基于直觉模糊和证据理论的混合型偏好信息集结方法  

Method to aggregate hybrid preference information based on intuitionistic fuzzy and evidence theory

作　　者：史超[1] 程咏梅[1] 潘泉[1] 

机构地区：[1]西北工业大学自动化学院,西安710072

出　　处：《控制与决策》2012年第8期1163-1168,共6页Control and Decision

基　　金：国家自然科学基金项目(61075029);航空科学基金项目(20100853010)

摘　　要：针对混合型偏好信息的集结问题,利用直觉模糊实现了5类混合型偏好信息的统一,提出利用冲突系数和Jousselme距离综合表示的证据冲突度计算专家权重的方法,最后利用Demsper组合规则对专家权重修正后的证据进行组合.实例分析表明,所提出的偏好信息统一方法能够保持转化前后方案的优先顺序不变,通过专家权重对专家意见的修正使得证据合成结果能体现多数专家的一致意见,具有较高的实用价值.Aiming to the aggregation of hybrid preference information, five kinds of hybrid preference information are unified by intuitionistic fuzzy. Then expert weight is computed by evidence conflict degree based on conflict coefficient and Jousselme distance. Finally, Demsper combination rule is used to combine the mass functions modified. An illustrative example shows that the proposed method can preserve the alternative＇s preference order, and the result of evidence aggregation is in accordance with the majority of experts because of mass function modified by expert＇s weight, which has high availability.

关 键 词：直觉模糊 证据理论 偏好信息 

分 类 号：C934[经济管理—管理学]