Social network analysis-based conflict relationship investigation and conflict degree-based consensus reaching process for large scale decision making using sparse representation

Large-Scale Decision Making (LSDM) scenarios, such as public participation events, are becoming increasingly common in human life. Decision makers (DMs) in LSDM events present different interest preferences, leading to different relationships being created between them. In LSDM scenarios, a conflict relationship, which is a type of negative relationship among DMs, has the biggest negative impact on reaching the consensus. The conflict relationships can be divided into two parts: the opinion conflict and the behavior conflict. In this paper, a Social network analysis-based Conflict Relationship Investigation Process (S-CRIP) is presented to detect the conflict relationships among DMs for LSDM events, in which sparse representation is used. Besides, a Conflict Degree-based Consensus Reaching Process (CD-CRP) is proposed for LSDM problems, which is using group conflict degree to check whether the consensus is reached or not. In the decision selection process, DMs' weights are calculated by their conflict performances, which can reduce the negative influence of those DMs that present conflict in the LSDM event. The proposed S-CRIP can not only investigate the conflict relationships among DMs, but can also recognize the two types of conflict relationships according to their features. The three processes constitute the S-CRIP and CD-CRIP-based LSDM model, which is suitable for any numerical representations. Illustrative experiments not only show the feasibility and veracity of S-CRIP in LSDM scenarios, but also prove the practicability and effectiveness of S-CRIP and CD-CRP-based LSDM model.