For more information about this meeting, contact Manfred Denker.
|Title:||Computational methods for continuous time Markov chains with applications to biological processes|
|Seminar:||Seminar on Probability and its Application|
|Speaker:||David Anderson, University of Wisconsin|
|I will discuss computational methods for continuous time Markov chains, which includes the large class of stochastically modeled biochemical reaction networks (and other population processes). I will show how different methods can be developed and analyzed by using path-wise representations for the processes, such as the random time change representation of Tom Kurtz. In particular, I will discuss a new multi-level Monte Carlo method that produces an unbiased estimator for the approximation of expectations, though with only a fraction of the computational cost of the traditional exact algorithms (such as Gillespie's algorithm) combined with crude Monte Carlo. Time permitting, I will also discuss some related work on computing parameter sensitivities for these stochastic models.|
Room Reservation Information
|Date:||01 / 13 / 2012|
|Time:||02:20pm - 03:20pm|