Reconfigurable Distributed Model Predictive Control for Decentralized IT/OT Systems

The rapid evolution of industrial environments, driven by the proliferation of dynamic devices such as wireless robots and smart sensors, has introduced unprecedented complexity. Navigating this landscape requires not only adapting to internal system changes but also aligning with managers' visions. This paper proposes a new approach that enables real-time reconfiguration of industrial systems. At its core is a modeling framework that captures node-level dependencies, revealing how coordination complexity can exponentially degrade system performance. By leveraging the flexibility offered by virtualization and IT/OT (Information Technology and Operational Technology) infrastructure, the proposed Distributed Model Predictive Control (DMPC) method evaluates a multitude of operational scenarios, allowing the system to adapt dynamically. The workflow culminates in the reconfiguration of Industrial Internet of Things (IIoT) nodes through WebAssembly (WASM), seamlessly bridging the gap between virtual parameters and physical reality. A mathematical process formulation, including an exponential latency model, provides a robust framework that manages complexity while ensuring responsive system behavior.