Elements of Engineering

Mathematics I
Chapter 1: ODEs
College of ICE
Sungkyunkwan University

Prof. Bongshik Song (송봉식 교수)

1
Flipped Class

2
3
Contents

• Chapter 1. First Order Ordinary Differential Equations

– 1.1 Basic Concepts. Modeling
– 1.2 Geometric Meaning of y′=f(x,y) : Directions Fields
– 1.3 Separable ODEs.
– 1.4 Modeling : Separable OEs.
– 1.5 Exact ODEs. Integrating Factors
– 1.6 Linear ODEs. Bernoulli Equation
– 1.7 Existence and Uniqueness of Solutions

4
1.1 Basic Concepts. Modeling

5
Basic Concepts. Modeling
• Differential Equations
– The most frequent encounters
in engineering problem solving
– Physical, mechanical, and
electrical phenomena

6
Basic Concepts. Modeling
• Ordinary Differential Equations (ODEs)
– Single independent variable
y = cos x
y + 9 y = 0
x 2 yy + 2e x y = ( x 2 + 2) y 2

• Partial differential equations (PDEs)

– Two or more independent variables involved.

 2u  2u
+ 2 =0
x 2
y

7
Basic Concepts. Modeling
• Order of ODE
– First derivative of the unknown function
– First-Order ODEs
• Form 1. (Implicit form)
F ( x, y, y ') 0

Ex) x 3 y ' 4 y 2 0

• Form 2. (Explicit form)

y' f ( x, y )

Ex) y ' 4 x3 y 2

8
Basic Concepts. Modeling
• Solution of ODE
– A solution of a given first-order differential equation, F(x,y,y′)=0, on
some open interval a<x<b, is a function y=h(x) that has a derivative
y′=h′(x) and satisfies F(x,y,y′)=0 for all x in that interval.

– Explicit solution : y=f(x)

xy ' 2y y x2

– Implicit solution :

yy ' x x2 y2 1

9
Basic Concepts. Modeling
• Example 2. Solution Curves

10
Basic Concepts. Modeling
• Solution Type
– General solution
• To satisfy differential equation, including the integration constant c
solution
y' cos x y sin x c

– Particular solution
• Determine the conditions of integration constant c
solution
y' cos x y sin x 2

– Singular solution
• Impossible to be represent as general solution
general solution
y '2 xy ' y 0 y c( x c)
1 2
y x is also a solution but not general solution. ⇒ Singular solution
4
11
Basic Concepts. Modeling
• Example 3 : Exponential Decay

y = ce ( c any constant )
at
⎯⎯ ⎯ ⎯
⎯→
Differenti ate
y =
dy
dt
( )
= a ce at = ay

y = ce at is a solution of y  = ay

y ce3t ( c: constant )
dy
y 3ce3t 3 y
dt
y ce3t a solution of y 3y
A general solution

12
Basic Concepts. Modeling
• Initial Value Problems
– ODE with initial condition

y f ( x, y ), y(x0 ) y0

– Unique solution (particular solution) is obtained from a general

solution with an initial condition y(x0)=y0.

– Which determines arbitrary constant c.

– Geometrical meaning: solution curve pass through the given point.

13
Basic Concepts. Modeling
• Example 4: Initial value problem
dy
y 3 y, y (0) 5.7
dx

– The general solution

y ( x) = ce3 x

– Determines c by the initial condition

y (0) ce0 c 5.7

– Particular solution
y ( x) 5.7e3 x

14
Basic Concepts. Modeling
• Modeling
– Step 1. The transition from the physical situation (the physical
system) to its mathematical formulation (its mathematical model)
• Sort of approximation can play a role here.
– Step 2. The solution by a mathematical method
– Step 3. The physical interpretation of the result

Physical Mathematical Physical

Model Set-up Manipulation Interpretation

Derivative Differential Equation

Multiple variables Partial Differential Equation

Computing Aide

15
Basic Concepts. Modeling
• Example 5. Radioactivity, Exponential Decay
– Step 1. Setting up a mathematical model (a differential equation)
of the physical process
dy
ky
dt

dy
Initial value problem given ky, y (0) 0.5
dt

– Step 2. Mathematical solution

Initial value
y(t) cekt y (0) c 0.5
y(t) 0.5ekt
dy
0.5kekt k·0.5ekt ky, y (0) 0.5e0 0.5
dt
16
Basic Concepts. Modeling
• Example 5. Radioactivity, Exponential Decay
– We can verify the solution using the original ODE and IC

