The Numerical Simulation Researches Of Comprehensive Evaluation Methods’Performances

Comprehensive evaluation is a method which evaluates the objects objectively, equitably and reasonably. With the increasing research of comprehensive evaluation, comprehensive evaluation methods enrich continuously. Such as Delphi, score evaluation, composite index evaluation, efficacy coefficient method, Multi statistical evaluation method, fuzzy comprehensive evaluation, gray system evaluation method, the analytic hierarchy process, and the recent years’comprehensive evaluation methods as sequence relation analysis method, data envelopment analysis, grad scale-up method, artificial neural networks and so on. The evaluate methods are constantly complex. With the development of computer technology, comprehensive evaluation method is also gradually developing its study by computer. Numerical simulation method is the gradual emerging of such development.Resorting to computer, relying on computer software, numerical simulation researches a problem or a process by numerical calculation and image analysis or other ways. This method can well display the research process and explain the research result, and it has made a lot of engineering research results, but is seldom used in the theoretical studies of comprehensive evaluation.This article tries to introduce numerical simulation method to comprehensive evaluation problems, finishes three-part work:First of all simulates the integrated non-dimensional methods, studies the nature of integrated approach. Secondly, simulates the stability of four kinds of differential comprehensive evaluation methods including grad scale-up method, mean square error method, maximizing deviation method and entropy method. Studies the relevance of evaluation results adding an extreme sample with the evaluation results without increasing extreme sample, and describes the methods processing extreme samples. The third section discusses using numerical simulation method to resolve the sequence relation analysis method’s sequencing problem when the ratio of the relative importance of the degree between indices are intervals.