The rapid development of industrial Internet has given rise to a large number of real-time computing demands, and traditional cloud computing cannot meet the low-latency requirements, while the cloud-edge cooperative architecture integrates the advantages of high computing power in the cloud and low latency at the edge, which has become an effective solution for resource scheduling in industrial Internet. In this paper, we design an optimization model for industrial Internet resource scheduling based on cloud-edge cooperative architecture, which realizes the cooperative processing of data and computation by integrating cloud computing and edge computing technologies. Firstly, an independent task resource scheduling model is constructed, and three evaluation indexes, namely, computing time, energy consumption and throughput rate, are defined; then a path planning method based on DDPG and a multilevel resource scheduling model at the cloud-edge end are designed to realize optimal allocation of tasks among multilevel resources. Simulation results show that the proposed CMRS model tends to converge at the 310th iteration, and the convergence speed is 30 iterations faster than GA-DDPG, and the reward value at convergence reaches -12.9, which is better than GA-DDPG’s -17.1. In terms of cache hit rate, the CMRS model is significantly higher than the comparison algorithm. The performance evaluation shows that the CMRS model achieves lower total system overhead and task processing latency when dealing with tasks of different complexity and different data volumes, proving that the model is able to effectively improve resource utilization and quality of service in the industrial Internet environment.