– Step 3. Interpretation of result

17
Basic Concepts. Modeling
• Example 6: A geometric application
– Find the curve through (1,1) having the slope –y/x
– ODEs
general solution
y y/ x y c/x
y(1) c /1 1 c 1

– Particular solution: y 1/ x

18
1.2 Geometric Meaning of
y′=f (x, y) : Direction Fields

19
Geometric Meaning of y′=f (x, y) : Direction Fields
• A First-Order ODE
y  = f ( x, y )

– A solution curve passed through a point (x0, y0) (initial condition)

→ y( x0 ) = f ( x0 , y0 )

– Solution finding using direction fields

• Direction Fields by a CAS (Computer Algebra System)

• Direction Fields by using Isoclines (the Older method)

f ( x, y ) k contant isoclines

Good numerial approach for non-solvable DE

20
Geometric Meaning of y′=f (x, y) : Direction Fields
• Two Direction Fields and induced solution
x 2 /2
y ce (c arbitrary)  General solution (it is solvable)
y x 2 /2
x(ce ) xy c 2/ e 1.213
Check!
x2 /2
( x, y ) (1, 2) y (1) 2 2 1/2
ce y ( x) 1.213e

21
Geometric Meaning of y′=f (x, y) : Direction Fields
• Direction Fields by using Isoclines (the Older method)
x
y 0.1(1 x 2 )
y
k 5, 3, 1/ 4, 1

– Linear elements generated

is generated by Computer
– 3 typical sol. Curves.

• On Numerics
– Euler-Cauchy method
– Runge-Kutta method

22
Separable ODEs. Modeling

23
Separable ODEs. Modeling
• Many practically useful ODEs can be reduced to the form

g (y )y = f (x)

 g ( y) ydx =  f ( x)dx + c
ydx = dy   g ( y )dy =  f ( x)dx + c

• Method of separating variables

– X appears only on the right side
– Y appears only on the left side

The separation make things easy!

24
 Separable ODEs. Modeling
• Example 1. A separable ODE
y = 1 + y 2

Separation
dy dy
= 1+ y2  = dx
dx 1+ y 2

Integration

arctan( y ) = x + c
 y = tan( x + c)

25
Separable ODEs. Modeling
• Example 2. Radiocarbon dating
– The iceman, Oetzi, found in Southern Tvrolia in 1991.
The ratio if radioactive carbon 6C14 to ordinary carbon
6C
12 was 52.5% of that of a living organism

y: amount of 6C14
k: rate exponent
H: half life, e.g., 5000 years.

26
Separable ODEs. Modeling
• Example 2. Radiocarbon dating
– Radioactive decay is governed by the ODE y’=ky

dy
= kdt , ln y = kt + c, y = y0 e kt
y

– The half life H=5715 determines k

ln 0.5 0.693
y0ekH = 0.5 y0 , ekH = 0.5, k = =− = −0.0001213
H 5715

– Determine the time t

ekt = e−0.0001213 = 0.525

ln0.525
t= = 5312
−0.0001213
27
Separable ODEs. Modeling
• Example 3. Self regulated water level
– Initially, 1000 gal containing 100 lb of salt in the tank
– Influx: 10gal/min, 5 lb/gal salt.
– Outflux: 10 gal/min

10
y ' = 50 − y = 50 − 0.01y = 0.01( y − 5000)
1000

– Separating variables
dy
= −0.01 dt , ln y − 5000 = 0.01t + c1 , y = c1e −0.01t + 5000
y − 5000

– The initial condition

100 = ce0 + 5000, c = 4900

– Solution
y (t ) = 5000 − 4900e−0.01t
28
Separable ODEs. Modeling
• Example 4. Heating office building(Newton’s law of cooling)
• Daytime temp. is 70oF. Heating is shut off 10PM turned on 6AM
• Someday inside temp. is 65oF at 2am.
• Outside is 50oF@10PM down to 40oF@6AM.
• What is inside temp. @ 6AM?

