The Elementary Theory of Initial–Value Problems - (7.

1)

In Chapter 7, we study the numerical methods for solving initial-value problems for ordinary differential
equations:
y ′ t  f t, y , a ≤ t ≤ b, ya  .

1. Lipschitz Condition:
Definition 1 A function ft, y is said to satisfy a Lipschitz condition in the variable y on a set D ⊂ R 2 if
there exists a constant L  0 such that
ft, y 1  − ft, y 2  ≤ L y 1 − y 2 ,
whenever both points t, y 1 and t, y 2 are in D. The constant L is called a Lipschitz
constant for f.

Example Let ft, y  1  ty 2 and D  t, y | 0 ≤ t ≤ 2, − 1 ≤ y ≤ 1 . Does f satisfy a Lipschitz

condition on D? If so, find its Lipschitz constant.
Let t, y 1 and t, y 2 be in D, i.e., t is in 0, 2 and y is in −1, 1 . Observe that
ft, y 1  − ft, y 2   |1  ty 21 − 1  ty 22 |  t |y 21 − y 22 |  t |y 1  y 2 y 1 − y 2 |  t y1  y2 y1 − y2 .
Because t ≤ 2 and y 1  y 2 ≤ 2, we have
ft, y 1  − ft, y 2  ≤ 22|y 1 − y 2 |  4 y 1 − y 2 .
So, f satisfies a Lipschitz condition and its Lipschitz constant is 4.
Note that the Lipschitz constant L is not unique, that is, for any L  4, the inequality
ft, y 1  − ft, y 2  ≤ L y 1 − y 2
also holds. So, in practice, we want to find a L as small as possible.

2. A Sufficient Condition for Lipschitz Condition:

Theorem 1 Suppose ft, y is defined on a convex set D ⊂ R 2 . If a constant L  0 exists with
∂f
t, y ≤ L, for all t, y ∈ D,
∂y
then f satisfies a Lipschitz condition on D in the variable y with Lipschitz constant L.

Example Let ft, y  1  ty 2 and D  t, y | 0 ≤ t ≤ 2, − 1 ≤ y ≤ 1 . Does f satisfy a Lipschitz

condition on D? If so, find its Lipschitz constant.
∂f
The partial derivative of f with respect to y is t, y  2ty. Because
∂y
∂f
t, y  2ty  2 t y ≤ 221  4,
∂y
f satisfies a Lipschitz condition with Lipschitz constant 4.

Example Let ft, y  1  t sinty. Determine if f satisfies a Lipschitz condition in

D t, y | 0 ≤ t ≤ 2, −   y   .

1
 ∂f
Compute the partial derivative of f with respect to y : t, y  t costy t  t 2 costy. Because
∂y
costy ≤ 1 and t ≤ 2,
∂f
t, y  t 2 costy  |t 2 | |costy| ≤ |t 2 | ≤ 4.
∂y
So, f satisfies a Lipschitz condition with a constant 4.

3. A Sufficient Condition for the Uniqueness of Solution of an Initial Value Problem:

Theorem 2 Suppose that D  t, y | a ≤ t ≤ b, −   y   and that f t, y is continuous on
D. If f satisfies a Lipschitz condition on D in the variable y, then the initial-value problem
y ′ t  f t, y , a ≤ t ≤ b, ya  
has a unique solution yt for a ≤ t ≤ b.

Example Determine if the initial - value problem

y ′ t  2 y  t 2 e t , 1 ≤ t ≤ 2, y1  0
t
has a unique solution for 1 ≤ t ≤ 2. If so, find the solution exactly or numerically.
Let D  t, y | 1 ≤ t ≤ 2, −   y   . Check if f t, y  2 y  t 2 e t satisfies a Lipschitz
t
condition in D. Because
∂f
 2 ≤ 2  2,
∂y t 1
f t, y satisfies a Lipschitz condition in D and therefore by Theorem 2 the initial-value problem has a
unique solution.
dy
Solve  2 y  t2et:
dt t
dy
(i) Solve the homogeneous equation − 2 y  0 by separation of variables:
dt t
1 dy  2 dt, ln|y|  2 ln|t|  C, e ln|y|  e ln|t| 2 C , y h  t 2 e C  t 2 C.
y t
dy
(ii) Find a particular solution of the homogeneous equation − 2 y  t2et :
dt t
Let y p  At  Bt  Ce . Then
2 t

y ′p  2At  Be t  At 2  Bt  Ce t

and
2At  Be t  At 2  Bt  Ce t − 2 At 2  Bt  Ce t  t 2 e t 
t
At 2 e t  2A  B − 2Ate t  B  C − 2Be t − 2 Ce t  t 2 e t
t
A  1, B  0, C − B  0, C  0
The solution is y p  t 2 e t and the general solution is: yt  y h  y p  t 2 C  t 2 e t .
(iii) Solve C by the initial value y1  0 : y1  C  e  0, C  −e.
The general solution: yt  t 2 e t − et 2 .

