Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

Numerical Methods 151B

Summer 2020

UCLA

August 3, 2020
Initial Value Problem

The problem is to find a function y (t) such that


⇢ 0
y (t) = f (t, y (t)) atb
(1)
y (a) = ↵ initial condition

for the ODE


y 0 (t) = f (t, y (t))
Initial Value Problem: motivation

I Whenever we do calculation using a computer, we introduce a


round-o↵ error
I Example: In Matlab, using double precision value, the first
number after 1 is roughly 1 + 10 6
I We need to make sure that our approximated solution will not
be a↵ected by a small error
I We require our IVP problem to be well-posed in order to
ensure this.
Well-Posedness

The IVP ⇢
y 0 (t) = f (t, y (t)) atb
(2)
y (a) = ↵ initial condition
is said to be well-posed if the following conditions are satisfied:
(i) The IVP has a unique solution.
(ii) There exists a k > 0 and ✏0 > 0 such 8✏ 2 (0, ✏0 ) with
continuous (t), where (t)  ✏ for all a  t  b, 0  ✏ , the
perturbed problem:
⇢ 0
z (t) = f (t, z(t)) + (t) atb
(3)
z(a) = ↵ + 0 initial condition

has a unique solution z(t) satisfying

|z(t) y (t)|  k✏8t 2 [a, b] (4)


Well-posedness in words...

The IVP for z(t) is a perturbed problem.


Due to round o↵ error we almost always solve a perturbed
problem!!
If the initial IVP is well-posed, the solution to the perturbed
problem will be close to the solution of the non-perturbed problem.
Or back to the equations...

There exists a k > 0 and ✏0 > 0 such 8✏ 2 (0, ✏0 ) with (t) as a


continuous with (t)  ✏ for all a  t  b, 0  ✏ , the perturbed
problem:
⇢ 0
z (t) = f (t, z(t)) + (t) atb
(5)
z(a) = ↵ + 0 initial condition

has a unique solution z(t) satisfying

|z(t) y (t)|  k✏8t 2 [a, b] (6)

You might also like