A new normalization-free method for ranking and clustering multiple alternatives

What is it about?

Most multiple criteria decision analysis methods hardly guide decision-makers beyond supplier evaluation (by providing them with a ranking), and supplier selection and classification is effectively the decision-makers' job. The relative difference between the performance of different suppliers is usually overlooked even though this fact is functionally important in real-life multi-sourcing decisions. The current study proposed a system that not only evaluates suppliers through the Dynamic Grey Relational Analysis (Dynamic GRA), a generalized form of Deng's GRA, but also helps the decision-makers to distinguish the most reliable suppliers from the least reliable ones using a novel Rank Product Score (RPS), which allows the decision-makers to visualize the performance of each supplier on an exponential curve. Unlike the benchmark models, the normalization of input data is not mandatory in the proposed method, and uncertainty in input can be quantified even when it is not explicitly visible in the input. The new framework enables the procurement managers looking for strategic global multi-sourcing to select the suppliers from the top cluster on this curve. It is tested on a practical case from a big engineering organization with diverse suppliers, demonstrating its validity. The comparative analyses on three dimensions confirm its effectiveness and efficiency. It is of immediate practical use for building intelligent decision support systems in the purchasing environment.

Featured Image

Photo by Gilly on Unsplash

Photo by Gilly on Unsplash

Why is it important?

A novel method of rank-based clustering is proposed. Generalized form of an influential multiple criteria decision making model (GRA) is reported.

Perspectives