Speaker: Dr. Ting Wang (University of Delaware)

Title:Parallel in time sampling for continuous time Markov chains with rare events

Time: Tue, Oct.9, 2018 , PM:4:30-5:30.

Location: Room 4318, Building No.4, Wushan Campus

Abstract:

    Continuous time Markov chain (CTMC) is an important tool in applications such as modeling the stochastic reaction kinetics, queueing systems, etc. In this talk, we consider the problem of sampling the stationary distribution of ergodic CTMCs that exhibit multiple time scales. This is often computationally demanding due to the rare events associated with the slow time scale. On top of the parallel replica (ParRep) dynamics originally designed by Art Voter, we develop a parallel in time method for steady state sampling of CTMC. We provide theoretical analysis of ParRep using the theory of quasi-stationary distribution. If time permits, we will also discuss the parametric uncertainty quantification of CTMC. To conclude, the application of ParRep to several biological examples such as the genetic switch will be demonstrated.  

Biography:

  Ting Wang is a postdoc researcher in the Department of Mathematical Sciences at University of Delaware, working with Professor Petr Plechac on uncertainty quantification. His research mainly focuses on stochastic methods and uncertainty quantification for complex physical systems with an application in molecular dynamics simulations. He received his Ph.D from the University of Maryland Baltimore County under the supervision of Professor Muruhan Rathinam.