TY - JOUR
T1 - Accelerating Simulation of Population Continuous Time Markov Chains via Automatic Model Reduction
AU - Feng,Cheng
AU - Hillston,Jane
PY - 2018/1/11
Y1 - 2018/1/11
N2 - We present a novel model reduction method which can signicantly boost thespeed of stochastic simulation of a population continuous-time Markov chain(PCTMC) model. Specifically, given a set of predefined target populations ofthe modellers' interest, our method exploits the coupling coefficients betweenpopulation variables and transitions with respect to those target populationswhich are calculated based on a directed coupling graph constructed for thePCTMC. Population variables and transitions which have high coupling coeffi-cients on the target populations are exactly simulated. However, the remainingpopulation variables and transitions which have low coupling coefficients can ei-ther be removed or approximately simulated in the reduced model. The reducedmodel generated by our approach has signicantly lower cost for stochastic sim-ulation, but still retains high accuracy on the statistical properties of the targetpopulations. The applicability and effectiveness of our method is demonstratedon two illustrative models.
AB - We present a novel model reduction method which can signicantly boost thespeed of stochastic simulation of a population continuous-time Markov chain(PCTMC) model. Specifically, given a set of predefined target populations ofthe modellers' interest, our method exploits the coupling coefficients betweenpopulation variables and transitions with respect to those target populationswhich are calculated based on a directed coupling graph constructed for thePCTMC. Population variables and transitions which have high coupling coeffi-cients on the target populations are exactly simulated. However, the remainingpopulation variables and transitions which have low coupling coefficients can ei-ther be removed or approximately simulated in the reduced model. The reducedmodel generated by our approach has signicantly lower cost for stochastic sim-ulation, but still retains high accuracy on the statistical properties of the targetpopulations. The applicability and effectiveness of our method is demonstratedon two illustrative models.
U2 - 10.1016/j.peva.2017.11.004
DO - 10.1016/j.peva.2017.11.004
M3 - Article
JO - Performance Evaluation
T2 - Performance Evaluation
JF - Performance Evaluation
SN - 0166-5316
ER -