

+by

2

,

dy

2

dx

= cy

1

+dy

2

.

We will rst consider the homogeneous case where B = 0.

Only in the case when the entries of A are constants is there an algorithmic

method of nding the solution. If A a is a constant, then we know that

y = Ce

ax

is a solution to dy/dx = ay. We will dene a matrix e

A

such that the

general solution of dY/dx = AY is Y = Ce

xA

, where C = [C

1

, C

2

, . . . , C

n

] is

a vector of arbitrary constants. We dene e

T

for an arbitrary n n matrix T

and apply T = xA. Indeed, we know that

e

t

= 1 +t +

t

2

2!

+

t

3

3!

+

t

4

4!

+ +

t

n

n!

+ .

Therefore, let us analogously dene

e

T

I +T+

T

2

2!

+

T

3

3!

+ +

T

n

n!

+ . (9.1)

121

122 CHAPTER 9. LINEAR SYSTEMS

We will consider systematically how to compute e

T

later, but rst we consider

an ad hoc example.

Example 9.1. Solve

dY

dx

=

_

1 1

0 1

_

Y.

A =

_

1 1

0 1

_

= I +N,

where

N =

_

0 1

0 0

_

so that NI = IN = N and N

2

= 0. Note that

A

n

=

Binomial exp.

..

(I +N)

n

= I

n

+

_

n

1

_

I

n1

N+

_

n

2

_

I

n1

N

2

+ +

_

n

n 1

_

IN

n1

+N

n

= I +nN =

_

1 n

0 1

_

.

Therefore, with A

n

properly dened, Equation (9.1) gives

e

Ax

= I +

_

1 1

0 1

_

x +

_

1 2

0 1

_

x

2

2

+ +

_

1 n

0 1

_

x

n

n!

+

=

_

1 +x +

x

2

2!

+ +

x

n

n!

+ 0 +x + 2

x

2

2!

+ 3

x

3

3!

+ +n

x

n

n!

0 1 +x +

x

2

2!

+ +

x

n

n!

+

_

=

_

_

e

x

x

_

1 +x +

x

2

2!

+ +

x

n

n!

+

_

0 e

x

_

_

=

_

e

x

xe

x

0 e

x

_

.

Therefore, the solution is

_

y

1

y

2

_

=

_

e

x

xe

x

0 e

x

__

C

1

C

2

_

=

_

C

1

e

x

+C

2

xe

x

C

2

e

x

_

,

9.2. COMPUTING E

T

123

that is,

y

1

= C

1

e

x

+C

2

xe

x

, y

2

= C

2

e

x

.

Is this in fact the solution of the equation? We have

d

dx

e

Ax

=

d

dx

_

I +Ax +A

2

x

2

2!

+ +A

n

x

n

n!

+

_

= A+A

2

x + +A

n

x

n1

(n 1)! +

= Ae

Ax

.

So if Y = Ce

Ax

, then

dY

dx

= ACe

Ax

= AY,

as required. Therefore, Y = Ce

Ax

is indeed the solution.

9.2 Computing e

T

The easiest case to do is a diagonal matrix. If

D =

_

2

.

.

.

n

_

_

,

then

e

D

=

_

_

e

1

e

2

.

.

.

e

n

_

_

.

First, we consider the eect of changing bases. Suppose T is diagonalizable, that

is, there exists a P such that P

1

TP is diagonalizable. Therefore, T = PDP

1

and

T

2

=

_

PDP

1

_ _

PDP

1

_

= PD

2

P

1

.

124 CHAPTER 9. LINEAR SYSTEMS

So in general, we have T

n

= PD

n

P

1

. It then follows that

e

T

= I +T+

T

2

2!

+ +

T

n

n!

= P

_

I +D+

D

2

2!

+ +

D

n

n!

+

_

P

1

= Pe

D

P

1

.

So if T is diagonalizable, we can compute e

T

.

Example 9.2. Compute e

T

, where

T =

_

4 1

3 2

_

.

4 1

3 2

= 0,

(4 ) (2 ) 3 = 0,

8 6 +

2

3 = 0,

2

6 + 5 = 0,

( 5) ( 1) = 0.

Therefore, = 1 or = 5. For = 1, we have

_

4 1 1

3 2 1

__

a

b

_

=

_

0

0

_

,

_

3 1

3 1

__

a

b

_

=

_

0

0

_

,

_

3a +b

3a +b

_

=

_

0

0

_

.

We see that b = 3a, so

_

a

b

_

=

_

a

3a

_

= a

_

1

3

_

.

9.2. COMPUTING E

T

125

For = 5, we have

_

4 5 1

3 2 5

__

a

b

_

=

_

0

0

_

,

_

1 1

3 3

__

a

b

_

=

_

0

0

_

,

_

a +b

3 3b

_

=

_

0

0

_

.

We see that a = b, so

_

a

b

_

=

_

a

a

_

= a

_

1

1

_

.

So let

P =

_

1 1

3 1

_

.

Therefore,

P

1

=

1

[P[

_

1 1

3 1

_

=

1

4

_

1 1

3 1

_

.

So

PTP =

1

4

_

1 1

3 1

__

4 1

3 2

__

1 1

3 1

_

=

1

4

_

1 1

3 1

__

1 5

3 5

_

=

1

4

_

1 0

0 20

_

.

What is P? For each eigenvalue

j

, we nd its corresponding eigenvector

v

j

. Then for P = [v

1

, v

2

, . . . , v

n

], which is an n n matrix, we have

A = P

_

2

.

.

.

n

_

_

P

1

.

126 CHAPTER 9. LINEAR SYSTEMS

For Y

= AY, we have

Y = e

Ax

C = e

PDP

1

x

C = Pe

xD

constant

..

P

1

C

= P

_

_

e

1x

e

2x

.

.

.

e

nx

_

_

C

= [v

1

, v

2

, . . . , v

n

]

_

_

e

1x

e

2x

.

.

.

e

nx

_

_

C

= C

1

e

1x

v

1

+C

2

e

2x

v

2

+ +C

n

e

nx

v

n

,

where C = P

1

C.

Example 9.3. Solve

Y

=

_

4 1

3 2

_

Y.

4 1

3 2

= 0,

(4 ) (2 ) 3 = 0,

8 6 +

2

3 = 0,

2

6 + 5 = 0,

( 5) ( 1) = 0.

Therefore, = 1 or = 5. For = 1, we have

_

4 1 1

3 2 1

__

a

b

_

=

_

0

0

_

,

_

3 1

3 1

__

a

b

_

=

_

0

0

_

,

_

3a +b

3a +b

_

=

_

0

0

_

.

9.2. COMPUTING E

T

127

We see that b = 3a, so

_

a

b

_

=

_

a

3a

_

= a

_

1

3

_

.

For = 5, we have

_

4 5 1

3 2 5

__

a

b

_

=

_

0

0

_

,

_

1 1

3 3

__

a

b

_

=

_

0

0

_

,

_

a +b

3 3b

_

=

_

0

0

_

.

We see that a = b, so

_

a

b

_

=

_

a

a

_

= a

_

1

1

_

.

Therefore,

P =

_

1 1

3 1

_

, D =

_

1 0

0 5

_

and we have

Y =

_

1 1

3 1

__

e

x

0

0 e

5x

__

C

1

C

2

_

=

_

e

x

e

5x

3e

x

e

5x

__

C

1

C

2

_

=

_

C

1

e

x

+C

2

e

5x

3C

1

e

x

+C

2

e

5x

_

= C

1

e

x

_

1

3

_

+C

2

e

5x

_

1

1

_

,

that is,

y

1

= C

1

e

x

+C

2

e

5x

, y

2

= 3C

1

e

x

+C

2

e

5x

.

If T does not have n distinct eigenvalues, computing e

T

requires greater

knowledge of linear algebra.

Property 9.4

1. e

PAP

1

= Pe

A

P

1

.

2. If AB = BA, then e

A+B

= e

A

e

B

= e

B

e

A

.

MATB24.

128 CHAPTER 9. LINEAR SYSTEMS

3. If

D =

_

2

.

.

.

n

_

_

,

then

e

D

=

_

_

e

1

e

2

.

.

.

e

n

_

_

.

