排队模拟的运行设计与误差修正

Design of Queue Simulation and It's Errors Correction

排队模拟实验的运行设计，一直是国际仿真模拟界研究的重要课 题［１－３０］。虽然已有一些被普遍接受的排队模拟实验运行设计方法和模拟运行结果数据的统 计学处理方法及准则［４，５］但模拟结果的误差问题至今仍未很好解决。这些方法在一定程 度上减小了运行误差，但这种减小仍然是有限的。为了使仿真模拟结果对系统分析提供 有价值的数据，必须对模拟结果进行修正。本文提出了三种误差修正方法。即１．松弛法“修正；２Ｍ／Ｍ／１型排队模拟误差曲线修正；３．置信区间修正。这些修正都是基于对 Ｍ／Ｍ／１型排队模拟结果与其理论值的误差的计算与分析的。

Design of queue simulation experiment has been an important research subject for more than two decades. There has been five major experiment methods for queue simulation which can't completely remove errors and bias in simulation. So the precision of queue simulation can't be improved only by design of simulation. It is necessary to analyze and correct the errors in queue simulation. Three corrections are presented.