– Step 1. Setting up model

T(t): the inside temperature, TA(t): the outside temperature
By Newton’s law dT
= k (T − TA )
dt
– Step 2. General solution
Golden rule: TA replace with the average (45°F) of the two known value
(40°F and 50°F)
dT
= k dt , ln T − 45 = kt + c *
T − 45
T (t ) = 45 + cekt (c = ec* )
29
Separable ODEs. Modeling
• Example 4 : Heating office building(Newton’s law of cooling)
– Step 3. Particular solution
choose 10 P.M. to be t=0, where t=4 is 2 A.M.
T(0)=70 and particular solution is Tp

T (0) = 45 + ce 0 = 70, c = 70 − 45 = 25, Tp (t ) = 45 + 25e kt

– Step 4. Determination of k
using T(4)=65

Tp (4) = 45 − 25e4 k = 65, e4 k = 0.8,

1
k = ln 0.8 = −0.056, Tp (t ) = 45 − 25e−0.056t
4

30
Separable ODEs. Modeling
• Example 4. Heating office building(Newton’s law of cooling)
– Step 5. Answer and interpretation

6 A.M. is t = 8.
Tp (8) = 45 − 25e−0.056*8 = 61[F ]

31
Example 5 (Leaking tank outflow… )
• Problem
– Initial height of the water: 2.25m
– Cylinder diameter: 2m
– Hole diameter: 1cm

• Torricelli’s Law
– Outflowing water has velocity of

v(t ) = 0.6 2 gh(t )

32
Example 5
• Solution
– Step 1. Setting up the model

V = Avt ( A = Area of hole, v=velocity of the water)

ΔV* = − BΔh (B = coross-sectional the area of tank)
Change of volume of the water in the tank.
− Bh = Avt
h A A
= − v = − 0.600 2 gh(t )
t B B
dh A
t → 0  = −26.56 h(t ) (26.56 = 0.600 2*980)
dt B

– Step 2. General solution

Separation of var.
dh A A
= −26.56 dt and 2 h = c * −26.56 t
h B B
A 2 0.52  2
(By h = (c − 13.28 t ) = (c − 13.28 t ) = (c − 0.000332t ) 2

B 1002 
33
Example 5

– Step 3. Particular solution

c 2 = 225, c = 15.00, h = 225 at t = 0

hp (t ) = (15.00 − 0.000332t ) 2

– Step 4. Tank empty

hp (t ) = 0 if t = 15.00/0.000332 = 45181[sec] = 12.6[hour]

34
Reduction to a Separable Form
• Nonseparable form → Separable form
– By transformations

• A type of ODE
– Turn separable by transformation
y ' = f ( xy )
– Transformation: y/x=u
y = ux → y ' = u ' x + u
– By substitution
u ' x + u = f (u ) or u ' x = f (u ) − u

– Separated
du dx
=
f (u ) − u x

35
Separable ODEs. Modeling
• Example 6. Reduction to separable form
2 xyy ' = y 2 − x 2
– Dividing by x2
y 2 − x2
y' =
2 xy
– By setting u=y/x
u 1
u'x+u = −
2 2u
– By separating variables
2u du dx
= −
1+ u2 x
– By integrating
ln (1 + u 2 ) = − ln x + c1 ,
c
1+ u2 =
x
2
 c c2
x 2 + y 2 = cx, x−  + y =
2

 2 4 36
 The Natural Response of an RL Circuits
- Techniques for analyzing the natural response of first-order circuits
abruptly disconnected from its dc source

1. Assume that the independent current source generates a constant current of Is, and that
the switch has been in a closed position for a long time.
2. All currents and voltages have reached a constant value.
3. Before the switch is being opened ; the inductor appears as a short circuit (v=Ldi/dt = 0)
prior to the release of the stored energy
4. No current in either R0 or R, and all Is in the inductive branch
5. After the switch has been opened ; v(t) and i(t) for t ≥0 ?.

