Balancing different objectives in multi-objective optimization consensus for social network large-scale group decision-making with unmanned aerial vehicle review filtering

For social network large-scale group decision-making (SN-LSGDM) with unmanned aerial vehicle (UAV) purchase, online reviews vary greatly in quality, making it difficult to ensure validity. Meanwhile, during the consensus-reaching process (CRP), it is often challenging to achieve a balance among decision-makers’ (DMs) satisfaction, resource investment, and the consensus degree. Therefore, a review filtering mechanism and a consensus model balancing different objectives are devised. First, a sentiment analysis (SA) model integrating the filtering of high-quality reviews is developed, which quantifies them into probabilistic linguistic term sets (PLTSs) and extends them to PL-incomplete information systems (PL-IISs). Next, a trust propagation mechanism integrating trust decay with PageRank is proposed. Moreover, a multi-neighbor-based K-nearest neighbor (KNN) strategy is explored to fill the PL-IISs. Then, the Infomap algorithm is utilized to cluster a comprehensive indicator that combines similarity with the trust matrix. Subsequently, a dual-path consensus model is established, combining the opinion evolution process based on the DeGroot and Hegselmann–Krause (HK) models with a multi-objective optimization process. This process dynamically adjusts DMs’ preferences based on personality traits while balancing satisfaction, cost, and consensus. Finally, experimental results show that the model effectively improves consensus efficiency and information reliability.