关于举行中山大学郭先平教授学术报告会的通知

报告题目：ConstrainedMarkov decision processes with varying discount factors

报告人：郭先平教授（中山大学）

报告时间：2017年1月10日（周二）下午16:00-17:00

报告地点：4号楼4141室

欢迎广大师生前往!

                                                     数学学院

                                                 2017年01月09日

Abstract:This talkfocuses on the constrained optimality problem of firstpassagediscrete-time Markov decisionprocesses in denumerable statesand compact action spaces with multi-constraints, state-dependentdiscount factors and possibly unbounded costs.By means of theproperties of a so-called occupation measure of a policy, we showthat the constrained optimalityproblem is equivalence to an(infinite-dimensional) linear programming on the set of occupationmeasures with some constraints, and thus prove the existence of anoptimal policy under suitable conditions. Furthermore, using theequivalence between the constrained optimalityproblem and the linearprogramming we obtain an exact form of an optimal policyfor the caseof finite

statesand actions. Finally, as an example, a controlled queueing system isgiven to illustrate our results.