学术报告:QR algorithm with two-sided Rayleigh quotient shifts

审核发布:数学与信息学院 来源单位及审核人: 发布时间:2022-10-26浏览次数:124

报告人:陈小山教授 华南师范大学

会议时间:2022年10月31日, 15:30-17:30 (GMT+08:00) 中国标准时间 - 北京
点击链接入会,或添加至会议列表:
https://meeting.tencent.com/dm/wSsP9vwjdJBw
#腾讯会议:243-698-970
复制该信息,打开手机腾讯会议即可参与。

【注:因疫情原因,原线下会议调整为线上会议】  

  

摘  要:We introduce the two-sided Rayleigh quotient shift for the QR algorithm for non-Hermitian matrices. For the singly shifted case, the two-sided Rayleigh quotient iteration is incorporated in the QR iteration in an economic manner. A modified version of the method is proposed, where the two-sided Rayleigh quotient is computed directly from current upper Hessenberg matrix. Based on the observation that the Francis double-shift QR iteration is related to a 2D Grassmann-Rayleigh quotient iteration, we propose the doubly shifted QR algorithm with the two-sided 2D Grassmann-Rayleigh quotient double-shift. A modified version of the method is also proposed. For semi-simple eigenvalues, we claim that the QR algorithm with either the two-sided Rayleigh quotient shift or the 2D Grassmann-Rayleigh quotient double-shift has a cubic local convergence rate.Numerical examples are presented to support the claim. The proposed algorithms may be used for computing the Schur form of submatrices needed in the multishift QR algorithm with aggressive early deflation.

  

报告人简介:陈小山,华南师范大学数学科学学院教授,博导。主要从事数值线性代数及其应用方面的研究工作,主持一项国家自然科学基金和二项省自然科学基金。2012年获广东省自然科学技术奖二等奖(排名第三)。在《Numersche Mathematik》,《SIAM J. Matrix Anal. Appl.》,《Numer. Linear Albebra Appl.》,《BIT Numer. Math.》和《Linear Algebra Appl.》等国际知名期刊上发表论文40多篇。

  

欢迎广大师生参加! 

 

返回原图
/