Kemeny's constant for singularly
perturbed Markov chains
报告专家：刘源远 教授（中南大学）
报告时间：2025年4月11日（星期五）下午15:00-16:00
报告地点：线上；腾讯会议：376-134-966 会议密码：250411
报告摘要：In this talk, we present a systematic procedure for computing Laurent series expansions of the fundamental matrix and Kemeny's constant for singularly perturbed Markov chains. We establish upper and lower bounds for Kemeny's constant under perturbation, and introduce a novel condition number formulation for sensitivity analysis. Our results extend existing perturbation frameworks and provide new tools for analyzing the Hamiltonian cycle problem. The results are demonstrated through a queueing example.
专家简介：刘源远，中南大学数学与统计学院教授、博士生导师、副院长（主持工作）。从事马氏过程的遍历性、扰动理论、排队网络的渐近性、以及带块结构的多维马氏过程领域的研究工作，论文发表在《SIAM Journal on Matrix Analysis and Applications》、《Advances in Applied Probability》、《Queueing Systems》、《Science China Mathematics》等刊物上。入选湖南省芙蓉特聘教授，现担任《应用概率统计》期刊编委。
邀请人：胡泽春
