A consensus degree based multiple attribute group decision making method

Consensus building is a hot but difficult problem in the current research area of group decision making. If the group decision algorithm is not well designed, it will lead to the experts' opinions compromise to each other, and the final result will be the experts' mean opinions. It may be kind of safe, but too conservative to receive the few but creative and right opinions. In this paper, a group decision making approach based on PAM (Partitioning Around Medoids) and Particle swarm optimization was proposed. With this approach, we can promote the effective interactions among experts and deepen their task understandings by analyzing their opinions' conflicts, thereby, accelerating the convergence of group opinions, improving the efficiency of group decision making and the reliability of the results. In some cases the experts cannot reach consensus independently, but we can solve this problem with this approach by searching the optimal solution within the acceptable range given by expert groups, fine-tuning the original evaluation matrix, thus improving the success rate of group decision making.