• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

Computer Engineering & Science ›› 2022, Vol. 44 ›› Issue (10): 1885-1892.

• Artificial Intelligence and Data Mining • Previous Articles     Next Articles

An extrapolated and generalized HSS method for non-Hermitian positive definite linear systems

WU Si-ting,BAO Liang   

  1. (School of Mathematics,East China University of Science and Technology,Shanghai 200237,China)
  • Received:2021-03-21 Revised:2021-05-20 Accepted:2022-10-25 Online:2022-10-25 Published:2022-10-28

Abstract: How to efficiently solve non-Hermitian positive definite linear systems is discussed, and an extrapolated and generalized Hermitian and skew-Hermitian splitting iteration method (EGHSS) is proposed. Firstly, the method constructs a new asymmetric two-step iterative scheme based on the generalized Hermitian and skew-Hermitian splitting iteration. Secondly, the convergence of the new method is theoretically analyzed, and the necessary and sufficient conditions for the convergence are given. Finally, numerical experiments show that the EGHSS iteration method is more effective than the GHSS iteration method and the EHSS iteration method in dealing with some problems.


Key words: non-Hermitian positive definite linear systems;GHSS iteration method;EHSS iteration method;convergence ,