4. If A

n+1

= 0, then

e

A

= I +A+

A

2

2!

+ +

A

n

n!

.

Proof. We have already proved (1), and (3) and (4) are trivial. What remains

is (2). So note that

e

A

e

B

=

_

I +A+

A

2

2!

+ +

A

n

n!

__

I +B+

B

2

2!

+ +

B

n

n!

_

= I +A+B+

1

2

_

A

2

+ 2AB+B

2

_

+ .

Also

e

A+B

= I +A+B+

(A+B)

2

2!

+

(A+B)

3

3!

+

= I +A+B+

1

2

_

A

2

+AB+BA+B

2

_

+ ,

which is the same as the above if AB = BA.

Theorem 9.5

Given T over C, there exists matrices P, D, and N with D diagonal,

N

k+1

= 0 for some k, and DN = ND such that T = P(D+N) P

1

.

9.3. THE 2 2 CASE IN DETAIL 129

Proof. Proof is given in MATB24.

Corollary 9.6

We have e

T

= Pe

D

e

N

P

1

(e

D

and e

N

can be computed as above).

Proof. Proof is given in MATB24.

The steps involved in nding P, N, and D are essentially the same as those

involved in putting T into Jordan Canonical form (over C).

9.3 The 2 2 Case in Detail

Theorem 9.7 (Cayley-Hamilton)

Let p() = [AI[. Then p(A) = 0.

Proof. Proof is given in MATB24.

The eigenvalues of A are the solutions of p() = 0. We now consider three

cases for the solutions of p() = 0.

Case 1: p() has two distinct real roots

1

and

2

. This is the diagonalizable

case considered earlier. We nd eigenvectors v and w for the eigenvalues

1

and

2

. Let

P =

_

v

1

w

1

v

2

w

2

_

.

Then

P

1

TP =

_

1

0

0

2

_

,

so

e

T

= P

_

e

1

0

0 e

2

_

P

1

.

The solution of Y

= TY would be

Y = P

_

e

1x

0

0 e

2x

__

C

1

C

2

_

130 CHAPTER 9. LINEAR SYSTEMS

in this case.

Case 2: p() has a double root r. Let S = TrI so that T = S +rI. Since

SI = IS, we have

e

T

= e

S

e

rI

= e

s

_

e

r

0

0 e

r

_

.

Then

p(T) = 0,

(TrI)

2

= 0,

S

2

= 0.

Therefore,

e

S

= I +S = I +TrI = T+ (1 r) I

and

e

T

= (T+ (1 r) I)

_

e

r

0

0 e

r

_

.

Case 3: p() has complex roots r iq, where q ,= 0. Let S = TrI so that

T = S +rI. Then

e

T

= e

S

_

e

r

0

0 e

r

_

and

p() =

2

2r +

_

r

2

+q

2

_

= ( r)

2

+q

2

.

We then have

p(T) = 0,

(TrI)

2

+ (qI)

2

= 0,

S

2

+ (qI)

2

= 0,

S

2

+q

2

I = 0.

9.3. THE 2 2 CASE IN DETAIL 131

Therefore, S

2

q

2

I and it follows that

e

S

= I +S +

S

2

2!

+

S

3

3!

+

= I +S

q

2

I

2!

q

2

S

3!

+

q

4

I

4!

+

q

4

S

5!

q

6

I

6!

+

= I

_

1

q

2

2!

+

q

4

4!

q

6

6!

+

_

+S

_

1

q

2

3!

+

q

4

5!

_

= I cos(q) +

S

q

sin(q).

Therefore,

e

T

=

_

cos(q)I +

_

TrI

q

_

sin(q)

_

_

e

r

0

0 e

r

_

.

Example 9.8 (Double root). Solve

Y

=

_

3 9

4 9

_

Y.

Solution. We have

p() = 0,

3 9

4 9

= 0,

(3 ) (9 ) + 36 = 0,

27 6 +

2

+ 36 = 0,

2

6 + 9 = 0,

( 3)

2

= 0.

Therefore, = 3 is a double root and we have Y = e

Ax

C. For A, the eigenvalue

132 CHAPTER 9. LINEAR SYSTEMS

is 3. Therefore, for T, the eigenvalue is r = 3x. Hence,

Y = (T+ (1 r) I)

_

e

r

0

0 e

r

_

= (Ax + (1 3x) I)

_

e

3x

0

0 e

3x

_

C

=

__

3x 9x

4x 9x

_

+

_

1 3x 0

0 1 3x

___

e

3x

0

0 e

3x

_

C

=

_

1 6x 9x

4x 1 + 6x

__

e

3x

0

0 e

3x

_

C

=

_

e

3x

6xe

3x

9xe

3x

4xe

3x

e

3x

+ 6xe

3x

__

C

1

C

2

_

=

_

C

1

_

e

3x

6xe

3x

_

9C

2

xe

3x

4C

1

xe

3x

+C

2

_

e

3x

+ 6xe

3x

_

_

= C

1

e

3x

_

1

4

_

+C

1

xe

3x

_

6

1

_

+C

2

e

3x

_

9

1

_

+C

2

xe

3x

_

9

6

_

.

Example 9.9 (Complex roots). Solve

Y

=

_

1 4

1 1

_

Y.

Solution. We have

p() = 0,

1 4

1 1

= 0,

(1 ) (1 ) + 4 = 0,

( + 1)

2

+ 4 = 0.

Therefore, = 1 2i. For A, the roots are 1 2i, so for T, the roots are

(1 2i) x, i.e., r = x and q = 2x. Note that

TrI

q

=

1

2x

_

0 4x

x 0

_

=

_

0 2

1

2

0

_

.

9.4. THE NON-HOMOGENEOUS CASE 133

Thus,

Y = e

Ax

C

=

_

cos(2x)I + sin(2x)

_

0 2

1

2

0

___

e

x

0

0 e

x

_

C

=

_

cos(2x) 2 sin(2x)

1

2

sin(2x) cos(2x)

__

e

x

0

0 e

x

_

C

=

_

e

x

cos(2x) 2e

x

sin(2x)

1

2

e

x

sin(2x) e

x

cos(2x)

__

C

1

C

2

_

=

_

C

1

e

x

cos(2x) + 2C

2

e

x

sin(2x)

1

2

C

1

e

x

sin(2x) +C

2

e

x

cos(2x)

_

.

9.4 The Non-Homogeneous Case

The general solution of Y

= AY + B is Y = Y

h

C + Y

p

, where Y

h

C is the

general solution of Y

= AY. Therefore, Y

h

C = AY

h

C for all C, giving us

Y

h

= AY

h

, (9.2a)

and Y

p

is a particular solution to Y

= AY +B, giving us

Y

p

= AY

p

+B. (9.2b)

To nd Y

p

, we use variation of parameters. Thus, we have

Y

p

= Y

h

V, (9.2c)

where

V =

_

v

1

v

2

_

.

134 CHAPTER 9. LINEAR SYSTEMS

To nd V, we have

Y

p

= Y

h

V,

Y

p

..

(9.2b)

= Y

h

..

(9.2a)

V+Y

h

V

,

AY

p

+B = AY

h

V

. .

(9.2c)

+Y

h

V

= AY

p

+Y

h

V

.

Therefore, Y

h

V

= B V

= Y

1

h

B.

Example 9.10. Solve

Y

=

_

4 1

3 2

_

Y +

_

e

5x

e

2x

_

.

1

= 1 and

2

= 5. Thus,

V

1

=

_

1 3

_

, V

2

=

_

11

_

.

Therefore,

P =

_

1 1

3 1

_

, D =

_

1 0

0 5

_

,

and the solution of Y

= AY is Y = Pe

Dx

C = Y

h

C. So we have

Y

h

= Pe

Dx

=

_

1 1

3 1

__

e

x

0

0 e

5x

__

e

x

e

5x

3e

x

e

5x

_

,

Y

1

h

= e

Dx

P

1

=

1

[P[

_

e

x

0

0 e

5x

__

1 1

3 1

_

=

1

4

_

e

x

e

x

3e

5x

e

5x

_

.

