34 de nition 568 a lmm is weakly stable if it is

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

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: rog scheme grew when h < 0 and decayed when h > 0. All parasitic solutions of a strongly stable method decay for su ciently small values of h as n increases, since there modulus is less than unity when h = 0. Example 5.6.4. Let us examine the stability of the Adams-Bashforth and AdamsMoulton methods as h ! 0. Recall (Sections 5.3 and 5.4) that the Adams' methods have the form (5.1.2) with 0 = 1, 1 = ;1, and i = 0, i = 2 3 : : : k, i.e., yn = yn;1 + h k X i=0 i fn;i : The coe cient 0 = 0 for the (explicit) Adams-Bashforth methods and is nonzero for the (implicit) Adams-Moulton methods. Using the test equation (5.6.4) yn = yn;1 + h Assuming that yn is proportional to n k X i=0 i yn;i: leads to (5.6.7) which, in this case, is ( );h ( )=0 ( ) = k;1( ; 1) ( )= k X i=0 i k ;i : Thus, the roots of the rst characteristic polynomial ( ( ) = 0) are 1 =1 2 = 3 35 = ::: = k = 0: The parasitic roots i(0), i = 2 3 : : : k, are all zero for both the Adams-Bashforth and Adams-Moulton methods thus, these methods are strongly stable. Computation is performed at nite values of h. The root condition and the notions of strong and weak stability all involve reasoning in the limit of vanishingly small step sizes. This shortcoming encourages us to further extend concepts of stability. De nition 5.6.10. A LMM is absolutely stable for those values of h where the roots of (5.6.7a) satisfy j (h )j 1 when applied to the test problem (5.6.4). De nition 5.6.11. A LMM is relatively stable for those values of h where the parasitic roots of (5.6.7) are less in magnitude than the principal root. The boundary of the region of absolute stability of a LMM may be calculated by the boundary locus method as described in Chapter 3 for one-step methods. Complex values of having unit modulus are written in polar form = ei and (5.6.7a) is imposed. The boundary of the region of absolute stability follows by calculating i 2 0 ]: h = (e i ) (e ) The regions of absolute stability of the Adams-Bashforth, Adams-Moulton, and backward di ere...
View Full Document

This document was uploaded on 03/16/2014 for the course CSCI 6820 at Rensselaer Polytechnic Institute.

Ask a homework question - tutors are online