Ensemble multi-attribute decision-making for material selection problems

• Published in: Soft computing (Berlin, Germany) Vol. 28; no. 6; pp. 5437 - 5460
• Main Author: Şahin, Mehmet
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01-03-2024; Springer Nature B.V
• Subjects: Application of Soft Computing; Artificial Intelligence; Computational Intelligence; Control; Correlation coefficients; Corrosion resistance; Criteria; Decision making; Engineering; Literature reviews; Materials selection; Mathematical Logic and Foundations; Mechatronics; Methods; Product design; Production costs; Ranking; Robotics; Sustainable materials; Weighting methods; Material selection; Hybrid approach; Group decision-making; Comparative analysis; Copeland
• ISSN: 1432-7643; 1433-7479
• DOI: 10.1007/s00500-023-09296-1

Material selection is influential in product design, manufacturing, and marketing. Appropriate material selection maximizes the performance of a product while minimizing its cost, whereas inappropriate material selection creates devastating results such as low performance, low quality, and high cost. Therefore, it is crucial how to choose the most suitable material. Unlike other studies, this study presents an ensemble multi-attribute decision-making approach for material selection. The approach involves four weighting methods—criteria importance through intercriteria correlation, Entropy, the method based on the removal effects of criteria, and statistical variance, five ranking methods—additive ratio assessment, combined compromise solution, multi-attributive border approximation area comparison, range of value, and the technique for order performance by similarity to the ideal solution, Spearman's correlation coefficients, and the Copeland method. Three different problems are considered to show the applicability of the proposed method and to reveal a comprehensive analysis. The results of each problem show valuable implications. The results of the ranking methods are sensitive to attribute weights. No ranking method alone can assure dependable selection for a given problem. Overall, the results reveal the importance of using multiple weighting and ranking methods and the superiority of the proposed integrated approach.