针对矛盾体的闭环综合评价体系建立研究与实证分析

项目名称： 针对矛盾体的闭环综合评价体系建立研究与实证分析

项目编号： No.61304183

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 自动化技术、计算机技术

项目作者： 马丽叶

作者单位： 燕山大学

项目金额： 24万元

中文摘要： 针对具有矛盾体的多属性系统，本项目建立了一套新的综合评价体系，有效地将矛盾体的两个对立因素结合起来评价。本项目抽象出"安全性与经济性"这一矛盾体进行描述：首先，基于群灰色模糊聚类建立一个对立因素（安全性）评价指标体系；并利用证据融合理论确定评价指标权重；同时基于可拓理论建立安全性的等级评价模型，并且根据数据类型分别构建了可拓云模型和区间可拓模型，最终实现对安全性的等级划分。其次，针对另一个对立因素（经济性）建立评价指标体系；基于支持向量域描述建立不同安全等级下经济性的最小包围超球；通过计算当前位置与最优经济超球面的距离，并给定距离判定准则，最终完成对确定安全等级下经济性的评价。接着，通过建立安全性等级划分的闭环调整机制，将两个对立因素再次结合，并确定最优的调整方案。最后，对本项目提出的评价体系进行实证分析，依据仿真结果对综合评价系统进行理论完善，同时验证其正确性和合理性。

中文关键词： 综合评价体系；最优安全经济域；供电能力；超球支持向量机；可拓云模型

英文摘要： For Multi-attribute system with contradiction, a new comprehensive evaluation system is proposed, which can combine two opposing factors of contradiction effectively. The contradiction of "security and economic" is abstracted to describe. Firstly, the evaluation index system for an opposing factor (security) is established based on group gray fuzzy clustering; and by using evidence fusion theory the evaluation index weights are determined; then the level evaluation model for security is builded based on extension theory, at the same time, extension cloud model and the interval extension model are constructed according to the data type to achieve the level of security division. Secondly, the evaluation index system for another opposing factor (economy) is established; and based on the support vector domain description the economic smallest enclosing hyperspheres for different security levels are established; then the distance between the current location and optimal economic hypersphere is calculated, meanwhile, the distance criteria is given to complete the economic evaluation of determined the level security. Thirdly, the closed-loop adjustment mechanism for security classification is established, which can combine the two opposing factors organically again, and the optimal adjustment program is given. Finally,

英文关键词： comprehensive evaluation system；the optimal security and economic region；supply capacity；hypersphere support vector machine；extension cloud model