基于实验设计的概率语言权重信息多准则决策方法

Multi-criteria Decision Making Method with Probability Linguistic Weight Information Based on Design of Experiment

作为模糊语言集的一种,概率语言集能够反映决策问题中评价准则权重重要程度及其概率,这在决策问题中较为实用。本文运用实验设计(Design of Experiment,DOE)的方法,对权重信息以概率语言给出的多准则决策问题进行研究:通过多次模拟权重重要程度及其概率,设计出多组实验权重,并用GRA法求得每组实验数据的得分,获得描述该决策问题准则值与评价结果相互影响关系的最适用回归方程,将决策问题的实际数据带入方程进行运算,得出方案排序结果。通过快速成型过程、柔性制造系统过程两个案例,进行对比验证。结果表明所提出的方法能较好地处理基于概率语言权重信息的多准则决策问题,计算简单且速度快,可重复性强且颇为实用。

As a kind of fuzzy language sets,the probability language sets can reflect the importance and the probability ofevaluation criteria weights,which is more practical in decision-making problems.In this paper,the design of experiment(DOE)method is used to study the multi-criteria decision-making problems given by probability language.Firstly,by simulating the importance and probability of weight repeatedly,the multiple weight of experiment can be designed.Secondly,the GRA method is used to obtain the scores of each experimental data and the most applicable regression equation describing the influence of relationship between the criteria value and the evaluation results of the decision-making problems can be obtained.Then the actual data of the decision-making problems is brought into the equation to calculate the ranking orders of alternatives.Finally,the proposed method is applied to two cases,includingrapid prototyping process andflexible manufacturing system process,to illustrate the application and feasibility of the method.The results show that the proposed method is appropriate to deal with the multi-criteria decision-making problem based on probability linguistic weight information.