采用优化模型指标筛选的马田系统综合评价方法研究

Development of a Methodology for Comprehensive Evaluation Method Using Improved Mahalanobis Taguchi System based on Feature Selection Optimization Model

传统的马田系统主要用于分类与诊断.将马田系统作为一种综合评价方法进行研究,分别研究了有基准空间和无基准空间两种情形下的马田系统综合评价方法及步骤.针对传统马田系统变量筛选存在的缺陷,构建多目标规划模型进行评价指标筛选,采用遗传算法求解模型.通过两个实际案例,将马田系统综合评价方法与一些常用的综合评价方法对比研究,结果表明,马田系统可以筛选评价指标和避免指标赋权问题,是一种实用且有效的综合评价方法.

Traditional Mahalanobis-Taguchi System （MTS） is mainly used for classification and diagnosis. The authors research the MTS as a comprehensive evaluation method in the paper. Two kinds of MTS comprehensive evaluation methods are studied in two different kind of situations that reference space exists or not. Aiming at the inadequacy of screening variables of MTS, the authors propose a multi-objective program model and Genetic Algorithm is employed to solve the optimization problem according to characteristics of the model. Through two cases, MTS is compared with other comprehensive evaluation methods. According to the result, MTS can screen useful variables and avoid weights of variables problem and is a practical comprehensive evaluation method.