It follows that

V

= Y

1

h

B =

1

4

_

e

x

e

x

3e

5x

e

5x

__

e

5x

e

2x

_

=

1

4

_

e

4x

e

x

3 +e

3x

_

,

V =

1

4

_

1

4

e

4x

e

x

3x

1

3

e

3x

_

9.5. PHASE PORTRAITS 135

and

Y

p

= Y

h

V =

1

4

_

e

x

e

5x

3e

x

e

5x

__

1

4

e

4x

e

x

3x

1

3

e

3x

_

=

1

4

_

1

4

e

5x

e

2x

+ 3xe

5x

1

3

e

2x

3

4

e

5x

+ 3e

2x

+ 9xe

5x

1

3

e

2x

_

= e

5x

_

1

16

3

16

_

+e

2x

_

1

4

3

4

_

+xe

5x

_

3

4

9

4

_

+e

2x

_

1

12

1

12

_

.

Therefore,

Y = Y

h

C+Y

p

= C

1

e

x

_

1

3

_

+C

2

e

5x

_

1

1

_

+e

5x

_

1

16

3

16

_

+e

2x

_

1

4

3

4

_

+xe

5x

_

3

4

9

4

_

+e

2x

_

1

12

1

12

_

,

where C

1

and C

2

are arbitrary constants.

9.5 Phase Portraits: Qualitative and Pictorial

Descriptions of Solutions of Two-Dimensional

Systems

Let V : R

2

R

2

be a vector eld. Imagine, for example, that V(x, y) represents

the velocity of a river at the point (x, y).

the path that a leaf dropped in the river at the point (x

0

, y

0

) will follow. For

example, Figure 9.1 shows the vector eld of V(x, y) =

_

y, x

2

_

. Let (t) =

(x(t), y(t)) be such a path. At any time, the leaf will go in the direction that

the river is owing at the point at which it is presently located, i.e., for all t, we

have

dx

dt

= F(x, y)

. .

y

,

dy

dt

= G(x, y)

. .

x

2

.

We are assuming here that V depends only on the position (x, y) and not also upon time

t.

136 CHAPTER 9. LINEAR SYSTEMS

-4 -2 0 2 4

-4

-2

0

2

4

x

y

(a) The vector eld plot of

`

y, x

2

.

- 3 - 2 - 1 0 1 2 3 4

x

- 3

- 2

- 1

0

1

2

3

4

y

(b) The trajectories of

`

y, x

2

_

y, x

2

_

and its trajectories.

In general, it will be impossible to solve this system exactly, but we want to

be able to get the overall shape of the solution curves, e.g., we can see that in

Figure 9.1, no matter where the leaf is dropped, it will head towards (, ) as

t .

Before considering the general case, let us look at the linear case where we

can solve it exactly, i.e., V = (ax +by, cx +dy) with

dx

dt

= ax +by,

dy

dt

= cx +dy,

or X

= AX, where

X =

_

x

y

_

, A =

_

a b

c d

_

.

Recall from Theorem 10.12 (p. 157) that if all the entries of A are continuous,

then for any point (x

0

, y

0

), there is a unique solution of X

= AX satisfying

x(t

0

) = x

0

and y(t

0

) = y

0

, i.e., there exists a unique solution through each

point; in particular, the solution curves do not cross.

The above case can be solved explicitly, where

X = e

At

_

x

0

y

0

_

is a solution passing through (x

0

, y

0

) at time t = 0. We will consider only cases

9.5. PHASE PORTRAITS 137

where [A[ , = 0.

9.5.1 Real Distinct Eigenvalues

Let

1

and

2

be distinct eigenvalues of Aand let v and wbe their corresponding

eigenvectors. Let P = (v, w). Then

P

1

AP =

_

1

0

0

2

_

. .

D

.

Therefore, At = P(Dt) P

1

and we have

X = e

At

_

x

0

y

0

_

= Pe

Dt

P

1

_

x

0

y

0

_

= (v, w)

_

e

1t

0

0 e

2t

__

C

1

C

2

_

= (v, w)

_

C

1

e

1t

C

2

e

2t

_

= C

1

e

1t

v +C

2

e

2t

w.

Dierent C

1

and C

2

values give various solution curves.

Note that C

1

= 1 and C

2

= 0 implies that X = e

1t

v. If

1

< 0, then

the arrows point toward the origin as shown in Figure 9.2a in which contains a

stable node. Note that

X = C

1

e

1t

v +C

2

e

2t

w = e

2t

_

C

1

e

(12)t

v +C

2

w

_

.

The coecient of v goes to 0 as t , i.e., as t , X (0, 0), approaching

along a curve whose tangent is w. As t , X = e

1t

_

C

1

v +C

2

e

(21)t

w

_

,

i.e., the curves get closer and closer to being parallel to v as t .

We have the degenerate case when

1

<

2

= 0, in which case X = C

1

e

1t

v+

C

2

w.

The case when

1

< 0 <

2

gives us the phase portrait shown in Figure 9.2b

in which contains a saddle point. This occurs when [A[ < 0. The case when

0 <

1

<

2

gives us the phase portrait shown in Figure 9.2c in which contains

an unstable node. We have

X = C

1

e

1t

v +C

2

e

2t

w = e

1t

_

C

1

v +C

2

e

(21)t

w

_

.

Therefore, as t , X approaches parallel to w asymptotically.

Note that in all cases, the origin itself is a xed point, i.e., at the origin,

x

= 0 and y

138 CHAPTER 9. LINEAR SYSTEMS

0

0

0

0

(a)

1

<

2

< 0

0

0

0

0

(b)

1

< 0 <

2

0

0

0

0

(c) 0 <

1

<

2

Figure 9.2: The cases for

1

and

2

.

9.5. PHASE PORTRAITS 139

is called an equilibrium point; in a stable node, if it is disturbed, it will come

back in an unstable node; if perturbed slightly, it will leave the vicinity of the

origin.

9.5.2 Complex Eigenvalues

Complex eigenvalues come in the form = i, where ,= 0. In such a case,

we have

X = C

1

Re

_

e

t

v

_

+C

2

Im

_

e

t

v

_

,

where v = p +iq is an eigenvector for . Then

e

t

v = e

t

e

it

(p +iq)

= e

t

(cos(t) +i sin(t)) (p +iq)

= e

t

(cos(t)p sin(t)q +i cos(t)q +i sin(t)p) .

Therefore,

X = e

t

[(C

1

cos(t)p C

1

sin(t)q) +C

1

cos(t)q +C

2

sin(t)q]

= e

t

_

k

1

cos(t) +k

2

sin(t)

k

3

cos(t) +k

4

sin(t)

_

= e

t

_

cos(t)

_

k

1

k

3

_

+ sin(t)

_

k

2

k

4

__

.

Note that tr(A) = 2.

So

= 0 = tr(A) = 0,

> 0 = tr(A) < 0,

< 0 = tr(A) > 0.

Consider rst = 0. To consider the axes of the ellipses, rst note that

X =

_

p

1

q

1

p

2

q

2

_

. .

P

_

C

1

C

2

C

2

C

1

_

. .

C

_

cos(t)

sin(t)

_

.

Recall that the trace of a matrix is the sum of the elements on the main diagonal.

140 CHAPTER 9. LINEAR SYSTEMS

Except in the degenerate case, where p and q are linearly dependent, we have

_

cos(t)

sin(t)

_

= C

1

P

1

X.

Therefore,

cos(t) + sin(t)

_

cos(t)

sin(t)

_

= X

t

_

P

1

_

t

_

C

1

_

t

C

1

P

1

X,

cos

2

(t) + sin

2

(t) = X

t

_

P

1

_

t

_

C

1

_

t

C

1

P

1

X,

1 = X

t

_

P

1

_

t

_

C

1

_

t

C

1

P

1

X.

Note that C = C

t

, so

CC

t

=

_

C

1

C

2

C

2

C

1

__

C

1

C

2

C

2

C

1

_

C

2

=

_

C

2

1

+C

2

2

0

0 C

2

1

+C

2

2

_

=

_

C

2

1

+C

2

2

_

I.

Therefore,

C

1

=

C

C

2

1

+C

2

2

,

_

C

1

_

t

= C

1

=

C

C

2

1

+C

2

2

,

_

C

1

_

t

C

1

=

_

C

1

_

2

=

_

C

2

1

+C

2

2

_

I

(C

2

1

+C

2

2

)

2

=

I

C

2

1

+C

2

2

.

