17.02.2014 Views

Mathematical Methods for Physicists: A concise introduction - Site Map

Mathematical Methods for Physicists: A concise introduction - Site Map

Mathematical Methods for Physicists: A concise introduction - Site Map

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

NUMERICAL METHODS<br />

approximate value obtained <strong>for</strong> y 1 , we obtain an improved value, denoted by<br />

…y 1 † 1<br />

:<br />

…y 1 † 1 ˆ y 0 ‡ 1 2 f f …x 0; y 0 †‡f …x 0 ‡ h; y 1 †g:<br />

…13:21†<br />

This process can be repeated until there is agreement to a required degree of<br />

accuracy between successive approximations.<br />

The three-term Taylor series method<br />

The rationale <strong>for</strong> this method lies in the three-term Taylor expansion. Let y be the<br />

solution of the ®rst-order ordinary equation (13.18) <strong>for</strong> the initial condition<br />

y ˆ y 0 when x ˆ x 0 and suppose that it can be expanded as a Taylor series in<br />

the neighborhood of x 0 .Ify ˆ y 1 when x ˆ x 0 ‡ h, then, <strong>for</strong> suciently small<br />

values of h, wehave<br />

<br />

y 1 ˆ y 0 ‡ h<br />

dy ! !<br />

‡ h2 d 2 y<br />

dx<br />

0<br />

2! dx 2 ‡ h3 d 3 y<br />

3! dx 3 ‡: …13:22†<br />

Now<br />

and<br />

d 3 y<br />

dx 3 ˆ<br />

0<br />

dy ˆ f …x; y†;<br />

dx<br />

d 2 y<br />

dx 2 ˆ @f<br />

@x ‡ dy @f<br />

dx @y ˆ @f<br />

@x ‡ f @f<br />

@y ;<br />

@<br />

@x ‡ f @ @f<br />

@y<br />

@x ‡ f @f<br />

@y<br />

<br />

ˆ @2 f<br />

@x 2 ‡ @f @f<br />

@y @y ‡ 2f @ 2 f<br />

@x@y ‡ f<br />

<br />

@f 2<br />

‡ f 2 @ 2 f<br />

@y @y 2 :<br />

Equation (13.22) can be rewritten as<br />

<br />

y 1 ˆ y 0 ‡ hf …x 0 ; y 0 †‡ h2 @f …x 0 ; y 0 †<br />

‡ f …x<br />

2 @x<br />

0 ; y 0 † @f …x <br />

0; y 0 †<br />

;<br />

@y<br />

where we have dropped the h 3 term. We now use this equation as an iterative<br />

equation:<br />

<br />

y n‡1 ˆ y n ‡ hf …x n ; y n †‡ h2 @f …x n ; y n †<br />

‡ f …x<br />

2 @x<br />

n ; y n † @f …x <br />

n; y n †<br />

: …13:23†<br />

@y<br />

That is, we compute y 1 ˆ y…x 0 ‡ h† from y 0 , y 2 ˆ y…x 1 ‡ h† from y 1 by replacing x<br />

by x 1 , and so on. The error in this method is proportional h 3 . A good approxima-<br />

472<br />

0

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!