The Task Scheduling Algorithm for Fog Computing in Intelligent Production Lines Based on DQN

The intelligent production line is the primary carrier in the intelligent manufacturing process, and ensuring fast and reliable processing of task requests is critical to realizing intelligent manufacturing. Fog computing, which is the extension of cloud computing in edge devices, has become an inevitable choice for the efficient processing of intelligent production line tasks with its distributed, low latency and high reliability features that respond to task requests from end devices quickly and reliably. However, task requests in intelligent production lines usually have high time sensitivity, so how to dynamically schedule these tasks to the most suitable fog nodes in resource-constrained fog nodes to reduce task response time and improve task completion rate has become a research hotspot for fog computing in intelligent production lines. To solve the above problems, we propose a DQN-based dynamic task scheduling algorithm to realize online scheduling in intelligent production line fog computing tasks. Finally, the proposed method is demonstrated by using the intelligent production line experimental verification platform, and the proposed algorithm has significant advantages over other algorithms in both latency and task completion rate in multiple task scenarios.