Research On The Deaggregation Modeling Of Generalized Choquet Integral Group Decision Under Social Interaction Behavio

Group decision making is a process of making decision by coordinating the preferences of all decision makers to obtain consistent preferences.There are two kinds of interactive group decision making theories: one is group interactive social network group decision making based on direct preference information,and the other is individual interactive preference learning group decision making based on indirect preference information.In social network group decision-making,the research on analyzing psychological factors among decision makers by means of social network analysis method is still very weak.In preference learning group decision making,there are often some problems such as low density and large granularity of preference information,so how to mine and reproduce richer and finer preference information is an urgent problem to be solved.In order to address the above issues,the main research contents and innovations of this paper are as follows:(1)Since Choquet integral can be utilized in both interactive group decision making theories,so this paper further improves Choquet integral theory,that is,develops a hierarchical-levelbidirectional generalized Choquet integral.This paper studies the generalized Choquet integral’s properties,expression forms,M?bius transform,k-additivity and some indices,discusses the mutual transformation between generalized Choquet integral and traditional Choquet integral,and analyzes the expression of parameters in two modeling methods,which provides a theoretical basis for the construction of two interactive group decision making models.(2)For social network group decision making,multi-stage fluctuation utility function is used to describe the psychological preference of decision maker at different stages,social network empathy opinion evolution model is developed to simulate the evolution of opinions,the synergy and redundancy effects of Choquet integral is employed to simulate the cooperation and competition among decision makers,and social network topology structures and characteristic indicators are introduced to simulate different group decision making environments.Therefore,based on the above,from the perspective of social choice and personal value,this paper uses the method of social network analysis to build minimum cost consensus models to study the evolution of opinions and the law of consensus convergence.(3)Aiming at preference learning in group decision making,this paper focuses on optimizing the disaggregation process of group preference: due to the lack of preference information caused by cognitive limitation,this paper induces decision makers to give rich preference information from the dimensions of criteria structure and alternatives level;due to the difficulty of reproduction by the complexity of preference,a hierarchical-level-bidirectional generalized Choquet integral is developed to reproduce high-dimensional complex preferences;due to the preferences conflict caused by the group knowledge heterogeneity,a preference adjustment model considering resource consumption,fairness and individual differences is constructed to explore the consensus region of consistent group preferences;due to the overfitting problem caused by the high dimension of preference parameters,a minimum criteria cardinal model is constructed to reduce the preference parameters and obtain a concise expression of preference;in order to output stable results,non-additive robust ordinal regression and stochastic multiobjective acceptability analysis are combined to output necessary class,possible class and class acceptable index.In short,the two theories of social network group decision making based on empathy interaction and preference learning group decision making based on disaggregation theory constructed in this paper provide new solutions and decision supports for emergency material storage and national regime classification in group decision making environment respectively.