Improved consensus evolution networks for supporting large-scale group decision making

Recently, studies focusing on large-scale group decision-making (LSGDM) have gained momentum. Unlike conventional group decision-making, decision-makers (DMs) in LSGDM tend to have different knowledge backgrounds, social statuses, and decision-making habits, causing the preferences of DMs to be relatively scattered and presenting challenges to consensus-reaching. To address these challenges, a growing body of research is dedicated to formulating consensus methods that are based on diverse association networks, thereby facilitating interactions among decision-makers. In this context, consensus evolution networks (CENs) have emerged as effective tools for delineating the consensus relationships among DMs and for supporting the consensus reaching process (CRP). Thus, we propose improved CENs to support LSGDM. First, we construct CENs with different consensus thresholds and conduct community detection to obtain a series of possible partitions, from which we select the best by applying four rules. Second, a novel consensus measurement and corresponding feedback mechanism are proposed to improve consensus within and among subgroups. The adjustment suggestions in each round are generated considering the DMs’ connection strength in the consensus network. After the CRP is finished, a selection process based on improved maximum consensus sequence mining (MCSM) is used to derive the priority order of alternatives. Finally, a numerical example is provided to illustrate the applicability and effectiveness of our method.