# Di erence approximations may be constructed in a

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: mations may be constructed in a variety of ways, but the use of Taylor's formula is probably the simplest for our present purposes. To begin, let us use Taylor's formula to express n+1 in terms of n and its derivatives as j j 2 1 ( )n 2 + + 1 ( k )n k n n n )j + 2! 2 j j +1 = j + ( ! kj k +1 + 1 ( k+1 )n+ k+1 (2.1.2) j ( + 1)! u u x x U t u t u u u @u @x x u @u @x x ::: @ k 1 @x @u k u x @x x : 2 Finite Di erence Methods t n (j,n) 2 1 -2 -1 0 1 2 j x Figure 2.1.1: A partition of the upper half of the ( )-plane into uniform cells of size . xt x t The last term in (2.1.2), the remainder, involves the evaluation of k+1 k+1 at = ( + ) and , with an unknown point on (0 1). As an example, suppose that we retain the rst two terms ( = 1) in (2.1.2) and solve for ( )n to obtain j @u j x n [email protected] x t k @[email protected] ( )= n ux j +1 ; uj n uj n ; 1 ( xx)n+ j 2 u x (2.1.3) x: Neglecting the remainder term, we get the formula for the rst forward nite di erence approximation of x as u +1 ; Uj ( )= n Uj n Ux j n x (2.1.4a) : The neglected remainder term in (2.1.3) n j n ; ( x)n ; j+1 j u u x n uj = ;1( 2...
View Full Document

{[ snackBarMessage ]}

Ask a homework question - tutors are online