An RL Circuit

37
38
Deriving the Expression for the current

KVL

39
i (t ) = I 0 e − ( R / L )t , t 0
Voltage across the resistor,
v = iR = I 0 Re −(R/L)t , t  0+
v(0 − ) = 0,
v(0 + ) = I 0 R
Power dissipated in the resistor,
v2
p = vi, p = i R, or p =2
R
p = I 02 Re −2 ( R / L )t , t  0 +
Energy delievered to the resistor after the switch has been opened.,
t t
w =  pdx =  I 02 Re −2 ( R / L ) x dx
0 0
1
= I 02 R(1 − e −2 ( R / L )t )
2( R / L )
1
= LI 02 (1 − e −2 ( R / L )t ), t  0
2

As t becomes infinite,
the energy dissipated in the resistor approaches the initial energy stored in the inductor
40
The significance of the time
constant
i (t ) = I 0 e − ( R / L )t , t 0
v(t ) = I 0 Re −(R/L)t , t  0+
p = I 02 Re −2 ( R / L )t , t  0+
1 2
w= LI 0 (1 − e − 2 ( R / L )t ), t  0
2
R / L determines the rate at which t he current or voltage approches "0"
L
 = time constant =
R
t
−
i (t ) = I 0 e  , t 0
t
−
v(t ) = I 0 Re  , t  0+
t
−2
p= I 02 Re  , t  0+
t
1 2 −2
w = LI 0 (1 − e  ), t 0
2
41
 When the elapsed time exceeds five time constants(5),
the current is less than 1% of its initial value. Thus, we
say that a circuit system reaches its final value after
elapsing for ≥ 5.
long time : elapsed for ≥ 5

t
 5

Steady-state response: the response that exists a long

time after the switching has taken place
Transient response: the existence of current in a RL circuit
with a momentary event

42
1.4 Exact ODEs. Integrating
Factors

43
Exact ODEs. Integrating Factors
• Exact differential equation
– If a function u(x,y) has continuous partial derivatives, its differential is
u u
du = dx + dy
x y
• u(x,y) =c=constant → du=0.
– Example: u=x+x2y3=c
du = (1 + 2 xy 3 )dx + 3 x 2 y 2 dy = 0
dy 1 + 2 xy 3
y = =−
dx 3x 2 y 2

– Exact differential equation

• If the following holds
u u
M ( x, y )dx + N ( x, y )dy = 0 = du = dx + dy
x y
u u
M ( x, y ) = , N ( x, y ) =
x y
44
Exact ODEs. Integrating Factors
• Solution of Exact differential equation
– u(x,y)=c → implicit solution (while Explicit Solution y=h(x))

• Condition for exact diff. equation

– For a function u,
M  2u 
u u =
= M, =N y yx  M N
x y   =
N  u 
2
y x
= Necessary and
x xy  sufficient cond.

u
M ( x, y ) =  u ( x, y ) =  Mdx + k ( y ) = 0
x Constant

u
N ( x, y ) =  u ( x, y ) =  Ndy + l ( x) = 0
y
45
Exact ODEs. Integrating Factors
• Example 1. Solve
cos( x + y )dx + (3 y 2 + 2 y + cos( x + y ))dy = 0

• Solution
– Step 1. Test for exactness.

M = cos( x + y )
N = 3 y 2 + 2 y + cos( x + y )
M 
= − sin( x + y ) 
y  The same!
N 
= − sin( x + y ) 
x 

46
Exact ODEs. Integrating Factors
• Example 1. An Exact ODE
– Step 2. Implicit general solution

u =  Mdx + k ( y ) =  cos( x + y )dx + k ( y ) = sin( x + y ) + k ( y )

u dk
= cos( x + y ) + = N = 3 y 2 + 2 y + cos( x + y ).
y dy
dk
 = 3 y 2 + 2 y  k = y3 + y 2 + c *
dy
 u ( x, y ) = sin( x, y ) + y 3 + y 2 = c

– Step 3. Checking an implicit solution

u u
du = dx + dy = cos( x + y )dx + (cos( x + y ) + 3 y 2 + 2 y )dy = 0
x y
47
Exact ODEs. Integrating Factors
• Example 2. An Initial value Problem
(cos y sinh x + 1)dx − sin y cosh xdy = 0, y (1) = 2

– Solution
u = −  sin y cosh xdy + l ( x) = cos y cosh x + l ( x)
u dl
= cos y sinh x + = M = cos y sinh x + 1
x dx
dl
 =1
dx
– By integral
l ( x) = x + c *
u( x, y) = cos y cosh x + x = c
– By initial condition

cos 2 cosh1 + 1 = 0.358 = c

 ( x, y ) = cos y cosh x + x = 0.358
48
Exact ODEs. Integrating Factors
• Example 3. Breakdown in the case of nonexact DE

