基于单值中智集和云聚类的产品造型设计决策方法

为了获得客观、合理的权重分配，综合考虑有限理性心理行为特征和策略操控行为，把决策专家进行科学聚类，提出基于单值中智集（SVNS）和云模型聚类的产品造型设计多属性决策方法.决策专家构造标准属性和备选方案的成对比较比率平方矩阵获得SVNS，映射真、假、不确定3个隶属度值到单值中智立方体（SVNC）中，筛选各标准属性下备选方案的评估结果，获得标准属性的相对权重.利用融合多粒度语言的云模型聚类方法集群决策专家，淘汰存在冲突和非理性的决策专家，获得有效的设计决策专家权重.由标准属性权重和决策专家权重，综合计算各备选方案的总体优先级得分并进行排序.以汽车造型设计方案为例，验证所提方法的可行性和有效性.结果表明，所提方法避免了恶意策略操作，有效地解决了复杂和不确定情况下的汽车造型设计多属性决策问题.

Decision method for product styling design based on single-valued neutrosophic sets and cloud clustering

A multi-attribute decision method for product styling design based on single-valued neutrosophic sets (SVNS) and cloud model clustering was proposed, in order to obtain an objective and reasonable weight distribution. Decision-making experts were scientifically clustered, considering the limited rational psychological behavior characteristics and strategic manipulation behaviors. The square matrix of the pairwise comparison ratios of the standard attributes and the alternatives was constructed by the decision-making experts to obtain the SVNS. Mapping the three membership values of true, false, and uncertain in the single-valued neutrosophic cube (SVNC), the relative weight of the standard attributes was obtained by screening the evaluation results of the alternatives under each standard attribute. The cloud model clustering method fused with multi-granularity language was used to cluster decision-making experts, conflicting and unreasonable decision-making experts were eliminated to obtain effective design decision-making expert weights. The overall priority score of each alternative was calculated and sorted through the standard attribute weights and the weights of decision-making experts. The feasibility and effectiveness of the proposed method was verified by an example of car styling design scheme. Results show that using the proposed method, dishonest strategic manipulation is avoided and the multiple attribute decision making problem for car styling design in complex and uncertain situations is effectively solved.