基于改进的证据推理方法的多指标决策模型

对于一个多指标决策问题，证据理论可以通过构造辨识框架和基本概率分配函数、采取递归的证据合成方法。计算出原始数据在反映多个指标联合作用的情况下对不同判别结果的支持程度，并可以在信息复杂或数据不完整的条件下做出评估决策。本文首先建立基于证据推理的多指标评估问题的基本模型，然后引入了模糊数据方法以处理具有模糊概念或推理关系的复杂问题，同时还考虑了实际问题中可能出现加权证据或者相关证据的情况，其目的是为了建立一套具有实用性的、准确有效的多指标评估模型。文章最后设计一个风险评估的算例，分析了该方法的优点以及需要进一步完善之处。

Improved Multi-criteria Decision Making Model Based on Evidential Reasoning Method

To solve a multi-criteria decision making problem, we can construct identification frames and basic probability assignment functions, use the recursive evidence combination method, calculate the supporting degrees of original data to various evaluating results that reflect the influence of multi evidences, and make decisions in conditions of complex information or incomplete data. In this paper, we first propose an ordinary multi-criteria decision making model based on evidential reasoning, then apply fuzzy theory to solve the problems with fuzzy concept or relations. We also set up some formulae to deal with weighted or correlative evidence combination problems. Finally, we design a simple example to analyze the application process. The goal of our work is to build up a precise, effective and practical model for solving multi-criteria evaluating problems.