Interval Sampling Method for Steady-State Simulation Output Analysis

A new algorithm namely the interval sampling method, applicable to the analysisof steady-state simulation output is proposed. This algorithm uses the time series analysisto carry out conrrelation analysis of the steady-state simulation output so as to obtain theobservation data which are actually uncorrelated in nature. On the basis of theseuncorrelated data gathered, some satisfactory deductions cam be made on the data under re search. A comparison between batch means method and the interval sampling method hasbeen performed by taking the M/M/l queuing system as an example. The results attestedthat the interval sampling method is mere accurate than the batch means method.

仿真输出分析中的BAYES方法

本文提出用Bayes方法对小样本或单次运行的仿真输出进行分析。给出了几种条件下的Bayes方法和伽玛分布置信区间的估计方法。并用M/M/I排队系统对几种方法进行了实例研究。实验结果表明,在处理小样本或单次仿真运行时Bayes方法有明显的优越性和应用前景。