Therefore,

1 = X

t

_

P

1

_

t

_

C

1

_

t

C

1

P

1

X =

1

C

2

1

+C

2

2

X

t

_

P

1

_

t

P

1

X.

Let T =

_

P

1

_

t

P

1

. Then X

t

TX = C

2

1

+ C

2

2

and T = T

t

(T is symmetric).

Therefore, the eigenvectors of T are mutually orthogonal and form the axes of

the ellipses. Figure 9.3 shows a stable spiral and an unstable spiral.

9.5. PHASE PORTRAITS 141

0

0

0

0

(a) < 0

0

0

0

0

(b) > 0

Figure 9.3: The cases for , where we have a stable spiral when > 0 and an

unstable spiral when > 0.

9.5.3 Repeated Real Roots

We have N = AI, where N

2

= 0 and A = N+I. So

e

At

= e

Nt+tI

= e

Nt

e

tI

= (I +Nt)

_

e

t

0

0 e

t

_

.

Therefore,

X = e

At

_

C

1

C

2

_

= (I +Nt)

_

e

t

0

0 e

t

__

C

1

C

2

_

= e

t

(I +Nt)

_

C

1

C

2

_

.

Note that N

2

= 0 [N[

2

= 0 [N[ = 0. Therefore,

N =

_

n

1

n

2

n

1

n

2

_

.

Also, N

2

= 0 tr(N) = 0 n

1

+n

2

= 0. Let

V =

_

1

_

.

Then

Nv =

_

n

1

+n

2

(n

1

+n

2

)

_

=

_

0

0

_

.

142 CHAPTER 9. LINEAR SYSTEMS

So Av = (N+I) v = v, i.e., v is an eigenvector for . Therefore,

X = e

t

(I +Nt)

_

C

1

C

2

_

= e

t

_

C

1

+ (n

1

C

1

+n

2

C

2

) t

C

2

+(n

1

+n

2

C

2

) t

_

= e

t

__

C

1

C

2

_

+ (n

1

C

1

+n

2

C

2

) t

_

1

__

= e

t

__

C

1

C

2

_

+ (n

1

C

1

+n

2

C

2

) tv

_

.

If < 0, we have

_

x

y

_

_

0

0

_

as t . If > 0, then

_

x

y

_

_

0

0

_

as t . What is the limit of the slope? In other words, what line is

approached asymptotically? We have

lim

t

y

x

= lim

t

C

2

+ (n

1

C

1

+n

2

C

2

) tv

2

C

1

+ (n

1

C

1

+n

2

C

2

) tv

1

=

v

2

v

1

,

i.e., it approaches v. Similarly,

lim

t

y

x

=

v

2

v

1

,

i.e., it also approaches v as t . Figure 9.4 illustrates the situation.

We encounter the degenerate case when N = 0. This does not work, but

then A = I, so

X = e

At

_

C

1

C

2

_

=

_

e

t

0

0 e

t

__

C

1

C

2

_

= e

t

_

C

1

C

2

_

,

which is just a straight line through

_

C

1

C

2

_

.

Figure 9.5 illustrates this situation.

9.5. PHASE PORTRAITS 143

0

0

0

0

(a) < 0

0

0

0

0

(b) > 0

Figure 9.4: The cases for , where we have a stable node when < 0 and an

unstable node when > 0.

0

0

0

0

(a) < 0

0

0

0

0

(b) > 0

Figure 9.5: The degenerate cases for when N = 0, where we have a stable

node when < 0 and an unstable node when > 0.

144 CHAPTER 9. LINEAR SYSTEMS

Chapter 10

Existence and Uniqueness

Theorems

10.1 Picards Method

Consider the general rst order ivp

_

_

_

dy

dx

= f(x, y),

y(x

0

) = y

0

.

()

Picards Method is a technique for approximating the solution to Equation ().

If y(x) is a solution of Equation (), then

_

x

x0

dy

dt

dt =

_

x

x0

f(t, y(t)) dt,

_

y

y0

dy = y y

0

,

giving us

y = y

0

+

_

x

x0

f(t, y(t)) dt. ()

Conversely, if y satises (), then dierentiating gives

dy

dx

= 0 +f(x, y(x)) = f(x, y)

145

146 CHAPTER 10. EXISTENCE AND UNIQUENESS THEOREMS

and y(x

0

) = y

0

+ 0 = y

0

. Therefore, the ivp () is equivalent to the integral

equation ().

If p(x) is a function, dene a new function P(p) by

(P(p)) (x) = y

0

+

_

x

x0

f(t, p(t)) dt.

So a solution to () is a function y(x) such that P(y) = y.

Picards Method is as follows. Start with a function y

0

(x) for which y

0

(x

0

) =

y

0

. Let y

1

= P(y

0

), y

2

= P(y

1

), . . . , y

n

= P(y

n1

), . . ., and so on. Under the

right conditions (discussed later), the function y

n

will converge to a solution y.

Intuitively, we let y = lim

n

y

n

. Since y

n+1

= P(y

n

), taking limits as n

gives y = P(y). Later, we will attempt to make this mathematically precise.

Example 10.1. Solve y

= y, where y(0) = 1.

Solution. We begin with y

0

(x) 1. With x

0

= 0, y

0

= 1, f(x, y) = y, we

have

y

1

(x) = y

0

+

_

x

x0

f

_

1

x

, y

0

(t)

_

dt

= 1 +

_

x

0

(1) dt = 1 x,

and

y

2

(x) = y

0

+

_

x

x0

f(t, y

1

(t)) dt

= 1 +

_

x

0

(1 t) dt = 1 x +

x

2

2

,

and

y

3

(x) = 1 +

_

x

0

_

1 t +

t

2

2

_

dt = 1 x +

x

2

2

x

3

3!

.

So in general, we have

y

n

(x) = 1 x +

x

2

2!

+ + (1)

n

x

n

n!

.

Therefore, we nally have

y(x) = lim

n

y

n

(x) =

n=0

(1)

n

x

n

n!

.

10.1. PICARDS METHOD 147

In this case, we recognize the solution as y = e

x

.

Example 10.2. Solve y

= x +y

2

, where y(1) = 1.

Solution. Let y

0

(x) 1. With x

0

= 1, y

0

= 1, and f(x, y) = x +y

2

, we have

y

1

(x) = 1 +

_

x

1

_

t + 1

2

_

dt = 1 +

_

x

1

[(t 1) + 2] dt

= 1 +

_

(t 1)

2

2

+ 2 (t 1)

_

x

1

= 1 +

(x 1)

2

2

+ 2 (x 1)

= 1 + 2 (x 1) +

(x 1)

2

2

and

y

2

(x) = 1 +

_

x

1

_

_

t +

_

1 + 2 (t 1) +

(t 1)

2

2

_

2

_

_

dt

= 1 + 2 (x 1) +

5

2

(x 1)

2

+

5

3

(x 1)

3

+

1

4

(x 1)

4

+

1

20

(x 1)

5

.

Picards Method is not convenient for actually nding solutions, but it is

useful for proving that solutions exist under the right conditions.

Theorem 10.3

If f : X R is continuous, where X R

n

is closed and bounded (compact),

then f(X) is closed and bounded (compact). In particular, there exists an

M such that [f(x)[ M for all x X.

Proof. Proof is given in MATB43.

Corollary 10.4

Let R be a (closed) rectangle. Let f(x, y) be such that f/y exists and is

continuous throughout R. Then there exists an M such that

[f(x, y

2

) f(x, y

1

)[ M [y

2

y

1

[

. .

Lipschitz condition with respect to y

for any points (x, y

1

) and (x, y

2

) in R.

148 CHAPTER 10. EXISTENCE AND UNIQUENESS THEOREMS

Proof. Let R be a (closed) rectangle. Let f(x, y) be such that f/y exists and

is continuous throughout R. By Theorem 10.3, there exists an M such that

f

y

(r)

M

for all r R. By the Mean Value Theorem, there exists a c (y

1

, y

2

) such that

f(x, y

2

) f(x, y

1

) =

f

y

(x, c)(y

2

y

1

).

Since R is a rectangle, we have (x, c) R, so

f

y

(x, c)

M.

