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.