Accelerated Modulus-Based Matrix Splitting Iteration Methods for Linear Complementarity Problem

Ning Zheng,Jun-Feng Yin

DOI: https://doi.org/10.1007/s11075-012-9664-9

IF: 2.37

2012-01-01

Numerical Algorithms

Abstract:For the large sparse linear complementarity problem, a class of accelerated modulus-based matrix splitting iteration methods is established by reformulating it as a general implicit fixed-point equation, which covers the known modulus-based matrix splitting iteration methods. The convergence conditions are presented when the system matrix is either a positive definite matrix or an H (+)-matrix. Numerical experiments further show that the proposed methods are efficient and accelerate the convergence performance of the modulus-based matrix splitting iteration methods with less iteration steps and CPU time.