决策信息为区间数的多属性群决策方法

摘要

为了解决群体决策过程中难以精准确定指标权重和专家权重,以及难以得到精确决策信息的问题,提出了一种基于集值统计和信息公理相结合的群决策方法.该方法将决策个体对各指标的决策信息采用区间数表示；用集值统计模型计算决策群体对各指标决策信息的综合评价值,在此基础上引入观点分歧度概念,以判定综合评价值的有效性；以有效的综合评价值作为决策群体期望的指标设计范围,根据信息公理计算各方案的信息含量来判断方案的优劣,并给出该决策方法的具体步骤.最后,通过实例说明了该方法的有效性；实验数据分析表明,该方法在群决策中避免了确定权重的复杂计算过程,降低了决策的复杂度,更加符合决策者的思维习惯.%In the process of group decision-making, it is difficult to determine accurate index weight and expert weight coefficients, and obtain precise decision-making information. To solve this problem, a group decision-making method was presented based on a set-valued statistics model and the information axiom. In this method, individual decision information for each index is expressed by interval numbers, the comprehensive evaluation value of each index is calculated using the set-valued statistics model for decision-making group, and the availability of comprehensive evaluation values is determined through the viewpoint divergence degree. Taking the effective comprehensive evaluation values as the index design range of decision group expectation, the information content of each alternative scheme is calculated by using the information axiom. According to the information content, the ranking of alternatives was judged, and the concrete evaluation steps are given for this method. Finally, the validity of the proposed method was demonstrated through a case study. The experimental data analysis shows that this method can avoid the complex calculation process of determining weight coefficients, reducing the decision-making complexity, and fitting decision makers' thinking habits.