Therefore, [f(x, y

2

) f(x, y

1

)[ M [y

2

y

1

[.

Theorem 10.5

Let f be a function. Then

1. f is dierentiable at x implies that f is continuous at x.

2. f is continuous on B implies that f is integrable on B, i.e.,

_

B

f dV

exists.

Proof. Proof is given in MATB43.

Denition (Uniform convergence)

A sequence of functions f

n

dened on B is said to converge uniformly to a

function f if for all > 0, there exists an N such that [f(x) f

n

(x)[ < for all

n N.

The point is that the same N works for all xs. If each x has an Nx that worked for it

but there was no N working for all xs at once, then it would converge, but not uniformly.

10.1. PICARDS METHOD 149

Theorem 10.6

Suppose that f

n

is a sequence of functions that converges uniformly to f

on B. if f

n

is integrable for all n, then f is integrable and

_

B

f(x) dV = lim

n

_

B

f

n

(x) dV.

Proof. Proof is given in MATB43.

Theorem 10.7

Let f

n

be a sequence of functions dened on B. Suppose that there exists

a positive convergent series

n=1

a

n

such that [f

n

(x) f

n1

(x)[ a

n

for

all n. Then f(x) = lim

n

f

n

(x) exists for all x B and f

n

converges

uniformly to f.

Proof. Let f

n

be a sequence of functions dened on B. Suppose that there

exists a positive convergent series

n=1

a

n

such that [f

n

(x) f

n1

(x)[ a

n

for all n. Let

f

n

(x) = f

0

(x) + (f

1

(x) f

0

(x)) + + (f

n

(x) f

n1

(x))

= f

0

(x) +

n

k=1

(f

k

(x) f

k1

(x)) .

Therefore

lim

n

f

n

(x)

. .

f(x)

= f

0

(x) +

k=1

(f

k

(x) f

k1

(x))

and

[f(x) f

n

(x)[ =

k=n+1

(f

k

(x) f

k1

(x))

k=n+1

a

k

<

for n suciently large, since

n=1

a

n

converges. Therefore, f

n

converges uni-

formly to f.

150 CHAPTER 10. EXISTENCE AND UNIQUENESS THEOREMS

10.2 Existence and Uniqueness Theorem for First

Order ODEs

Theorem 10.8 (Existence and uniqueness theorem for rst order odes

)

Let R be a rectangle and let f(x, y) be continuous throughout R and satisfy

the Lipschitz Condition with respect to y throughout R. Let (x

0

, y

0

) be an

interior point of R. Then there exists an interval containing x

0

on which

there exists a unique function y(x) satisfying y

0

) = y

0

.

To prove the existence statement in Theorem 10.8, we will need the following

lemma.

Lemma 10.9

Let I = [x

0

, x

0

+] and let p(x) satisfy the following:

1. p(x) is continuous on I.

2. [p(x) y

0

[ a for all x I.

Then

_

x

x0

f(t, p(t)) dt exists for all x I and q(x) = y

0

+

_

x

x0

f(t, p(t)) dt

also satises (1) and (2).

Proof. Let I = [x

0

, x

0

+] and let p(x) satisfy the following:

1. p(x) is continuous on I.

2. [p(x) y

0

[ a for all x I.

Immediately, (2) implies that (t, p(t)) R for t I. We have

I R

f

R

t (t, p(t))

so f is continuous, hence integrable, i.e.,

_

x

x0

f(t, p(t)) dt exists for all x

I. Since q(x) is dierentiable on I, it is also continuous on I. Also, since

10.2. EXISTENCE ANDUNIQUENESS THEOREMFOR FIRST ORDER ODES151

[f(t, p(t))[ M, it follows that

[q(x) y

0

[ =

_

x

x0

f(t, p(t)) dt

M [x x

0

[ M a.

let f(x, y) be continuous throughout R that satisfy the Lipschitz Condition with

respect to y throughout R. Let (x

0

, y

0

) be an interior point of R. Since f is

continuous, there exists an M such that [f(r)[ M for all r R. Let a be the

distance from (x

0

, y

0

) to the boundary of R. Since (x

0

, y

0

) is an interior point,

we have a > 0. Let = min(a, a/M). We will show that y

y(x

0

) = y

0

, has a unique solution on the domain I = [x

0

, x

0

+].

Inductively dene functions (1) and (2) of Lemma 10.9 by y

0

(x) y

0

for all

x I and

y

n

(x) = y

0

+

_

x

x0

f(t, y

n1

(t)) dt.

By Lemma 10.9, the existence of y

n1

satisfying (1) and (2) guarantees the

existence of y

n

. By hypothesis, there exists an A such that [f(x, v) f(x, w)[

A[v w[, whenever (x, v), (x, w) R. As in the proof of Lemma 10.9, we have

[y

1

(x) y

0

(x)[ = [y

1

(x) y

0

[ M [x x

0

[

for all x I. Likewise, we have

[y

2

(x) y

1

(x)[ =

_

x

x0

[f(t, y

1

(t)) f(t, y

0

(t))] dt

_

x

x0

[f(t, y

1

(t)) f(t, y

0

(t))[ dt

_

x

x0

A[y

1

(t) y

0

(t)[ dt

_

x

x0

AM [t x

0

[ dt

MA

[x x

0

[

2

2

,

152 CHAPTER 10. EXISTENCE AND UNIQUENESS THEOREMS

and

[y

3

(x) y

2

(x)[ =

_

x

x0

[f(t, y

2

(t)) f(t, y

1

(t))] dt

_

x

x0

[f(t, y

2

(t)) f(t, y

1

(t))[ dt

_

x

x0

_

AMA

[t x

0

[

2

2

_

dt

MA

2

[x x

0

[

3

3!

.

Continuing, we get

[y

n

(x) y

n1

(x)[

MA

n1

n

n!

.

We have

y

n

(x) = y

0

(x) + (y

1

(x) y

0

(x)) + (y

2

(x) y

1

(x)) + + (y

n

(x) y

n1

(x))

= y

0

(x) +

n

k=1

(y

k

(x) y

k1

(x)) .

Since

[y

k

(x) y

k1

(x)[

MA

k1

k

k!

and

k=1

MA

k1

k

k!

=

M

A

_

e

A

1

_

,

converges, we must have

lim

n

y

n

(x) = y

0

(x) +

k=1

(y

k

(x) y

k1

(x))

converging for all x I. Therefore, y(x) = lim

n

y

n

(x) exists for all x I

and y

n

converges uniformly to y.

Note that

[f(x, y

n

(x)) f(x, y

n1

(x))[ A[y

n

(x) y

n1

(x)[

10.2. EXISTENCE ANDUNIQUENESS THEOREMFOR FIRST ORDER ODES153

implies that f(x, y

n

(x)) converges uniformly to f(x, y(x)). With

y

n

(x) = y

0

+

_

x

x0

f(y, y

n1

(t)) dt,

we then have

y(x) = lim

n

y

n

(x) = lim

n

_

y

0

+

_

x

x0

f(t, y

n1

(t)) dt

_

= y

0

+

_

x

x0

_

lim

n

f(t, y

n1

(t))

_

dt = y

0

+

_

x

x0

f(t, y(t)) dt

implying that y satises y

0

) = y

0

.

To prove the uniqueness statement in Theorem 10.8, we will need the fol-

lowing lemma.

Lemma 10.10

If y(x) is a solution to the de with y(x

0

) = y

0

, then [y(x) y

0

[ a for all

x I. In particular, (x, y(x)) R for all x I.

Proof. Suppose there exist an x I such that [y(x) y

0

[ > a. We now consider

two cases.

Suppose x > x

0

. Let s inf(t : t > x

0

and [y(t) y

0

[ > a). So s < x

x

0

+ x

0

+a. By continuity, we have

a = [y(s) y

0

[ = [y(s) y(x

0

)[ = [y

(c) (s x

0

)[ = [f(c, y(c)) (s x

0

)[

. .

M(sx0)

for some c (x

0

, s). However,

s I = s x

0

= M (s x

0

) M a.

Therefore,

M (s x

0

) = Ma = s x

0

= a = s = x

0

+a,

which is a contradiction because we have s < x

0

+ a. The case with x < x

0

is

similarly shown.

