Decision-Making Modeling in Engineering Problems Using Machine Learning and Data Envelopment Analysis

Decision-making plays a critical role in engineering systems, particularly in complex environments characterized by multiple inputs and outputs, uncertainty, and rapidly growing data volumes. Traditional analytical and optimization-based approaches often rely on simplifying assumptions that limit their applicability in real-world engineering problems. In response, data-driven methods have gained increasing attention as effective tools for supporting engineering decision-making. Data Envelopment Analysis (DEA) is a well-established non-parametric technique for performance evaluation and efficiency analysis of multi-input–multi-output systems. DEA enables benchmarking and relative efficiency assessment without requiring explicit functional relationships between inputs and outputs, making it particularly suitable for engineering applications. In parallel, Machine Learning (ML) techniques have emerged as powerful tools for predictive modeling, pattern recognition, and handling nonlinear and high-dimensional data. However, while ML models provide strong predictive performance, they often lack interpretability and benchmarking capabilities. In recent years, hybrid DEA–ML approaches have been proposed to integrate the strengths of both methodologies. These hybrid models combine the efficiency evaluation and interpretability of DEA with the predictive and learning capabilities of ML, offering enhanced decision support for complex engineering systems. This paper presents a comprehensive review of DEA, ML, and hybrid DEA–ML models applied to engineering decision-making. Classical and advanced DEA models, major ML techniques, and different integration strategies are systematically reviewed and discussed. Furthermore, key challenges and future research perspectives related to hybrid DEA–ML models are identified, including standardization, scalability, data uncertainty, dynamic system modeling, and the trade-off between interpretability and predictive accuracy. The findings of this review highlight the potential of hybrid DEA–ML frameworks as effective and robust tools for data-driven decision-making in engineering sciences and provide valuable insights for future research and practical applications.

کلیدواژه ها:

Decision-Making ، Engineering Sciences ، Data Envelopment Analysis ، Machine Learning (ML) ، Performance Evaluation ، Hybrid Models