REAL-TIME CYBER-PHYSICAL DEPLOYMENT AND VALIDATION OF H-DEABSF: MODEL PREDICTIVE CONTROL, AND DIGITAL-TWIN–DRIVEN PROCESS CONTROL IN SMART FACTORIES

Abstract

The evolution of intelligent manufacturing has necessitated the development of integrated simulation and control frameworks capable of synchronizing physical operations with digital decision-making environments. This study introduces and empirically validates a Hybrid Discrete-Event and Agent-Based Simulation Framework (H-DEABSF) augmented by Model Predictive Control (MPC) and Digital Twin (DT) technologies for real-time cyber-physical process control in smart factory environments. The research addresses the limitations of single-paradigm simulation models by establishing a hybrid architecture that unifies the event-driven precision of Discrete-Event Simulation (DES) with the autonomous, adaptive decision-making capabilities of Agent-Based Simulation (ABS). Through the incorporation of MPC and real-time DT synchronization, the framework achieves continuous bidirectional communication between physical equipment and virtual models, enabling predictive decision support and dynamic reconfiguration under stochastic production conditions. A quantitative experimental design was employed using a cyber-physical testbed comprising interconnected programmable logic controllers, IIoT-enabled sensors, and a virtual simulation layer that replicates factory operations. Empirical data were collected across twelve operational trials under varying workload intensities and analyzed using descriptive, inferential, and multivariate statistical methods including ANOVA, regression, MANOVA, and correlation modeling. The hybrid configuration achieved a 22.8% increase in throughput efficiency, a 39% reduction in response latency, and a 96.2% predictive accuracy rate, outperforming traditional DES-only and ABS-only control architectures. Furthermore, fault recovery time decreased by 53%, while overall machine utilization improved by 11.7%, coupled with a 15.8% reduction in energy consumption, demonstrating the hybrid system’s efficiency and sustainability advantages. The integration of predictive control and digital twin feedback enhanced both operational adaptability and stability, ensuring robust performance across variable manufacturing conditions. The results substantiate that H-DEABSF constitutes a validated and scalable architecture for intelligent process optimization, fusing simulation modeling, predictive analytics, and cyber-physical synchronization into a single self-regulating control ecosystem. This research contributes a significant advancement to the domain of smart manufacturing by providing empirical evidence of how hybrid simulation frameworks can operationalize the core principles of Industry 4.0, promoting data-driven autonomy, resilience, and sustainable production intelligence in next-generation industrial systems.