HYBRID DISCRETE-EVENT AND AGENT-BASED SIMULATION FRAMEWORK (H-DEABSF) FOR DYNAMIC PROCESS CONTROL IN SMART FACTORIES

Abstract

This paper introduces a Hybrid Discrete-Event and Agent-Based Simulation Framework (H-DEABSF) designed for dynamic process control in smart factories. The framework integrates the advantages of discrete-event simulation (DES)—noted for its efficiency in modeling system flows and queues—with the adaptability and autonomy of agent-based simulation (ABS), which captures decentralized decision-making and interactions among heterogeneous entities. By combining these paradigms, the H-DEABSF addresses the limitations of each when applied in isolation, enabling both macroscopic process optimization and microscopic behavior modeling. The proposed framework is specifically developed to support the evolving needs of Industry 4.0, where factories must continuously adapt to fluctuating demand, real-time disruptions, and resource constraints. It leverages smart sensors, IoT-enabled devices, and cyber-physical systems to feed real-time data into the hybrid model, ensuring that simulations reflect operational realities. Through dynamic control loops, the H-DEABSF facilitates adaptive scheduling, predictive maintenance, and production-line reconfiguration, thereby enhancing responsiveness and resilience. A case study in a digital twin–enabled smart factory environment demonstrates the applicability of H-DEABSF for dynamic production scheduling under stochastic conditions. The results show improvements in system throughput, reduction of idle time, and optimized allocation of resources when compared to conventional single-method simulation models. Furthermore, the integration of human operators as autonomous agents highlights the framework’s ability to capture socio-technical interactions critical in real-world factory operations. This research contributes to the field of smart manufacturing by providing a comprehensive simulation framework that enhances real-time decision-making and supports sustainable operational strategies. It also offers an extensible platform for future integration with machine learning algorithms, enabling data-driven decision support for next-generation intelligent factories.