Proof (Proof of uniqueness statement (Theorem 10.8)). Suppose y(x) and z(x)

154 CHAPTER 10. EXISTENCE AND UNIQUENESS THEOREMS

both satisfy the given de and initial condition. Let (x) = (y(x) z(x))

2

. Then

(x) z

(x))

= 2 (y(x) z(x)) (f(x, y(x)) f(x, z(x))) .

Therefore for all x I,

[

= 2A(y(x) z(x))

2

= 2A(x).

Therefore, 2A(x)

(x) 2A(x) =

(x) 2A(x) 0

= e

2Ax

(

(x) 2A(x)) 0,

(x) =

d

_

e

2Ax

(x)

_

dx

= e

2Ax

(x) decreasing on I

= e

2Ax

(x) e

2Ax0

(x

0

)

= (x) 0

if x x

0

. Similarly,

(x) 2A(x) =

(x) + 2A(x) 0

= e

2Ax

(

(x) + 2A(x)) 0,

(x) =

d

_

e

2Ax

(x)

_

dx

= e

2Ax

(x) increasing on I

= e

2Ax

(x) e

2Ax0

(x

0

)

= (x) 0

if x x

0

. Therefore, (x) 0 for all x I. But obviously, (x) 0 for all

x from its denition. Therefore, (x) 0, i.e., y(x) z(x). Therefore, the

solution is unique.

10.3. EXISTENCE ANDUNIQUENESS THEOREMFOR LINEAR FIRST ORDER ODES155

10.3 Existence and Uniqueness Theorem for Lin-

ear First Order ODEs

Here, we consider the special case of a linear de

dy

dx

= a(x)y +b(x).

Theorem 10.11 (Existence and uniqueness theorem for linear rst order des)

Let a(x) and b(x) be continuous throughout an interval I. Let x

0

be an

interior point of I and let y

0

be arbitrary. Then there exists a unique

function y(x) with y(x

0

) = y

0

satisfying dy/dx = a(x)y + b(x) throughout

I.

0

be

an interior point of I and let y

0

be arbitrary. Let A = max([a(x)[ : x I).

Inductively dene functions

on I by y

0

(x) = y

0

and

y

n+1

(x) = y

0

+

_

x

x0

(a(t)y

n

(t) +b(t)) dt.

Then

[y

n

(x) y

n1

(x)[ =

y

0

+

_

x

x0

(a(t)y

n1

(t) +b(t)) dt

y

0

_

x

x0

(a(t)y

n2

(t) +b(t)) dt

_

x

x0

[a(t) (y

n1

(t) y

n2

(t))] dt

.

Assuming by induction that

[y

n1

(x) y

n2

(x)[ A

n2

[x x

0

[

n

(n 1)!

,

out I and thus integrable, so the denition of y

n+1

(x) makes sense. Then y

n+1

(x) is also

dierentiable, i.e., y

n+1

(x) = a(x)yn(x) + b(x), and thus continuous, so the induction can

proceed.

156 CHAPTER 10. EXISTENCE AND UNIQUENESS THEOREMS

we then have

_

x

x0

[a(t) (y

n1

(t) y

n2

(t))] dt

_

x

x0

_

AA

n2

[t x

0

[

n1

(n 1)!

_

dt

A

n1

[x x

0

[

n

n!

,

thereby completing the induction. Therefore,

[y

n

(x) y

n1

(x)[

A

n1

n

n!

,

where is the width of I. The rest of the proof is as before.

10.4 Existence and Uniqueness Theorem for Lin-

ear Systems

A system of dierential equations consists of n equations involving n functions

and their derivatives. A solution of the system consists of n functions having

the property that the n functional equations obtained by substituting these

functions are their derivatives into the system of equations, and they hold for

every point in some domain D. A linear system of dierential equations has the

form

dy

1

dx

= a

11

(x)y

1

+a

12

(x)y

2

+ +a

1n

(x)y

n

+b

1

(x),

dy

2

dx

= a

21

(x)y

1

+a

22

(x)y

2

+ +a

2n

(x)y

n

+b

2

(x),

.

.

.

dy

n

dx

= a

n1

(x)y

1

+a

n2

(x)y

2

+ +a

nn

(x)y

n

+b

n

(x).

We can write the system as

dY

dx

= AY +B,

1

, y

2

, . . . , yn and their derivatives appears linearly.

10.4. EXISTENCE ANDUNIQUENESS THEOREMFOR LINEAR SYSTEMS157

where

A =

_

_

a

11

(x) a

1n

(x)

a

21

(x) a

2n

(x)

.

.

.

.

.

.

.

.

.

a

n1

(x) a

nn

(x)

_

_

, Y =

_

_

y

1

(x)

y

2

(x)

.

.

.

y

n

(x)

_

_

, B =

_

_

b

1

(x)

b

2

(x)

.

.

.

b

n

(x)

_

_

.

Theorem 10.12 (Existence and uniqueness theorem for linear systems of n des)

Let A(x) be a matrix of functions, each continuous throughout an interval

I and let B(x) be an n-dimensional vector of functions, each continuous

throughout I. Let x

0

be an interior point of I and let Y

0

be an arbitrary

n-dimensional vector. Then there exists a unique vector of functions Y(x)

with Y(x

0

) = Y

0

satisfying

dY

dx

= A(x)Y +B(x)

throughout I.

terval I and let B(x) be an n-dimensional vector of functions, each continuous

throughout I. Let x

0

be an interior point of I and let Y

0

be an arbitrary

n-dimensional vector.

Let = max(|A(x)| : x I). Inductively dene Y

0

(x) = Y

0

and

Y

n+1

(x) +Y

0

+

_

x

x0

(A(t)Y

n

(t) +B(t)) dt

. .

vector obtained by

integrating componentwise

.

158 CHAPTER 10. EXISTENCE AND UNIQUENESS THEOREMS

Then

[Y

n

(x) Y

n1

(x)[ =

_

x

x0

[A(t) (Y

n1

(t) Y

n2

(t))] dt

_

x

x0

_

n1

[t x

0

[

n1

(n 1)!

_

dt

=

n

[x x

0

[

n

n!

n

n

,

where is the width of I. The rest of the proof is essentially the same as

before.

Recall Theorem 8.5 (p. 90) restated here as follows.

Theorem 10.13 (Theorem 8.5)

Let x = p be an ordinary point of y

+ P(x)y

distance from p to the closest singular point of the de in the complex plane.

Then the de has two series y

1

(x) =

n=0

a

n

x

n

and y

2

(x) =

n=0

b

n

x

n

which converge to linearly independent solutions to the de on the interval

[x p[ < R.

+ P(x)y

+ Q(x)y = 0. Let R be

the distance from p to the closest singular point of the de in the complex plane.

We rst claim that if f(z) is analytic at p and R is the radius of convergence

of f, then for r < R, there exists an M (depending on r) such that

1

n!

f

(n)

(p)

M

r

n

()

for all n N. To prove this claim, since

n=0

1

n

f

n

(p)r

n

= f(r)

converges absolutely, let

M =

n=0

1

n!

[f

n

(p)[ r

n

.

10.4. EXISTENCE ANDUNIQUENESS THEOREMFOR LINEAR SYSTEMS159

Then

1

n!

f

(n)

(p)r

n

M,

so Equation () holds. Therefore, our claim holds.

Let w(z) be a solution of y

+ P(x)y

the de, w is twice dierentiable. Furthermore, dierentiating the de gives a

formula for w

in terms of w, w

, w

for all n, so w has a Taylor series. We now wish to show that the Taylor series

of w converges to w for [z p[ < R.

Since [z p[ < R, there exists an a such that [z p[ < a < R. As above,

nd constants M such that

1

n!

P

(n)

(p)

M

a

n

,

1

n!

Q

(n)

(p)

N

a

n

for all n N. Let

A(z) =

M

1

zp

a

, B(z) =

N

1

zp

a

.

Note that

A(z) = M

_

1 +

z p

a

+

_

z p

a

_

2

+ +

_

z p

a

_

n

+

_

,

so

A(p) = M,

A

(p) =

M

a

,

A

(p) = 2

M

a

2

,

.

.

.

A

(n)

(p) = n!

M

a

n

P

(n)

(p)

.

