Combination Evaluation Algorithm Study Based On PSO

In the evaluation of the indicators, there are many methods that can be used, and each method has their advantages and disadvantages, therefore in the process of evaluation, it's best to use multiple methods to assess. Due to different evaluation methods, the sorting results may be different, it is necessary to carry out combination evaluation. However, the combination evaluation results may also be different. How to use the information getting from these various combination evaluation methods is a worthwhile issue.In order to make full use of sorting information from the combination evaluation, the rational and integrated optimization methods should be used in determining the final ranking results to avoid attending to one thing and losing another. Therefore this thesis constructs a nonlinear programming model based on PSO algorithm for comprehensive optimization to resolve such issue.In this thesis, firstly single evaluation method and its classification as well as four combination evaluation methods: average, Borda, Copeland and fuzzy Borda method are introduced; secondly present the PSO algorithm origins and concepts, theories analysis and algorithm structure, parameters choice and multi-objective optimization problems, and then discuss the PSO algorithm applicability and advantages in solving issues. Finally after considering some single evaluation methods results and four combinations evaluation sorting results, an example is given to analyze and calculate, and it develops a comprehensive optimization model, then uses PSO parallel calculation to simulate in order to get the final sorting results which provide specific ideas, methods and experimental basis for the inconsistency of sorting results from different evaluation methods.Example shows that Combination evaluation algorithm study based on PSO proposed in this thesis has many advantages, such as rapid convergence and stability, and it is an effective and practical evaluation method. Its finall sorting programme will be credible. This method can provide policy makers with a reliable basis for evaluation.