Large-scale analytic hierarchy process method based on fuzzy-rough-advantage relation

In the big data era, the complexity of decision-making problems on a large scale has risen, particularly when confronted with large-scale alternatives (objects). Traditional multi-attribute decision-making methods, however, are becoming less effective in handling these problems. To address this, the present study proposes a Large-scale Analytic Hierarchy Process (LAHP) approach predicated on fuzzy-rough-advantage relations. Initially, an unsupervised learning technique, known as K-means++, is applied to cluster the alternatives (objects). Subsequently, the fuzzy-rough-advantage relationship is utilized to form the judgment matrix for these clusters. Following this, the Analytic Hierarchy Process (AHP) principles are employed to determine the weights of each cluster based on various evaluation criteria. These weights are then utilized to compute the overall scores of each cluster and identify the optimal group. Finally, the applicability of the LAHP method is confirmed through case analysis, with the ranking outcomes being subsequently analyzed and discussed.