Intelligent Collaboration Mechanism on Computing, Communication and Control for Industrial Networked Automation System

5G delivers ultra-reliable low-latency communications and mobile edge computing (MEC) for industrial applications, enabling distributed computing resource collaboration and promoting traditional automation systems toward networked control paradigms. However, industrial networked automation systems face significant challenges in optimally allocating limited computing and communication resources to meet strict QoS requirements for massive control tasks while maintaining system stability and efficiency. Thus, we propose a computing-communication-control collaboration mechanism to enhance coordination between MEC and local controllers, enabling complex control tasks to be processed despite limited local computing resources. To jointly optimize hybrid control tasks migration with communication constraints and computing resources allocation with incomplete information, we design a dual double deep Q-network embedded with the Stackelberg game model. Simulation results demonstrate that the proposed mechanism achieves better performance compared with other benchmarks, and extremely decreases the convergence time compared with the classical Stackelberg game solution.