Abstract: Interaction is a difficult and inevitable problem in group decision making (GDM). We should be clear about whom we need to negotiate with and whether the interaction is effective. This paper proposes a method to explain how to interact in GDM using hesitant fuzzy elements. We take the GDM as a network in which experts are nodes and theirs' interactions are links. There are three assumptions and two parts to describethe network value and interaction process, includingthe random dominant matrix and network utility,using the Bellman equation and hesitant fuzzy elements. More specifically, this method gives advice in two levels, and both group and experts could improve their strategies. Therefore, to promote the GDM efficiency, the method could compare the utility, network value and links simultaneously.

Key words: comprehensive evaluation, domain links, hesitant fuzzy element, stochastic domain

摘要： 基于评价网络结构,分析连接行为在群组评价中的作用。面向具有一定规模的群体,通过犹豫模糊元刻画其连接行为并测算评价网络效用期望。假设连接行为有互动成本;评价者越相似,互动的必要性越弱;连接行为是犹豫模糊的。首先由随机模拟连接计算网络效用并生成随机连接优胜矩阵;其次计算评价网络迭代价值函数;最后解释满足假设生成的评价网络价值,给出连接行为和互动过程终止建议。

关键词: 综合评价, 优胜连接, 犹豫模糊, 随机占优