Distributed reinforcement learning-based memory allocation for edge-PLCs in industrial IoT

Abstract The exponential device growth in industrial Internet of things (IIoT) has a noticeable impact on the volume data generated. Edge-cloud computing cooperation been introduced to IIoT lessen computational load cloud servers and shorten processing time for data. General programmable logic controllers (PLCs), which have playing important roles control systems, start gain ability process large amount share workload servers. This transforms them into edge-PLCs. However, continuous influx multiple types concurrent production streams against limited capacity built-in memory PLCs brings huge challenge. Therefore, reasonably allocate resources edge-PLCs ensure utilization real-time become one core means improving efficiency processes. In this paper, tackle dynamic changes arrival rate over at each edge-PLC, we propose optimize allocation with Q-learning distributedly. simulation experiments verify that method can effectively reduce loss probability while system performance.