一种多准则纯语言群决策方法

针对权系数信息和方案的准则值为确定语言等级,或位于两个语言等级之间,甚至缺失的群决策问题,提出一种新的决策方法.该方法利用证据推理算法对准则权系数和方案值在准则下进行群体集成,采用二元语义对方案进行语言集结,并用方案与理想方案的二元语义间距离和群体集成信息等构建非线性规划模型,使用遗传算法求解优化模型,进而得到方案的排序.最后通过实例说明该方法的可行性和有效性.

Multi-criteria group decision-making approach with linguistic assessment information

For group decision-making problems that the decision-makers' weights and the criteria's weights and the decision-makers' preference information are represented by the form of linguistic grade,or between two linguistic grade,or even default,a new approach based on evidential reasoning(ER) and two-tuple is proposed.Using the ER algorithm,the criteria's weights and the alternatives values are aggregated,and their belief degrees belonging to each linguistic grade are worked out and transformed into two-tuple.The nonlinear programming model is constructed based on the two-tuple distances between alternatives and ideal alternative,and the information of weights on decision-makers and criteria.By using genetic algorithms,the optimal weights on decision-makers and criteria is gained,and then the rank of the alternatives is obtained.Finally,an illustrative example shows the feasibility and the effectiveness of the method.