考虑决策要素关联性的犹豫模糊非平衡语言多属性群决策方法及应用研究

Due to increasing complexity in decision making environments, a new type of more complex multiple-attribute decision making (MADM) problems have evolved from traditional ones, which hold complexity characteristics of ill-structure, uncertainty of decision hesitance, interdependences among fundamental components in MADM, and group decision making settings, thus cannot be resolved by direct use of existing approaches. Therefore, to tackle these complex problems from real decision making activities, this research project aims to construct effective linguistic multiple attribute group decision making (MAGDM) approaches with consideration of interdependences among main three decision components (namely, attributes, attributes input arguments and decision makers), which is of significant importance for both MAGDM theories and real-world applications. Firstly, to depict preferences of decision makers more effectively and flexibly, we focus on investigating comprehensive type of fuzzy information expression tool based on hesitant fuzzy unbalanced linguistic set (HFULS). We analyze the potential irrationality in conventional fuzziness measures for hesitant fuzzy set, then we develop an effective fuzziness measures for HFULS in order to reduce information loss or information distortion. Next, we investigate different types of interdependent relations (interrelations) among decision components and utilize the developed fuzziness measures for HFULS to gauge the degrees of various interdependence. Then, we model those interdependent relations among decision components on weighting information for both attributes and decision makers in MAGDM. Finally, we put forward effective approaches for MAGDM based on HFULS with comprehensive consideration of interdependencies among decision components, and further carry out studies on real-world applications.

日益复杂的决策环境使得传统多属性决策问题演化出一类具有高度非结构化、决策偏好犹豫不确定、决策要素关联性、群体参与决策等综合复杂性特征的多属性决策问题,而现有决策方法不能直接应用求解,故从实际出发,本项目聚焦研究更具优势和适应性的复杂偏好信息表征工具—犹豫模糊非平衡语言集,深入研究三类决策要素(属性、属性值与决策者)上关联性信息及其度量方法,进而科学构建能综合考虑决策要素关联性的犹豫模糊非平衡语言多属性群决策模型方法,具有重要理论与实际意义.项目首先针对传统犹豫模糊信息测度方法可能造成的信息损失和扭曲现象,研发犹豫模糊非平衡语言集下的信息测度体系;后对决策要素关联性进行分析建模,以构建属性与专家权重的综合确定模型,进一步融合到信息集成算子和选优方法中;最后,研究形成综合考虑决策要素关联性的犹豫模糊非平衡语言多属性群决策框架、模型与方法,并展开应用研究.

日益复杂的决策环境使得传统多属性决策问题演化出一类具有高度非结构化、决策偏好犹豫不确定、决策要素关联性、群体参与决策等综合复杂特征的多属性群决策问题，而现有的决策方法不能够直接求解，故从实际问题出发，本项目定义了一类更具优势和适应性的复杂偏好信息表征形式——双边犹豫模糊非平衡语言集及其拓展形式，针对实际决策过程中可能存在的不同决策情形，基于此种混合信息形式的主观和客观信息测度来解析决策建模要素（属性、属性值与决策者等）的关联性特征，从而构建不同决策要素的权重确定模型并形成更加科学的犹豫模糊非平衡语言决策模型。.项目首先针对传统的犹豫模糊信息测度方法可能造成的信息损失和扭曲等现象，研发了犹豫模糊非平衡语言集及其拓展形式下的信息测度体系，包括距离测度、熵测度、相似性测度等。进而利用基于信息测度的Power法和DEMATEL法分别从客观和主观两个方法对决策要素间的关联性进行分析，构建了具有普遍适应性的属性和专家权重综合确定模型。籍此，项目研究依据实际决策过程中存在的不同决策情形，提炼出决策要素间存在优先级关系、决策要素间存在支持关系、决策者存在额外态度特征信息这三类特殊决策问题，将决策要素权重确定模型融入到信息融合算子和选优方法中，构建针对这三类特殊决策问题的决策模型。此外，项目研究利用上述模型与方法应用于绿色供应商的选择问题、应急方案的选择问题等实际复杂决策问题，并通过详实的实验研究验证了所开发方法的实际有效性。.总体来看，项目研究所开发的信息形式及其信息融合算子与模糊测度方法、决策关联要素权重确定方法以及多属性群体决策方法能够形成较为系统的双边犹豫模糊非平衡语言多属性群决策框架，为应对高度复杂环境下的科学决策提供了有效途径。