不完全信息下多属性决策的证据推理方法

Evidential Reasoning Approach for Multi-Attribute Decision Making with Incomplete Information

针对权系数未知、属性值具有缺失的不完全信息下多属性决策问题,给出了一种基于证据推理的方法。该方法通过对决策矩阵的规范化处理、不同属性下焦元的确定及评价值计算,获得证据的基本概率分配,然后利用证据合成,并通过信任区间的计算与排序,获得整个方案集的排序。最终,实例表明了该方法的有效性和可行性。

A kind of multi-attribute decision making problems with unknown weights and missing values is studied. A method based on evidential reasoning for the mentioned multi-attribute decision making problems is proposed. Firstly, the basic probability assignment is gotten by standardizing the decision matrix values, confirming focal element under different attribute and computing their score values. Then, by employing evidential combination defined in Dempster-Shafer theory, the ranking of the whole aheruatives can be attained through trust interval computing and ranking. Lastly, an example is given to explain the feasibility and availability of this method.