A Simulation-AI Integrated Approach for Production Scheduling: Modeling and Exploring Priority-Based Decision Spaces

Production scheduling in complex industrial environments faces significant challenges due to combinatorial complexity and multiple conflicting objectives. This work presents an integrated framework combining Discrete-Event Simulation (DES) with AI-based optimization for production scheduling in complex industrial environments. The proposed approach features an optimization module (MIA) that implements multiple metaheuristic algorithms (GA, SA, TS, ALNS, MCTS) integrated with a pre-existing DES system (See The Future - STF) through an offline API-based interface. A key contribution is the explicit modeling of priority-based decision spaces, enabling systematic exploration of sequencing policies through simulation. The framework was applied to a real-world-inspired flexible job shop case study with 41 orders, 21 machines, and 292 jobs. Preliminary results indicate that the automated approach identifies higher-profit solutions than manual scheduling, with Genetic Algorithms achieving a 114.24\% improvement over the initial solution. The modular architecture provides a flexible foundation for integrating intelligent optimization into industrial simulation-based scheduling systems.