Research On Intelligent Evaluation Model And Expert System

This thesis gave an intelligent evaluation model and developed a software package (CEC).It consisted of five principal departments as follows:Firstly, new arithmetics were designed based on the knowledge of reachable matrix in graph theory. By using these, hiberarchy and opinion matrix could be automatically formed. As a result, the design of hiberarchy is predigest and opinion matrix can be also used as recommendation to opinion expert.Secondly, we pointed out and analyzed the errors in [11-14] and gained several properties for convex combination and hadamard convex Combination of judgment matrix. The conclusion is useful for group decision in AHP.Thirdly, an index of the compatibility of a group decision was given. Based on the idea of minimizing, the compatibility of a group decision an algorithm of sorting for group decision was gained. Then the convergence theorem for this algorithm was established. At the same time an NLP model of sorting for group decision was gained based on "two win". The NLP was worked out by using Genetic Simulated Annealing. An example was given at last.Fourthly, via applying the method of ANP, the evaluation system on innovative technical ability was founded and the relative weight of each guide line was gained. By the evaluation system, the innovative technical ability of six provinces (Beijing, Shuanghai, Guangdong, Sichuan and Shanxi) was appraised. Five aspects(knowledge innovation,knowledge fluxion, innovative technical ability of corporation, technical innovation circumstance, the economic effect on technical innovation ability) of the six provinces were studied comparatively. Research results provided quantitative basis for policy and layout on science and technology of the East and West China.Finally, the Chinese Expert Choice System (CEC) was developed, which was applied to research on ticket price scheme of Chongqing light rail transit.