A hierarchical selection algorithm for multiple attributes decision making with large-scale alternatives

We consider a multiple attributes decision making (MADM) problem in the presence of large-scale alternatives. Considering the large number of alternatives, we first try to identify if there exist some alternative sets with dominative patterns by determining a hyper-plane. If so, the superior alternatives can be easily selected. We then exploit the divide and conquer idea and develop a hierarchical MADM algorithm, which selects locally superior alternatives iteratively until the globally best alternative is reached. Specifically, we first divide the large-scale alternatives into several clusters, and determine the attribute weights at each round. We then select the locally superior alternative in each cluster. The attribute weights are updated based on the former attributes weights after each clustering, so as to remain consistent with the attributes weights throughout the hierarchical MADM algorithm. Finally, numerical experiments are conducted to demonstrate the effectiveness of the proposed method. (C) 2020 Published by Elsevier Inc.