Preconditioners based on random sampling for solving least squares problems

发布者:王丹丹发布时间:2023-05-09浏览次数:269


江苏省应用数学(中国矿业大学)中心系列学术报告

题目:Preconditioners based on random sampling  for solving least squares problems

报告人:殷俊锋教授单位:同济大学数学科学学院

间:2023512日(周五)下午16:00-17:00

地点:数学院 A310

报告人简介:殷俊锋,同济大学数学科学学院教授,博导,创新创业学院副院长。主要研究方向数值代数与科学计算,计算金融,大数据和人工智能。主持及参与国家自然科学基金、上海市及教育部等科研项目10余项,发表高水平SCI学术论文30余篇,2009年入选上海市“浦江人才”,2010年荣获中国数学会计算数学分会应用数值代数奖,2019年获中国数学会计算数学分会青年创新奖(提名),现为中国工业与应用数学学会副秘书长,中国工业与应用数学学会大数据与人工智能专业委员会委员,中国高等教育学会教育数学委员会常务理事。

 

 

Abstract:

For the solution of large sparse least squares problems, preconditioned Krylov subspace methods are usually the fist choice and the preconditioners play the important roles in accelerating the convergence of the iteration. After the study on incomplete QR decomposition preconditioners based on Givens rotation, we proposed the preconditioners on random sampling and QR decomposition for solving least squares problems. Theoretical analysis and numerical experiments are presented to show the efficiency of the preconditioners, compared with the existing preconditioners.