报 告 人:黄荣
报告题目:A periodic qd-type reduction for computing eigenvalues of structured matrix products to high relative accuracy
报告时间:2017年12月9日(周六)下午16:00—17:00
报告地点:静远楼204学术报告厅
主办单位:数学与统计学院、科技处
报告人简介:
黄荣,博士,教授,博士生导师,湖南省杰出青年基金获得者、湖南省普通高校学科带头人培养对象、湖南省新世纪121人才工程人选、湖南省青年骨干教师培养对象。主要研究方向为数值代数等。主持国家自然科学基金面上项目、国家自然科学基金青年项目、教育部博士点基金、湖南省杰出青年基金等课题10余项。学术论文以第一作者发表在SIAM Journal on Matrix Analysis and Applications、Journal of Scientific Computing、BIT Numerical Mathematics、Numerical Linear Algebra with Applications、Numerical Algorithms、Journal of Computational and Applied Mathematics、Linear Algebra and its Applications、Linear and Multilinear Algebra等学术期刊上。
报告摘要:
In this paper, we consider the product eigenvalue problem for a wide class of structured matrices containing the well-known Vandermonde and Cauchy matrices. A periodic qd-type reduction method is developed for computing eigenvalues of products of these rectangular matrices so that no subtraction of like-signed numbers occurs. Consequently, all the eigenvalues of such a product are computed to high relative accuracy in a preferable complexity. Error analysis and numerical experiments are provided to confirm the high relative accuracy.