考虑偏好反转的区间不确定多属性决策方法  

Approach for multiple attribute decision making under interval uncertainty considering preference reversal

作　　者：张美璟[1,2] 王应明[1] 陈圣群[2] 李锴[2] 

机构地区：[1]福州大学决策科学研究所,福州350116 [2]福建江夏学院电子信息科学学院,福州350108

出　　处：《控制与决策》2016年第11期2019-2024,共6页Control and Decision

基　　金：国家杰出青年基金项目(70925004);福建省自然科学基金项目(2015J01279)

摘　　要：针对实际决策中的不确定性和偏好反转问题,提出一种区间不确定多属性决策方法.该方法通过证据推理方法集结区间不确定评估信息,采用累积前景理论代替主观期望效用理论构建方案的综合前景价值,从而应对不确定环境下可能的决策偏好反转,将区间可能度用于方案综合前景价值排序.介绍了决策过程,给出了求解方案综合前景价值的非线性规划模型,并通过实例验证了方法的可行性、合理性和有效性.Aiming at the uncertainty and preference reversal in practical decision making problems, an approach for multiattribute decision making under interval uncertainty is proposed. The interval assessments are integrated by the evidential reasoning. Then the prospect value of each alternative is worked out by using cumulative prospect theory, substituting for the subjective expected utility theory, to handle with preference reversal in possibility under uncertainty, and the interval possibility degree is used to order all alternatives. The process of decision making is provided, and a pair of nonlinear programming models is built up to calculate the overall prospect value of each alternatives. An example is given to illustrate the feasibility, rationality and effectiveness of the proposed approach.

关 键 词：累积前景理论 证据推理 多属性决策 区间不确定 偏好反转 

分 类 号：C934[经济管理—管理学]