A novel approach to method selection in multi-criteria decision-making

MCDM (Multi-Criteria Decision-Making) involves the systematic process of ranking alternative options based on multiple criteria, a task frequently encountered across diverse fields. Within MCDM operations, the choice of both the ranking method for alternatives (MCDM method) and the weighting method for criteria significantly impacts the final ranking outcomes. This study introduces an entirely novel approach for identifying the most suitable pair combining an alternative ranking method with a criterion weighting method. The investigating three prominent alternative ranking methods: TOPSIS (Technique for Order Preference by Similarity to Ideal Solution), MOORA (Multiobjective Optimization On the basis of Ratio Analysis), and RAM (Root Assessment Method). For criterion weighting, three methods were considered: RS (Rank Sum), RR (Rank Reciprocal), and ROC (Rank Order Centroid). To facilitate comprehensive comparison, a Taguchi experimental matrix was designed, yielding nine distinct experimental combinations, each pairing one ranking method with one weighting method. This resulted in the comparison of nine specific combinations: TOPSIS-RS, TOPSIS-RR, TOPSIS-ROC, MOORA-RS, MOORA-RR, MOORA-ROC, RAM-RS, RAM-RR, and RAM-ROC. The DEAR (Data Envelopment Analysis-based Ranking) method was employed to compare these combined pairs across four distinct examples. Two primary criteria were used to determine the most suitable combination: the average Spearman's rank correlation coefficient between each pair and the ratio of alternative scores for each pair. The use of the DEAR method to compare the combinations between each MCDM method and each weighting method is first introduced in this study. This simple approach allows for the comparison of multiple combinations with minimal computation, helping to determine the best combination of an MCDM and a weighting method for the criteria. The findings indicate that the RAM-RS and MOORA-RS combinations demonstrated superior performance compared to the others, while TOPSIS-RR and TOPSIS-ROC proved to be the least effective. Therefore, RAM-RS and MOORA-RS are two recommended combinations for use in MCDM problems