Example Determine if the initial - value problem

2
 y ′ t  1  t sinyt, 0 ≤ t ≤ 2, y0  0
has a unique solution for 0 ≤ t ≤ 2. If so, find the solution exactly or numerically.
Let D  t, y | 0 ≤ t ≤ 2, −   y   . Clearly, f t, y  1  t sinyt is continuous on D.
From an earlier example, we know f t, y  1  t sinyt satisfies a Lipschitz condition on D in the
variable y. By Theorem 2, we know the initial value problem has a unique solution.
dy
Solve the initial value problem:  1  t sinyt, y0  0. We cannot solve it exactly. The
dt
numerical solution:

1.5

0.5

0
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

y ′  1  sinyt, 0 ≤ t ≤ 2, y0  0

4. Picard’s Method:
Picard’s method is a method approximate the solution yt of the initial-value problem:
y ′  f t, y , a ≤ t ≤ b, ya  
by a sequence of functions y k t where y k t are functions:
t
y k t     f x, y k−1 x dx, k  1, 2, . . . .
a
That is, y 0 t  ,
t t
y 1 t     f x,  dx, y 2 t     f x, y 1 x dx, . . . .
a a
Derivation: For k ≥ 1,

t t
 a fx, yxdx   a y ′ xdx  yx| ta  yt − ya  yt − 
implies
t
y k t     f x, y k−1 x dx.
a

Example Find approximations y k t of the solution to the initial-value problem:

dy
 y  t, y0  −1, by Picard’s method.
dt
For this problem, ft, y  y  t and y 0 t  −1. Then
t
y 1 t  −1   −1  xdx  1 t 2 − t − 1
0 2

3
 t
y 2 t  −1   1 x 2 − x − 1  x dx  −1   t 1 x 2 − 1 dx  1 t 3 − t − 1
0 2 0 2 6
y k t  1 t k1 − t − 1.
k  1!

x
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
-1.0
y
-1.2

-1.4

-1.6

-1.8

-2.0

-2.2

-2.4

-2.6

-2.8

.-.- y 4 t, – y 5 t, - - y 6 t

Example Find approximations y 1 t and y 2 t of the solution to the initial-value problem:
dy
 1  t sinyt, y0  0, by Picard’s method.
dt
For this problem, ft, y  1  t sinyt and y 0 t  0. Then
t t
y 1 t  0   1  x sinx0dx   0 dx  t.
0
t
y 2 t  0   1  x sinx 2 dx  t  t 2 sin t 2 .
0
t
y 3 t  0   1  x sinxx  x 2 sinx 2 dx  ?
0

Example Find approximations y 3 t of the solution to the initial-value problem:

dy
 y 2  t, y0  1, by Picard’s method.
dt

t
y 0 t  1, y 1 t  1   1  xdx  t  1 t 2  1
0 2
t
y 2 t  1  
2
x  1 x2  1  x dx  t  3 t 2  2 t 3  1 t 4  1 t 5  1
0 2 2 3 4 20
t
y 3 t  1  
2
x  3 x 2  2 x 3  1 x 4  1 x 5  1  x dx
0 2 3 4 20
 t  3 t 2  4 t 3  13 t 4  49 t 5  13 t 6  233 t 7  29 t 8  31 t 9  1 t 10  1 t 11  1
2 3 12 60 30 1260 480 2160 400 4400

4
 14
y

12

10

0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
x

black - y 0 , blue - y 1 , red - y 2 , green - y 3

5. Approximate the solution by Taylor series:

Example Determine the first 5 terms in Taylor series expansion of the solution to the initial-value
dy
problem:  y  t, y0  −1.
dt
Let yt  y0  y ′ 0 t  12 y ′′ 0 t 2 . . .  n!1 y n t t n . . . . and y ′ t  ft, y  y  t.
y0  −1, y ′ 0  y0  0  −1
y ′′ t  y ′ t  1, y ′′ 0  y ′ 0  1  −1  1  0
y ′′′ t  y ′′ t, y ′′′ 0  0
yt ≈ −1 − t

Example Determine the first 5 terms in Taylor series expansion of the solution to the initial-value
problem: y ′  t 2 − 2y 2 − 1, y0  0.
Let yt  y0  y ′ 0 t  12 y ′′ 0 t 2 . . .  n!1 y n t t n . . . . and ft, y  t 2 − 2y 2 − 1.
y0  0, y ′ 0  −1
y ′′ t  2t − 22yy ′   2t − 4yy ′ , y ′′ 0  0 − 49  0
y ′′′ t  2 − 4y ′  2 − 4yy ′′  2 − 4−1 2  −2
y 4 t  −8y ′ y ′′ − 4y ′ y ′′  yy ′′′ , y 4 0  0 − 40  0  0
yt ≈ −t − 2t 3