Quantitative decision making in reverse logistics with a hybrid decision support system integrating agentic AI and evolutionary optimization

This research presents an innovative framework based on autonomous agents (Agentic AI) for multi-objective optimization, which is designed to improve accuracy, reduce execution time, and increase stability in the face of dynamic changes. The proposed model uses an adaptive learning mechanism to dynamically adjust parameters, integrates data-driven prediction models to guide the search, and a scalability evaluation module to analyze performance in different dimensions of the problem. Experimental results show that this approach, compared to traditional methods, significantly reduces the computation time, improves the quality of solutions, and has higher resistance to uncertainties. Also, examining the relationship between accuracy and execution time showed that intelligent agents are able to change this balance in favor of efficiency. This framework has broad applicability in areas such as supply chain management, production planning, and energy optimization.