Enhancing PROMETHEE-Based Decision Support Systems Through Evolutionary Weight Optimization

Abstract

Multi-Criteria Decision Making (MCDM) is a crucial issue in various strategic domains, including the energy and utilities sectors, involving multiple alternatives and criteria with varying degrees of importance. One widely used MCDM method is PROMETHEE, but its performance relies heavily on the determination of criteria weights, which are generally subjective and static. This study aims to improve the performance of a PROMETHEE-based decision support system through optimization of criteria weights. The research dataset is sourced from InfluxDB for Energy and Utilities and consists of 219 alternatives with a number of operational parameters represented in the form of a decision matrix. The proposed method integrates PROMETHEE with a weight optimization mechanism, where weights are treated as decision variables and PROMETHEE is used as an evaluation function in the optimization process. The performance of the proposed method is compared with the baseline PROMETHEE using indicators of net flow value, alternative ranking, ranking changes, and net flow variance. Experimental results show that weight optimization significantly improves the system's discrimination power, as indicated by an increase in net flow variance from 0.189 in the baseline PROMETHEE to 1.118 in the optimized PROMETHEE. Furthermore, weight optimization results in a more structured and adaptive ranking to data characteristics, without disrupting the stability of extreme alternatives. These findings confirm that criterion weight optimization is capable of improving the objectivity, reliability, and quality of decisions in a PROMETHEE-based decision support system, thus the proposed method has the potential to be applied to various complex multi-criteria decision-making problems. From the perspective of Informatics and Computer Science, this research contributes to the development of intelligent decision support systems by integrating computational optimization techniques with MCDM methods to support more adaptive and data-driven decision-making processes.