Robust optimization of best-worst multi-criteria decision-making method

Abstract

The Best-worst multi-criteria decision-making method can determine optimal weight value of each criteria. It uses two vectors for pairwise comparisons in multi-criteria decision-making problem. This paper improves the original method from the perspective of robust optimization. Four robust counterpart constraints instead of two linear constraints in original optimization model are proposed. The decision-making problem can divide into full consistent and non-full consistent problems by classifying parameter value. We can achieve a unique set of interval solution in full consistent decision-making problem. Non-full consistent problem can result in multiple sets of optimal interval solution. The result which we get from this method is more effective than the original method. Each criterion can achieve optimal weight interval value. Then we take quantile value of each interval as optimal weight. This is effectively illustrated in the numerical test at the end of the paper.