Abstract: For the problem of large-scale group decision-making in the social network environment where decision experts have overlapping social networks and evaluation information is in the form of interval numbers, a new social network group consensus decision-making method considering overlapping networks and interval information is proposed. Firstly, in the expression of preferences, considering the fuzzy judgment of decision experts and the uncertainty of decision problems, expert preferences are represented by interval information. Secondly, in the clustering and weight calculation process, a method for community division and weight determination considering overlapping networks is proposed. Through the LFM algorithm, precise clustering is achieved, and the influence is quantified by double centrality to determine the weights, effectively improving the scientificity and scene adaptability of clustering and weight allocation. Finally, a new consensus feedback model considering the overlapping community division of social networks is proposed. This model not only allows decision experts to refer to the suggestions of "bridging experts" with strong information advantages, but also allows decision experts to refer to the suggestions of decision experts with similar opinions. Through this multi-dimensional adjustment, the decision dynamics of the expert group are optimized, thereby more effectively promoting the formation of consensus. Numerical examples and comparative analysis further demonstrate the feasibility and effectiveness of the model.

Key words: Social network, consensus reaching, interval number, overlapping communities, large groups