基于单峰偏好理论的动态模糊偏序多属性群决策方法

Dynamic Fuzzy Partial Order Multi-attribute Group Decision-making Method Based on Single-peak Preference Theory

考虑多属性群决策问题的动态性、模糊性及不同决策阶段消费者偏好的偏序性,文章基于属性值独立于个体偏好的性质提出单峰形式的动态模糊偏序多属性群决策研究的新方法。首先,将消费者在不同决策阶段的模糊多属性偏序矩阵转化为单峰性投票数矩阵,依据各属性权重进行属性集结,根据消费者的影响权重进行群体集结,利用各阶段的影响权重进行阶段集结;然后,以三次集结的单峰性投票数矩阵为约束条件,构建消费群体异议最小的备选方案排序的目标函数;最后,收集大众点评网2018年12个月消费群体在线真实的评分和点评数据进行模型求解,同时,将所建模型的求解结果与已有算法所得结果进行对比分析,从而说明所建模型的可行性及有效性。

Considering the dynamics and fuzziness of multi-attribute group decision-making and the partial order of consumer preferences in different decision-making stages, this paper proposes a new method of dynamic fuzzy partial order multi-attribute group decision-making with single peak form based on the property that attribute values are independent of individual preferences. Firstly, the fuzzy multi-attribute partial order matrix of consumers in different decision-making stages is transformed into single-peak vote number matrix;attribute aggregation is carried out according to the weight of each attribute;group aggregation is carried out according to the influence weight of consumers, and stage aggregation is carried out by using the influence weight of each stage. Then, the objective function of ordering the alternatives with the minimum objections is constructed with the single-peak voting matrix of three aggregates as the constraint condition. Finally, the real online ratings and comments from consumer groups of Dianping in 12 months of 2018 are collected to solve the model. At the same time, the solution of the model is compared with the results of the existing algorithms, so as to demonstrate the feasibility and effectiveness of the proposed model.