– Solution
-ydx + xdy = 0 is not exact because M = −y and N = x, so that
M N
= −1 but = 1.
y x
u k
u =  M dx + k ( y ) = − xy + k ( y ), hence = −x + .
y y

u
Now, should equal N=x, however k(y) can depend on only on y.
y

49
Integrating Factors
• Reduction to exact form: Integrating Factors
– ODE : − ydx + xdy = 0 → Multifly it by 1/ x 2

− ydx + xdy y 1 y
2
= − 2 dx + dy = d ( ) = 0
x x x x
1 1
→M =− 2, N =
x x
M 1 N 1
→ =− 2, = − 2  Exact!!
y x x x

• Integrating Factor: F
FP( x, y )dx + FQ( x, y )dy = 0,
where P( x, y )dx + Q( x, y )dy = 0 : a nonexact ODE
FP( x, y)dx + FQ( x, y) dy = 0 : an exact ODE

50
Integrating Factors
• Finding Integrating Factors
– The exact condition
 
( FP) = ( FQ)
y x
 Fy P + FPy = Fx Q + FQx

• Too complicated → ?

– Golden rule
• Assume F is a function of only variable x.
dF
Let F = F ( x). Then Fy = 0, and Fx = F  = .
dx
 FPy = F Q + FQx

1 dF 1 P Q
= R, where R = ( − ).
F dx Q y x
51
Exact ODEs. Integrating Factors
• Theorem 1
– Integrating Factor F(x)
• If (12) is such that the right side R of (16), depends only on x, then (12)
has an IF F=F(x), which is obtained by integrating (16) and taking
exponents on both sides.

(12) P( x, y )dx Q( x, y )dy 0

F ( x) = exp  R( x)dx.
1 dF 1 P Q
(16)
F dx Q y x

1 P Q
※ Note R= ( − ).
Q y x

52
Exact ODEs. Integrating Factors

• Theorem 2
– Integrating Factor F*(y)
• If (12) is such that the right side R* of (18) depends only on y, then
(12) has an integrating factor F*=F*(y), which is obtained from (18) in
the form.

F *( y ) = exp  R( y )dy.

(12) P( x, y )dx Q( x, y )dy 0

1 dF 1 Q P
(18) R*
F dy P x y

53
Example 5 (IF with Initial Value Condition)
• Application of Theorems 1 and 2.
– Using theorem 1 and 2 find an integrating factor and solve the
initial value problem.

(e x y
ye y )dx (xe y 1)dy 0, y (0) 1

• Solution
– Step 1. Nonexactness. The exactness check fail:

P  x + y x+ y Q 
= (e + ye ) = e + e + ye but,
y y y
= ( xe y − 1) = e y
y y x x

54
Example 5 (IF with Initial Value Condition)
– Step 2. Integrating Factor. General solution
• Try two theorems
Too complicated!
1 P Q 1
R= ( − )= y (e x + y + e y + ye y − e y )
Q y x xe − 1
Try for F(y)
1 Q P 1 x+ y
R* = ( − ) = x+ y ( e y
− e − e y
− ye y
) = −1
P x y e + ye y
• By using THEOREM 2, integrating factor is

F * ( y) = e− y
• Test for exactness by integration
(e x + y )dx + ( x − e − y )dy = 0

u =  (e x + y )dx = e x + xy + k ( y )
u dk dk
 = x+ = N = x − e− y , = −e − y , k = e − y + c *
y dy dy
 u ( x, y ) = e x + xy + e − y = c
55
Example 5
– Step 3. Particular solution by initial value

u ( x, y ) = e x + xy + e − y = c y (0) = −1
u (0, −1) = 1 + 0 + e = 3.72
 u ( x, y ) = e x + xy + e − y = 3.72

– Step 4. Checking

Particular solution

56
1.5 Linear ODEs. Bernoulli
Equation

57
First-Order Linear ODE
• Standard Linear Form
y: output r: input
– Linear in y and y’
y + p ( x) y = r ( x)
– Example
• A linear DE: y 'cos x + y sin x = x → y '+ y tan x = x sec x
Standard form
• Non-linear form y '+ p( x) y = r ( x) y 2

• Homogeneous Linear ODE

– r(x) is zero → Homogeneous Linear ODE
dy
y + p( x) y = 0 = − p( x)dx
y
– Thus,
ln y = −  p ( x)dx + c *
y ( x) = ce 
− p ( x )dx

