A Look at Another Approach to Aggregation in the Analytical Hierarchy Process

Abstract

The Analytic Hierarchy Process (AHP), is a multi-criteria group decision making (MCGDM) methodology. It generally required aggregating the Individual Judgments (IJ) or Individual Priorities (IP) of a group of experts. Until now, in the scientific community has accepted the Geometric Mean (GM) and the Arithmetic Mean (AM) as aggregation variables. In this paper we offer the statistical Mode as a possible aggregation variable. This could simplify the process of finding the aggregated priorities/weights in the AHP process, particularly in the context of large, homogeneous groups of experts. We argue that the mode offers a viable, efficient, and representative solution under specific distributional constraints. Our analysis includes both empirical validations based on our long-term observations in various applications of AHP and theoretical, mathematical justification. It is complemented by using simulated datasets based on a controlled range around the Mode. Since we have not yet encountered AI that applies Mode as an aggregation variable, our subsequent efforts will be aimed at training an AI model and/or agent that works with Mode as an aggregation variable. The ultimate goal is for the agent to acquire not only the ability to apply AHP with the Mode as an aggregation variable, but also to analyze which is the most appropriate aggregation approach depending on the collected, experts output data.