Series of Academic Reports of Jiangsu Applied Mathematics (China University of Mining and Technology) Center

Report Title: Preconditioned iterative methods with Anderson acceleration for solving multilinear PageRank and applications

Presenter: Associate Professor Liu Dongdong, School of Mathematics and Statistics, Guangdong University of Technology

Report Time: 10:00-11:00 a.m. on Friday, May 9, 2025

Report Location: Room A321, School of Mathematics

Speaker Profile: Liu Dongdong, Ph.D., is a master's supervisor and associate professor. She obtained her Master's degree in Applied Mathematics from South China Normal University in July 2015 and her Ph.D. in Philosophy (Mathematics) from the University of Macau in April 2018. Her main research direction is tensor computation. In June 2018, he/she was introduced to Guangdong University of Technology under the "Hundred Young Talents" talent Program. Two papers on the research of tensor complementarity problem and the splitting algorithm of tensor equations are ESI highly cited papers (top 1%). He/She is a council member of the Operations Research Society of Guangdong Province and the Computational Mathematics Society of Guangdong Province. He/She has presided over one National Natural Science Foundation of China Youth Fund project, one Regional Joint Fund project and one general project of the Natural Science Foundation of Guangdong Province, and two projects of the Guangzhou Science and Technology Plan. Have presided over two provincial and ministerial-level laboratory projects.

Report Summary: In this talk, combining with the relaxation technique, we propose new preconditioned splitting methods for solving the multilinear PageRank problem. Besides, the preconditioned splitting iterative methods with Anderson acceleration are also given. Furthermore, we provide the convergence analysis of the proposed methods. Numerical experiments including both the synthetic data and the real-world data demonstrate that the proposed methods perform well.