Robust Consensus: A New Measure for Multicriteria Robust Group Decision Making Problems Using Evolutionary Approach

Abstract

AbstractIn fuzzy group decision making problems, we often use multiobjective evolutionary optimization. The optimizers search through the whole search space and provide a set of nondominated solutions. But, sometimes the decision makers express their prior preferences using fuzzy numbers. In this case, the optimizers search in the preferred soft region and provide solutions with higher consensus. If perturbation in the decision variable space is unavoidable, we also need to search for robust solutions. Again, this perturbation affects the degree of consensus of the solutions. This leads to search for solutions those are robust to their degree of consensus. In this work, we address these issues by redefining consensus and proposing a new measure called robust consensus. We also provide a reformulation mechanism for multiobjective optimization problems. Our experimental results show that the proposed method is capable of finding robust solutions having robust consensus in the specified soft region.KeywordsConsensusevolutionary algorithmsfuzzy group decision makingmultiobjective optimizationrobustness