Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

BLOCK ITERATIVE METHODS FOR TWO-CYCLIC MATRIX EQUATIONS WITH SPECIAL APPLICATION TO THE NUMERICAL SOLUTION OF THE SECOND-ORDER SELF-ADJOINT ELLIPTIC PARTIAL DIFFERENTIAL EQUATION IN TWO DIMENSIIONS

Technical Report ·
OSTI ID:4812735

An investigation was made to determine whether a cyclically reduced system of equations or the original system may be solved by an iterative method in the most efficient manner. Sohroder and Varga used the special form of a 2- cyclic matrix to obtain improvements in the convergence rate for certain point iteration techniques. It is shown that improvements in the rate of convergence may also be obtained for various block iteration techniques and that the cyclic reduction process leads to an efficient method of solution for the two- dimensional self-adjoint second-order elliptic difference equations. A discussion of iterative techniques and the concept of rate of convergence is given. For various block iteration methods, sufficient conditions were obtained which ensure a higher asymptotic rate of convergence for the cyclically reduced system than for the original system. The use of cyclic reduction in the numerical solution of two-dimensional second-order self-adjoint elliptic differential equations is illustrated. (M.C.G.)

Research Organization:
Westinghouse Electric Corp. Bettis Atomic Power Lab., Pittsburgh
DOE Contract Number:
AT(11-1)-GEN-14
NSA Number:
NSA-16-025691
OSTI ID:
4812735
Report Number(s):
WAPD-TM-327
Country of Publication:
United States
Language:
English

Similar Records

Preconditioned polynomial iterative acceleration methods for block tridiagonal systems
Conference · Mon Sep 01 00:00:00 EDT 1986 · OSTI ID:7181980

P-CYCLIC MATRICES: A GENERALIZATION OF THE YOUNG-FRANKEL SUCCESSIVE OVERRELAXATION SCHEME
Technical Report · Thu Aug 01 00:00:00 EDT 1957 · OSTI ID:4317093

Preconditioning by incomplete block cyclic reduction
Journal Article · Sat Mar 31 23:00:00 EST 1984 · Math. Comput.; (United States) · OSTI ID:6209063