计算与应用数学拔尖博士生系列论坛(2017 SERIES 12)——Hierarchical Eigensolver Based on Multriesolutionl Matrix Decomposition
主 题: 计算与应用数学拔尖博士生系列论坛(2017 SERIES 12)——Hierarchical Eigensolver Based on Multriesolutionl Matrix Decomposition
报告人: Zijun Zhang (Mathematical Sciences, Peking University)
时 间: 2017-12-15 12:00-13:30
地 点: Room1418, Sciences Building No. 1
12:00-12:30 lunch；12:30-13:30 Talk
Abstract: We propose a new iterative method to hierarchically compute a relatively moderate/large number of smallest eigenvalues and eigenvectors of a sparse symmetric positive-definite matrix, including the well-known graph Laplacian. We exploit the properties of an adaptive operator compression scheme which lead to a multiresolution matrix factorization. The decomposition is integrated into the implicitly restarted Lanczos methods to form an extension-refinement iterative scheme. Theoretical analysis and numerical illustration are reported to show the effectiveness of the algorithm.
欢迎各位同学积极参加，报名链接https://www.wjx.top/jq/18390631.aspx，报名截止时间为12月14日16:00，我们将为报名同学提供午餐。
