报告一：Measure-valued Generator of Piecewise Deterministic Markov Processes

报告人：刘国欣（石家庄铁道大学教授、副校长）

摘 要：The purpose of this talk is to introduce a new concept on generators of piecewise deterministic Markov processes (PDMPs, for short), the measure-valued generator, which is an extension of the extended generator for PDMPs. The domain of measure-valued generator is extended to the class of (path-)locally finite of variation functions. The PDMPs here are just as the jumping Markov processes introduced by Jacod and Skorohod (1996) which is greatly generalized from Davis' PDMPs. Roughly speaking, the PDMP here is a Markov process with discrete-type natural filtration. We begin with a brief survey of the development of the theory of PDMPs and give the characteristic triple (\phi; F; q) for a PDMP, which plays the same role as Q-matrix for a Q-process. As a preparation, we define the additive functionals of the semi-flow $\phi$ and discuss their properties. Basing on these, we introduce the measure-valued generator and present the characterization of its domain. By the way, we get the corresponding Ito type formula. As the application of the measure-valued generator, we present the measure integro-differential equation satisfied by the expectation of an additive functional of a PDMP.

报告二：Stochastic control and stochastic filtering in insurance

报告人：张帅琪 （广东工业大学助理教授）

摘 要：In this talk, three classes of stochastic control problems, including optimal dividend, reinsurance and investment problems are introduced. Dynamic programming principle and maximum principle are employed to deal with the problems. When we consider the partial information case, non-linear filtering will be used to solve the problem. We give several examples to illustrate the application of stochastic control in actuarial science.

时间： 2015年12月28日（星期一） 10：00—12:00

地点：中央财经大学学术会堂南楼506会议室

欢迎各位老师同学参加！

(责任编辑：网站编辑)