基于理想解的复杂系统非线性评估模型

Non-linear evaluating method for complex system based on ideal point

为解决非线性评估中可能存在的层次较多、属性混合且关联性强的问题,提出了一种基于理想解的非线性综合评价法。该方法首先利用灰色关联度法作指标筛选,用熵方法分析各指标重要程度并得到各底层指标的权重值,用AHP法得到底层以上的指标权重;然后,利用改进的TOPSIS方法得到各一级指标的评估分值;再由定义的优先等级隶属函数求得各一级指标的评估矩阵;最后,定义合成函数,利用模糊综合评分法,总合各个一级指标得分,得到被评对象的综合等级。案例分析表明:该方法具有良好的适用性及可操作性。

In order to solve the problems of multi-layers,mixed-attributes and strong relativity which exist in the non-linear evaluation,a non-linear integrated evaluating method is proposed based on ideal point.First,this method used the grey relationship method to filter indexes,the entropy analysis method to analyze the importance of bottom indexes and get their results.Other index results were obtained by using AHP method.Then the improved TOPSIS method was used to get every stair index evaluation points.According to the defined priority-grade subjection function,an evaluation matrix was obtained for each index.Finally,by defining compound function,using fuzzy integrate grade method and adding up the scores of all first-grade indexes,the integrated level of the evaluated objects was acquired.The example indicates this method is applicable and feasible.