排队仿真定量建模与χ~2-检验的有效性评估

Quantitative Modeling of Queuing Simulation and Assessment of χ~2-Test

定量建模一直是国际仿真领域着力探讨的一个重要课题。国际上采用的主要有所谓“拟合理论分布法”和“提取经验分布法” ,并存在激烈争论。其核心是在基于概率方法进行模型推断中 ,先验知识的合理性和有效性。文中提出随机变量X的总体分布与理论分布的定义。讨论了 χ2 -检验和K检验的特点。提出了评价 χ2

The quantitative modeling has been an important problem studied in international simulation field. There are two main approaches to it, one of which is fitting theoretical distribution, the other is abstracting empirical distribution. The core of the controversy on them is the reasonability and validity of a prior information. Two definitions for an overall distribution and theoretical one for a random variable are provided respectively in this paper. The characteristics of χ2-test and K-test are discussed. A assessment measure for the validity of χ2-test is offered.