Similarly,

B

(n)

(p) = n!

N

a

n

Q

(n)

(p)

.

Consider

y

= A(z)y

+B(z)y. ()

160 CHAPTER 10. EXISTENCE AND UNIQUENESS THEOREMS

Let v(z) be a solution of Equation () satisfying v(p) = [w(p)[ and v

(p) =

[w

(p)[. In general, w

(n)

(p) = C

1

w

(p) +C

2

w(p), where C

1

and C

2

are arbitrary

constants depending on P and Q and their derivatives at p, e.g.,

w

= P(p)w

(p) +Q(p)w(p),

w

= P

(p)w

(p) +P(p)w

(p) +Q

(p)w(p) +Q(p)w

(p)

= P

(p)w

(p) +Q(p)w(p)) +Q

(p)w(p) +Q(p)w

(p)

=

_

P

2

(p) +P

(p) +Q(p)

_

w

(p) + (P(p)Q(p) +Q

(p)) w(p).

Similar formulas hold v

(n)

(p) involving derivatives of A and B at p.

Using

P

(n)

(p)

A

(n)

(p) and

Q

(n)

(p)

B

(n)

(p) gives

w

(n)

(p)

v

(n)

(p).

The Taylor series of W about p is

n=0

a

n

(z p)

n

, where a

n

= w

(n)

(p)/n!;

the Taylor series of v about p is

n=0

b

n

(z p)

n

, where b

n

= v

(n)

(p)/n!.

Since we showed that [a

n

[ b

n

, the rst converges anywhere the second does.

So we need to show that the Taylor series of w(x) has a radius of convergence

equal to a (this implies that the Taylor series for w converges for n pk, but

a < R was arbitrary). We have

v

=

M

1

zp

a

v

+

N

1

zp

a

v.

Let u = (z p) /a. Then

v

=

dv

du

1

a

dv

du

.

v

= a

d

2

v

du

2

du

dz

=

1

a

2

d

2

v

du

2

.

Then we have

1

a

2

d

2

v

du

2

=

M

1 u

1

a

dv

du

+

N

1 u

v,

(1 u)

d

2

v

du

2

= aM

dv

du

+a

2

Nv.

Write v =

n=0

n

u

n

. Using its rst radius of convergence, we have

v =

n

u

n

=

n

_

z p

a

_

n

=

n

a

n

(z p)

n

= b

n

(z p)

n

,

10.4. EXISTENCE ANDUNIQUENESS THEOREMFOR LINEAR SYSTEMS161

which implies that

n

= b

n

a

n

> 0. Note that

dv

du

=

n=1

n

n

u

n1

. .

adjust indices

=

n=0

(n + 1)

n+1

u

n

,

d

2

v

du

2

=

n=1

(n + 1) n

n+1

u

n1

. .

adjust indices

=

n=0

(n + 2) (n + 1)

n+2

u

n

.

Now we have

(1 u)

n=0

(n + 2) (n + 1)

n+2

u

n

=

n=0

aM (n + 1)

n+1

u

n

+

n=0

a

2

N

n

u

n

,

_

_

_

_

_

_

n=0

(n + 2) (n + 1)

n+2

u

n

n=0

(n + 2) (n + 1)

n+2

u

n+1

_

_

_

_

_

_

=

_

_

_

_

_

_

n=0

(n + 2) (n + 1)

n+2

u

n

n=0

(n + 1) n

n+1

u

n

.

_

_

_

_

_

_

Therefore,

n=0

(n + 2) (n + 1)

n+2

u

n

=

n=0

_

(n + 1) (n +aM)

n+1

+a

2

N

n

u

n

,

so it follows that

(n + 2) (n + 1)

n+2

= (n + 1) (aM +n)

n+1

+a

2

N

n

,

n+2

=

aM +n

n + 2

n+1

+

a

2

N

n

n + 2

,

n+2

n+1

=

n +aM

n + 2

+

a

2

N

n + 2

n+1

for all n N. Since

n

/

n+1

< 1, we have

lim

n

a

2

N

n + 2

n+1

= 0.

Therefore,

lim

n

n+2

n+1

= 1 + 0 = 1.

So by the ratio test (Equation (8.2) of Theorem 8.2, p. 88), the radius conver-

162 CHAPTER 10. EXISTENCE AND UNIQUENESS THEOREMS

gence of

n=0

n

u

n

=

n=0

n

a

n

(z p)

n

=

n=0

b

n

_

z p

a

_

n

= 1

is a.

The point of Theorem 8.5 is to state that solutions are analytic (we already

know solutions exists from early theorems). The theorem gives only a lower

bound on the interval. The actual radius of convergence may be larger.

Chapter 11

Numerical Approximations

11.1 Eulers Method

Consider y

0

) = y

0

. The goal is to nd y(x

end

). Figure

11.1 shows this idea. We pick a large N and divide [x

0

, x

end

] into segments of

actual value

approximate value

x

0

x

0

+h

= x

1

x

0

+2h

= x

2

...

x

end

= x

N

y

0

y

1

y

2

y

N

Figure 11.1: Graphical representation of Eulers method.

length h = (x

end

x

0

) /N. Then

y(x

1

) y(x

0

) +y

(x

0

) (x

1

x

0

)

. .

h

163

164 CHAPTER 11. NUMERICAL APPROXIMATIONS

is the linear/tangent line approximation of y(x

1

). Set

y

1

= y

0

+hf(x

0

, y

0

) y

1

f(x

1

),

y

2

= y

1

+hf(x

1

, y

1

) y

2

f(x

2

),

y

3

= y

2

+hf(x

2

, y

2

) y

3

f(x

3

),

.

.

.

y

answer

= y

N1

+hf(x

N1

, y

N1

) y

answer

f(x

end

).

Example 11.1. Consider y

using a step size of h = 0.1.

Solution. First note that x

0

= 0. With a step size of h = 0.1, we have

x

1

= 0.1, x

2

= 0.2, x

3

= 0.3.

Therefore,

y

1

= y

0

+hf(x

0

, y

0

) = 1 + 0.1 (0 2)

= 1 0.2 = 0.8,

y

2

= 0.8 +hf(0.1, 0.8) = 0.8 +h(0.1 1.6)

= 0.8 0.1 1.5 = 0.8 0.15 = 0.65,

y

3

= 0.65 +hf(0.2, 0.65) = 0.65 +h(0.2 1.3)

= 0.65 0.1 1.1 = 0.65 0.11 = 0.54.

Therefore, y(0.3) 0.54.

What about the actual value? To solve the de, we have

y

+ 2y = x,

e

2x

y

+ 2e

2x

y = xe

2x

,

ye

2x

=

_

xe

2x

dx =

1

2

xe

2x

_

1

2

e

2x

dx

=

1

2

xe

2x

1

4

e

2x

+C,

11.1. EULERS METHOD 165

where C is an arbitrary constant. Therefore

y =

1

2

x

1

4

+Ce

2x

.

Note that

y(0) = 1 =

1

4

+C = 1 = C =

5

4

,

so the general solution is

y =

1

2

x

1

4

+

5

4

e

2x

.

Therefore,

y(0.3) = 0.15 0.25 +

5

4

e

0.6

0.1 + 0.68601 0.58601,

so y(0.3) = 0.58601 is the (approximated) actual value.

11.1.1 Error Bounds

The Taylor second degree polynomial of a function y with step size h is given

by

y(a +h) = y(a) +hy

(a) +h

2

y

(z)

2

for some z [a, h], where h

2

y

in each step. We have

h

2

y

(z)

2

M

2

h

2

if [y

(z)[ M on [a, h]. Therefore, dividing h by, say, 3 reduces the local trun-

cation error by 9. However, errors also accumulate due to y

n+1

being calculated

using the approximation of y

n

instead of the exact value of f(x

n

). Reducing h

increases N, since N = (x

end

x

0

) /h, so this eect wipes out part of the local

gain from the smaller h. The net eect is that

(overall error global truncation error)

Mh,

e.g., dividing h by 3 divides the overall error only by 3, not 9.

