基于层次分析法和信息熵的多准则指标选取

在新装备测试指标选取过程中,由于列装时间短,测试样本少,测试条件不足,通常选取工作无据可依。通过分析层次分析法、Dijkstra法、粗糙集法的优缺点,提出一种基于层次分析法、群组G1法和信息熵理论的指标多准则选取方法,将指标按几个基本准则两两比较获得原始数据,用群组G1法通过专家评判和信息熵法分析原始数据来确定评价准则的权值,最后对原始数据进行加权来对指标进行权重判断。文章结尾给出一个实际算例。方法依赖于两两比较数据,具有一定客观性,且无需进行样本测试,简化了指标选取工作。

Multi-criterioned Index Selecting Method Based on AHP and Entropy Theory

Index selecting of a new equipment is often cumbered by the lack of testing samples and resources. A multi-criterioned index selecting method based on AHP and entropy theory is put forward after contrast Analytic Hierarchy Process (AHP), Dijkstra and rough set. In this method, indexes are compared by each other according to several basic criterions to build up judging matrix, which will be analyzed under Entrop theory to decide the power of each criterion. Subjective attitude of experts is also collected to be analyzed under G1 method to decide the power of criterions. Then original data will be summed by power of criterions. A example is given in the end of the paper. This method simplifies the index selecting by its objective comparison and none-sample-demand attribute.