– If c = 0 then y(x) = 0  Trivial solution

58
Non-homogeneous Linear ODE
• Conditions
– r(x) is not always zero, move r to the left.
( py − r )dx + dy = 0 → Pdx + Qdy = 0, where P = py − r and Q = 1
– Integrating factor F(x)
1 dF dF
= p( x) → = pdx and ln F =  pdx
F dx F
 F ( x) = e 
pdx

– Multiply F(x)
e ( y + py ) = (e  y ) = e  r → e y =  e
pdx pdx pdx pdx pdx
rdx + c

 y ( x) = e− h (  eh rdx + c), h =  p( x)dx

– Total Output = Response to the Input r + Response to the Initial data

y ( x) = e − h  e h rdx + ce − h

59
Example 1
• First-order ODE, general solution
y − y = e 2 x

– General solution

p = −1, r = e2 x , h =  pdx = − x

y ( x) = e x (  e− x e 2 x dx + c) = e x (e x + c) = ce x + e 2 x

– Directly, multiplying the given equation by e h = e − x

( y − y)e− x = ( ye− x ) = e2 x e− x = e x
ye− x = e x + c, hence y = e2 x + ce x

60
Example 2. First-order ODE, Initial Value Problem
• Problem y + y tan x = sin 2 x, y (0) = 1

Note p = tan x, r = sin 2 x = 2sin x cos x,

h =  pdx =  tan xdx = ln sec x
e h = sec x, e − h = cos x, e h r = (sec x)(2sin x cos x) = 2sin x
y ( x) = cos x(2  sin xdx + c) = c cos x − 2 cos 2 x

– From initial condition,

1 = c 1 − 2 12 , c=3
– Hence
 y = 3cos x − 2 cos 2 x

61
Example 3. Hormone Level (optional)
– Assume that level of a certain hormone in the blood varies with
time. Time rate of change is the difference between a sin. input of
a 24-hour period from the thyroid gland and a continuous removal
rate proportional to the level present. Set up a model for the
hormone level in the blood and find its general solution. Find the
particular solution satisfying a suitable initial condition

– Step 1. Setting up a model

• Let y(t) be the hormone level at time t, The removal rate is ky(t).
• The input rate is A+Bcos(2πt/24) , where A is the average input rate (A
≥ B to make the input nonnegative.)
t t
y(t ) = In − Out = A + B cos( ) − ky (t ) or y + ky = A + B cos( )
12 12
• The initial condition
y part (0) = y0 with t = 0

62
Example 3. Hormone Level (optional)
– Step 2. General solution
1
p = K = const , h = kt , r = A + B cos( t )
12
t
y (t ) = e − Kt  e Kt ( A + B cos )dt + ce − Kt
12
A B t t
= e − Kt e Kt [ + (144 K cos + 12 sin )] + ce − Kt

K 144 K 2 +  2 12 12
A B t t
= + (144 K cos + 12 sin ) + ce − Kt

K 144 K 2 +  2 12 12

– Step 3. Particular solution

A B A B
y (0) = +  144 K + c = 0 Thus c = − − 144 K
K 144 K 2 +  2 K 144 K 2 +  2
A B t t A 144 KB
y part (t ) = +  144 K (cos + 12 sin ) − ( + ) e − Kt

K 144 K 2 +  2 12 12 K 144 K 2 +  2

63
Reduction to Linear Form, Bernoulli Equation
• Bernoulli Equation
– Nonlinear → Linear form

y + p ( x) y = g ( x) y a (a any real number)

– If a = 0 or a = 1, this equation is linear. Otherwise it is nonlinear.
– We set
u ( x) = [ y ( x)]
1-a

– Differentiate

u = (1 − a) y − a y = (1 − a) y − a ( gy a − py )
u = (1 − a)( g a − py1− a )

→ u + (1 − a) pu = (1 − a ) g since y1− a = u
: Linear ODE for u

64
Example 4. Logistic Equation
• Solve the following Bernoulli eq.
y = Ay − By 2
• Solution y=0 is also a solution

y − Ay = − By 2
a = 2 → u = y1− a = y −1

– Diff. and substitution

u = − y −2 y = − y −2 ( Ay − By 2 ) = B − Ay −1
→ u + Au = B since − Ay −1 = u
– General solution
u = ce− At + B / A
1 1
Since y = u −1 , y= = − At
u ce + B / A

