Self-Optimization of Industrial Technological Processes Based on Digital Twin and Edge AI: From RealTime Monitoring to Predictive Control

Abstract

The convergence of Digital Twin (DT) technology and Edge Artificial Intelligence (Edge AI) is transforming industrial automation from reactive monitoring into predictive, selfoptimizing control. This paper proposes an integrated framework in which a real-time digital twin continuously mirrors the physical process while an edge-deployed AI agent executes adaptive optimization locally. The model is built upon state-space representation, reinforcement-learning-based parameter tuning, and predictive fault diagnostics. MATLAB simulations demonstrate that coupling DT with Edge AI reduces steady-state error by more than 70 %, improves response speed by 2.4 times, and decreases energy consumption by approximately 25 %. Real-world evidence from process industries confirms the feasibility of this hybrid approach, marking a critical step toward resilient, autonomous manufacturing systems under the paradigm of Industry 5.0