2024 CSU DISTINGUISHED SCHOIARS FORUM

CONFLUENCE OF WISDOM  MEETING OF TALENTS

Speaker Profile

-          B.S. (1982), M.S. (1985, supervisor: Academician Shi Zhongci), University of Science and Technology of China.

-          Ph.D. (1991, supervisor: Professor Ivo Babuska), University of Maryland, College Park.

-          Professor at Wayne State University since 2002.

-          He was selected for the National Talents Program (2010) and the National Program for Introducing Overseas High-level Talents (2012).

-          He served or currently serves on the editorial boards of 10 domestic and international mathematics journals, including Mathematics of Computation, Journal of Scientific Computing, Numerical Methods for Partial Differential Equations, Journal of Computational Mathematics, Communications on Applied Mathematics and Computation, CSIAM Transaction on Applied Mathematics, etc., and has published over 200 SCI papers.

-          Before 2014, he chaired 10 NSF (National Science Foundation, US) projects. Since 2014, he has chaired 8 NSFC (National Natural Science Foundation of China) projects under the Key Program, General Program, Tianyuan Fund Program, and International Collaboration Program.

Professor Zhang Zhimin has long been engaged in the research of computational methods, especially the finite element method. He has achieved a number of innovative results in the areas of element construction, superconvergence, a posteriori error estimation, and adaptive algorithms. He was the first to establish the popular mathematical theory of ZZ discrete reconstruction format in the world. His Polynomial Preserving Recovery (PPR) method was adopted by the major commercial software COMSOL Multiphysics in 2008 and has been used ever since.

Report Abstract

Why spectral methods are preferred in PDE eigenvalue computations---- in some cases? When approximating PDE eigenvalue problems by numerical methods such as finite difference and finite element methods, it is well-known that only a small portion of the numerical eigenvalues are reliable. In contrast, spectral methods can perform extremely well in some situations, especially for one-dimensional problems and certain special higher-dimensional cases. Furthermore, we demonstrate that spectral methods can outperform traditional methods and the state-of-the-art methods for two-dimensional problems, even those with singularities.

Organized by: Department of Human Resources, Central South University

Hosted by: School of Mathematics and Statistics, Central South University

Original article link: https://math.csu.edu.cn/info/1736/13349.htm