Replication splitting and variance for simulating discrete-parameter stochastic processes
暂无分享,去创建一个
Previous results for stationary continuous-time processes concerning allocation of a fixed amount of simulation effort across independent replications are extended both to stationary and certain non-stationary discrete-time processes. In particular, in the presence of positive autocorrelation, variance is reduced if more short replications are designed. The magnitude, however, of the variance reduction is not great as long as the computation budget is not tight, suggesting that a good strategy is to design for a moderate number of replications in any case, which also mitigates potential bias problems.
[1] Joseph M. Sussman,et al. Toward guidelines for designing experiments in queuing simulation , 1977 .
[2] C. J. Ancker,et al. Mean Value Estimation from Digital Computer Simulation , 1966, Oper. Res..
[3] Averill M. Law,et al. An Analytical Evaluation of Alternative Strategies in Steady-State Simulation , 1984, Oper. Res..
[4] George S. Fishman,et al. Bias Considerations in Simulation Experiments , 1972, Oper. Res..