Trilevel Multi-criteria Decision Analysis Based on Three-Way Decision

Abstract

Based on the principles of three-way decision as thinking in threes, this paper proposes a framework of trilevel multi-criteria decision analysis (3LMCDA). The main motivations and contributions are a new understanding of multi-criteria decision analysis at three levels: 1) single-criterion-level analysis with respect to individual criteria, 2) multi-criteria-level aggregation by pooling results from many criteria, and 3) multi-methods-level ensemble of different multi-criteria decision analysis methods. A full understanding of multi-criteria decision analysis requires understandings at the three levels. While the existing studies have been focused on the first two levels, new efforts at the third level may enable us to take advantages of many different multi-criteria decision analysis methods.