Sensitivity of RIM, SPOTIS, and COMET methods under criteria removal process

Multi-Criteria Decision Analysis (MCDA) methods are widely applied in complex decision-making scenarios involving multiple, often conflicting criteria. While considerable research has focused on their performance under varying numbers of alternatives, the effect of changes in the number and importance of criteria, particularly in rank reversal-resistant methods, remains underexplored. This study investigates the stability of three rank reversal-resistant MCDA methods, namely the Reference Ideal Method (RIM), Stable Preference Ordering Towards Ideal Solution (SPOTIS), and the Characteristic Objects Method (COMET), under systematic changes in criteria inclusion. Through a series of 1,000 simulations with varying numbers of alternatives and criteria, the methods were tested for ranking consistency as criteria were removed in both ascending and descending order of importance. Results show that COMET was the most stable in high-dimensional problems, RIM performed best in mid-sized scenarios, and SPOTIS was more robust when more important criteria were removed. The findings provide valuable insights for selecting appropriate MCDA methods in dynamic decision environments and fulfill the study’s goal of assessing method sensitivity to criteria variation.