A Novel Approach to Weighting Criteria Based on Rank Stability

Abstract

In Multi-Criteria Decision Making (MCDM), the weighting of the criteria is at least as important as the normalization of the decision matrix and the procedure for ranking the alternatives. The slightest uncertainty in the weights of criteria poses a major problem for decision-makers, especially when there is little difference between the assessments of alternatives. One crucial question is which approaches should be applied to a given MCDM problem? This question may be answered by considering the insensitivity of rankings of the alternatives to changes in the weights of criteria. We propose a novel approach to criteria weighting that focuses solely on the insensitivity of rankings to changes in the weights of criteria. Simulations reveal that this approach is superior to other existing methods in terms of ranking stability. A real-life example concerning the digital readiness of the European Union reveals that this new approach yields beneficial results in different aspects of decision-making.