166 CHAPTER 11. NUMERICAL APPROXIMATIONS

11.2 Improved Eulers Method

To improve upon Eulers method, we consider

y

n+1

= y

n

h,

where, instead of using f

(x

n

), it is better to use the average of the left and

right edges, so

=

f

(x

n

) +f

(x

n+1

2

=

f(x

n

, y

n

) +f(x

n+1

, y

n+1

)

2

.

Figure 11.2 illustrates this idea. However, we have one problem in that we do

x

n

x

n+1

Figure 11.2: A graphical representation of the improved Eulers method.

not know y

n+1

(that is what we are working out at this stage). Instead, we ll

into the inside y

n+1

from the original Eulers method, i.e.,

1

2

(f(x

n

, y

n

) +f(x

n+1

, y

n

+hf(x

n

, y

n

))) ,

i.e.,

y

n+1

= y

n

+

h

2

[f(x

n

, y

n

) +f(x

n+1

, y

n

+hf(x

n

, y

n

))] .

This improves the local truncation error from h

2

to h

3

.

11.3. RUNGE-KUTTA METHODS 167

11.3 Runge-Kutta Methods

We again modify Eulers method and consider

y

n+1

= y

n

h .

At each step, for , we use some average of the values over the interval [y

n

, y

n+1

].

The most common one is x

n

+h/2. Figure 11.3 illustrates this idea. Set

x

n

x

n + h/2

x

n + 1

= x

n

+ h

Figure 11.3: A graphical representation of the Runge-Kutta Methods.

k

n1

= f(x

n

, y

n

),

k

n2

= f

_

x

n

+

1

2

h, y

n

+

1

2

hk

n1

_

,

k

n3

= f

_

x

n

+

1

2

h, y

n

+

1

2

hk

n2

_

,

k

n4

= f

_

x

n

+

1

2

h, y

n

+

1

2

hk

n3

_

.

Use

y

n+1

= y

n

+

h

6

(k

n1

+ 2k

n2

+ 2k

n3

+ 3k

n4

) .

This gives a local truncation error proportional to h

5

.

168 CHAPTER 11. NUMERICAL APPROXIMATIONS

- 2013FA-MATH-2420-81001Uploaded bySpoodie
- Fundamentals of Ordinary Differential EquationsUploaded byAshoka Vanjare
- Gate 2018 SyllabusUploaded byAjeet Kumar
- B. Arch. (I - X).pdfUploaded bysilvinus clisson
- using ode45Uploaded byhmsedighi459
- L6 IA 2011 U1 M1Uploaded byRonaldo95163
- Indivi_syllabi_I sem.docxUploaded bySiva Raman
- 2ndUploaded byUMER HANEEF
- Theory of ODE - HuUploaded byFavio90
- EquationsUploaded bySushobhan Sanyal
- 4-1.Differential EquationsUploaded byLutfi Amin
- CIV340 Mock ExamUploaded byapi-3707273
- EE203_04_ode3_jan13_14sUploaded byAsif Khan
- 02_6243 new B__pp 15-23 new2Uploaded byTarig Elzaki
- Linear Algebra and Ordinary Differential Equations (Riaz) (1)Uploaded byMunir Shah
- Math 311 Numerical Solutions of Ordinary Differential Equations.docUploaded bymartinariel79
- Differential Equations - Ordinary Differential Equations - Integrating FactorsUploaded byMaxEconomics
- It Curriculum BookUploaded bysrinureddy2014
- Project 309Uploaded byAditi Mondal
- Matematici speciale. ProblemeUploaded bysg8_live
- RK4SystemEquations.pdfUploaded bySylas Trovarius
- Derivadas de Cualquier Orden 1f1Uploaded bymaferio
- Math 53 Study GuideUploaded byBabuji
- j.euromechsol.2012.12.008_adbce7de4b70cf372e5a6562b73532cdUploaded byasgharhabibi
- Chapter 1 ExercisesUploaded bytharunenjp
- Modelsystems CourseNotes CompiledNotes Week1 Lec1 3Uploaded byAnonymous y05RLoEDV
- NS SolutionUploaded bykgupta27
- Treatise on Cal Cul 00 GeorUploaded bynewtonfeyn
- EquationsUploaded byPrabhawana Shailesh
- Finite DifferencesUploaded byhlvijaykumar

- About the Various Definitions of Life - Pier Luigi LuisiUploaded byKris Cowell Llorens
- f1040Uploaded byarnodiaz
- Electronic Devices and CircuitsUploaded byKing Ashok
- a.4.8Uploaded bymemrah2955
- BiomedicineUploaded bymemrah2955
- Transistor 3Uploaded byMo Ibra
- ECS203 - handout 3B.pdfUploaded bymemrah2955
- I TutorialUploaded byUday Kumar
- Burrows2_90_FourierTransform UnitStepFunctionUploaded byVivek Kumar
- SIUnits NoteUploaded byKarthikeyan Krishnan
- 2 Steady-State Response of Electrodes FullUploaded bymemrah2955
- Simulink HandoutUploaded byCinthia Farfán
- Nanoparticle FuntionalizationUploaded bymemrah2955
- Ash Basic Probability Theory (Dover)Uploaded byCosmin Burtea
- Routh Hurwitz AnalysisUploaded byilg1
- FP_356Uploaded bymemrah2955
- HN Pub 4 - Resp ModelsUploaded bymemrah2955
- 200608A12Uploaded bymemrah2955
- 2006MXu PA ReviewUploaded bymemrah2955
- 23. BIOMEDUploaded bymemrah2955
- JPD_36_2003_R167Uploaded bymemrah2955
- Bergere ThUploaded bymemrah2955
- Vectors SummaryUploaded bymemrah2955
- Bachelorthesis Van Anton Delawari en Reinier Doelman Juni 2010Uploaded bymemrah2955
- Intro to LV in 3 Hours for Control and Sim 8_5Uploaded bymemrah2955
- duality(2).pdfUploaded bymemrah2955
- InnerUploaded bymemrah2955
- DiyUploaded bysphericaly
- Paper1for10-18Uploaded bymemrah2955
- Lab4 u10 Multsim OScopeUploaded bynonreplica

- 5 Quantification in PC.pdfUploaded byMilosh Petrovic
- FAWithOutputChapter8.pptUploaded byvuslat88
- r Eschke 2004Uploaded byanthonymcfay9
- sinuous antenna by duhamel, reanalyzed, and builtUploaded byjoop
- Needle PenetrationUploaded byGopal Krishan
- Ship Margin Line Code of Federal RegulationsUploaded byAnonymous JSHUTp
- Basis for RadrommetUploaded byRagnhild Kosmo Holm
- Neural Nets and Deep LearningUploaded byManu Carbonell
- Para Pblms KeyUploaded bysuperkoolkat
- Matlab CodeUploaded byAvinash Singh
- Sine Cosine Based Algorithm for Data ClusteringUploaded byAnonymous lPvvgiQjR
- Lectures 06`Uploaded byPetre Oara
- Matrix Converter Induction Motor Drive- Modeling- SimulationUploaded byjoseph86177
- Improving learning in mathsUploaded byReinaldo Ramirez
- Wiley_ Differential Equations With Control Theory_ an Introduction to Modern Methods and Applications, 3rd Edition - James R. Brannan, William EUploaded bybhar
- Satellite Services Workshop II Volume 2Uploaded byBob Andrepont
- E CAD NOTES.docxUploaded byRahulBalu
- Acrostic Dictionary of Wordplay TermsUploaded byd-fbuser-41605158
- Template Matching for Detection of Starry Milia-like Cysts in Dermoscopic ImagesUploaded byViswanaath Subramanian
- Reasons For Proc SQLUploaded byPaulo Cristóvão
- Penyelesaian Persamaan LinearUploaded bynsa_da
- Position, Orientation and Coordinate TransformationsUploaded byVeronicaNistor
- Topic 3 - Flexible Pavement Stress Analysis HandoutsUploaded byjulspd
- Scour Around Coastal StructuresUploaded byJesus Campo
- f(1).txtUploaded byAadityaBadhan
- Visual Basic Language Specification 9.0Uploaded bynemrakm
- Q Ch 9Uploaded byJhon F Sinaga
- UTRAN Cell User Throughput(Day) P612 (2)Uploaded byRaju Sutradhar
- Tcs Sample QuestionsUploaded bySilva Panda
- UML NotationsUploaded bymk_valiant