Multi-domain Resources Scheduling in Edge Computing Power Network for IIoT

The Industrial Internet of Things (IIoT) imposes strict requirements on task processing delays. To ensure low-latency services, the edge Computing Power Network (CPN) integrates various multi-domain resources, including computing, storage and communication, enabling fast processing and responsiveness. However, the performance of the CPN is significantly influenced by the scheduling scheme for these multi-domain resources. To explore an optimal scheduling scheme, this paper proposes a mathematical optimization model for the resource scheduling problem. The problem is then transformed into a Markov Decision Process (MDP) and the Discrete Probabilistic Deep Deterministic Policy Gradient (DP-DDPG) algorithm, which modifies the actor network of the original DDPG to output a discrete probability vector, is introduced to obtain the optimal scheme. Experimental results demonstrate that the proposed approach reduces latency within the constraints of limited multi-domain resources, validating its feasibility and effectiveness in IIoT scenarios.