65
Population Dynamics (Optional)

– Malthus’s law : y = Ay − By 2

– If B = 0, y = dy = Ay ⇒ y = 1 e At (exponential growth)
dt c
B
y = Ay − By 2 = Ay (1 − y)
A

– If y  A , then y  0 , so that an initially small population keeps

B
growing as long as y  A .
B
– If y  A , then y  0 and population is decreasing as long as y  A
B B
– The limit is the same in both cases, namely, A .
B

66
Example 5
• Stable and unstable Equilibrium Solutions. “Phase Line Plot”

The ODE y = ( y − 1)( y − 2) has the stable equilibriu m solution y1 = 1and the unstable y2 = 2,
as the direction field in Fig.19 suggests. The values y1 and y2 are the zeros of the parabola
f ( y ) = ( y − 1)( y − 2) in the figure. Now, since the ODE is autonomous, we can " condense" the
direction field to a " phase line plot" giving y1 and y2 , and the direction( upward or downward)
of the arrows in the filed, and thus giving informatio n about the stability or instabilit y of the
equilibriu m solutions.

67
 the sudden application
vs = Vm cos(t +  )
di
I0 = 0 (initial current in inductor) vL = L
dt

: It becomes infinitesimal
as time elapses

: It exists
as long as the switch remains

68
Existence and Uniqueness of
Solutions

69
Existence and Uniqueness of Solutions
• Initial value problem

– y + y = 0, y (0) = 1 has no solution!

– y = 2 x, y (0) = 1 has precisely one solution!

 y = x2 +1

– xy = y − 1, y (0) = 1 has infinitely many solution!

 y = 1 + cx

70
Existence and Uniqueness of Solutions
• Initial value problem
y = f ( x, y ), y ( x0 ) = y0 One, No, or Many
solutions can exist.
• Theorem 1. Existence Theorem
– Let the right side f(x,y) of the ODE in the initial value problem
(1) y' f ( x, y ), y ( x0 ) y0

– be continuous at all points (x,y) in some rectangle

R :| x x0 | a, |y y0 | b (Fig. 26)
– and bounded in R; that is, there is a number K such that

(2) | f ( x, y) | K for all ( x, y) in R

→ Then the initial value problem (1) has at least one solution y(x). This
solution exists at least for all x in the subinterval |x − x0| < α of the
interval |x − x0|<a; here, α is the smaller of the two numbers a and b/K.

71
Existence and Uniqueness of Solutions
• Theorem 2. Uniqueness Theorem
– Let f and its partial derivative f y f / y be continuous for all (x,y)
in the rectangle R (Fig. 26) and bounded, say,

(3) (a) | f ( x, y) | K , (b) | f y ( x, y) | M for all ( x, y) in R

→ Then the initial value problem (1) has at most one solution y(x).
Thus, by Theorem 1, the problem has precisely one solution. This
solution exists at least for all x in that subinterval |x − x0| < α.

Fig 26. Rectangular R of theorem over existence and uniqueness

72
Example 1. Choice of a Rectangle
• Consider the initial value problem
y' 1 y 2 , y (0) 0

– Take the rectangle R; |x|<5, |y|<3, Then a=5, b=3 and

f ( x, y ) 1 y2 K 10,
f
2y M 6,
y
b
0.3 a.
K

73
Existence and Uniqueness of Solutions
• Lipschitz condition
– From mean value theorem, we have
f
f ( x, y2 ) f ( x, y1 ) ( y2 y1 )
y y y

– where (x,y1) and (x,y2) are assumed to be in R, and y is a suitable

value between y1 and y2.

– Strong conditions are transformed to a weaker one.

– Lipschitz condition

f ( x, y2 ) f ( x, y1 ) M y2 y1

• Weaker condition than | f y ( x, y) | M

74
Example 2. Nonuniqueness
• Consider the initial value problem
– The initial value problem Continuity of f(x,y)
doesn’t guarantee
y y , y(0) 0 the uniqueness.

– The two solutions

x2
if x 0
4
y 0 and y*
x2
if x 0
4

– Although f ( x, y) y is continuous for all y.

– For y1=0 and positive y2

f ( x, y2 ) f ( x, y1 ) y2 1
, ( y2 0)
y2 y1 y2 y2
This can be very large
By the choice of y2.
Violates the Lib.. Cond.
75
Continued
Page 44a

Continued
Continued