报告时间：11月23日 (星期三) 10:30-12:00              

报告地点：信息楼 228

邀 请 人：系统工程研究所  黄敏教授

Abstract:

This presentation contains two parts: The first one is to introduce our research on numerical computation in general stochastic models from 1997 to 2009 (and even to now). To that end, we found and developed two types (abbreviated as UDL-type and LDU-type) of RG-factorizations for any irreducible Markov process (and further for other key stochastic processes) through a constructive censoring technique under some higher skills. The second one of this presentation is to introduce our works (2010 to now) on Nonlinear Markov Processes in Big Networks together with the resultant applications of mean-field theory and RG-factorizations. On this research line, we have applied the mean-field theory as well as the RG-factorizations to discuss such nonlinear Markov processes from various practically large-scale stochastic systems including supermarket models, work stealing models, bike-sharing systems and healthcare systems. Based on this, we found a key conjecture that Black Hole Effect is a basic phenomenon in Big (Economy) Networks.

Biography:

李泉林教授，2003.12至2009.10曾任清华大学工业工程系副教授；2009至今任燕山大学经济管理学院教授、博士生导师。李泉林教授先后在加拿大Winnipeg大学、加拿大Carleton大学、香港大学、香港科技大学、香港中文大学、西班牙Complutense University of Madrid与澳门大学等国外院校从事于随机模型、排队网络、计算机网络、网络安全、网络资源管理、网络熵决策、超市模型、RFID技术与应用、物联网、大数据、云计算、数据中心网络、智慧能源网络、医疗服务系统、供应链管理等方面的合作研究工作。在理论研究方面，李泉林教授在国际上提出并发展了随机模型的RG-分解方法和理论应用，并由Springer出版英文专著《Constructive Computation in Stochastic Models with Applications: RG-Factorizations》。近五年来，李泉林教授在大型网络（资源管理）领域开展了系统性的研究工作并取得了关键性的理论进展，包括非线性马氏过程、多稳定域与跨域转换、黑洞效应与资源耗损、白洞资源涌现等等。他系统地研究了超市模型、负载调配模型、资源共享系统、网络熵决策、网络博弈以及大型服务系统的机制设计等重要理论问题。