An Improved Similarity-based Fuzzy Group Decision Making Model through Preference Transformation and K-Means Clustering Algorithm

Abstract

Group decision making plays a crucial role in organizational and community contexts, facilitating the exchange of expert opinions to arrive at effective decisions. The concept of preference, reflecting an individual's subjective evaluation of criteria or alternatives, forms a foundational element in this process. This study focuses on transforming non-fuzzy preferences, such as preference ordering and utility functions, into fuzzy preference relations (FPR) to address the uncertainty and uniformity inherent in expert preferences. To further enhance decision-making, we assess and visualize the similarity among the experts' uniform preferences. Integrating the K-means clustering algorithm into the fuzzy group decision making model allows for the predetermination of an appropriate number of groups based on the available alternatives. By aggregating individual preferences, we present a final ranking of alternatives. The enhanced methodology, as demonstrated through comparative analysis, showcases its ability to yield positive benefits when applied to decision-making applications.