IoT-Enabled Model Predictive Control for Perishable Goods Production With Dynamic Spoilage Mitigation

The rapid spoilage of perishable goods poses a major challenge to production and inventory management systems, where inaccurate state information often leads to suboptimal decisions and significant economic losses. This paper proposes an IoT-enabled Model Predictive Control (MPC) framework that integrates real-time sensor data with dynamic optimization and Predictive Technology Interventions (PTIs) to reduce spoilage and operational costs. A layered IoT architecture, comprising temperature, humidity, and RFID sensors, wireless communication, edge/cloud processing, and digital twins, enables continuous state estimation through Kalman filtering. These real-time updates allow the MPC controller to adapt production, storage, and spoilage-mitigation interventions dynamically, improving responsiveness to environmental fluctuations and uncertainty. Through comparative simulations, we analyze three scenarios: baseline MPC, MPC with raw IoT overwrites, and MPC with IoT-driven state estimation. Monte Carlo experiments under varying sensor noise levels demonstrate that IoT feedback combined with PTIs significantly reduces spoilage rates and improves profitability compared to conventional MPC. The results highlight the potential of IoT-enabled MPC with PTIs to enhance sustainability, efficiency, and resilience in perishable supply chains.