Akademik Veri Yönetim Sistemi
Applied Soft Computing, cilt.167, 2024 (SCI-Expanded)
Ensemble multi-criteria decision-making (ensemble MCDM) aggregates results from various MCDM techniques into a cohesive ranking, aiming to mitigate individual technique weaknesses and provide a robust approach to MCDM challenges. However, literature lacks in-depth analysis and a comprehensive guide defining ensemble MCDM. Therefore, this study aims to clarify ensemble MCDM terminology, conduct a systematic literature review, and outline a roadmap for future research in this domain. A systematic review protocol with clear inclusion and exclusion criteria was implemented, selecting relevant articles from IEEE, ScienceDirect, Web of Science, and Google Scholar. Out of 730 collected articles, only 10 met the criteria for ensemble MCDM techniques. Four main data extraction techniques were employed: defining ensemble MCDM terminology, mapping the literature, identifying ensemble decision-making techniques, and exploring ensemble levels within the MCDM framework. From the review, three primary ensemble MCDM techniques emerged: statistical means, voting technique, and Half-Quadratic programming. These techniques were not only described theoretically but also implemented and demonstrated through a numerical example, supported by provided code. Furthermore, the study mathematically describes several other ensemble techniques for potential future applications. The implications of this research establish a foundational understanding of ensemble MCDM and propose a roadmap for future investigations. By addressing individual MCDM technique weaknesses and offering a reliable aggregation method, this research serves as a fundamental reference for future studies. The accompanying codes, tutorials, and mathematical illustrations aim to advance knowledge and facilitate practical applications in the field of ensemble MCDM