Sensing-Communication-Computing Resource Interaction Method for Network Hierarchical Dispatch in Power Wireless Network

With the increasing penetration of distributed renewable energy and flexible loads, distribution network dispatch is evolving toward multi-level collaborative control. Power wireless networks, serving as the link between edge gateways and massive devices, require integrated and dynamic coordination of sensing, communication, and computing resources to support millisecond-level fault isolation and differentiated quality of services (QoS) requirements. However, existing approaches lack effective cross-layer interaction optimization. This paper proposes a sensing-communication-computing resource interaction method for hierarchical dispatch in power wireless networks. A three-layer architecture spanning device, edge, and cloud is constructed. To minimize end-to-end queuing delay under long-term constraints of energy consumption, data acquisition, and access rates, the problem is formulated and decomposed via Lyapunov optimization into three stages. First, opportunity-balanced upper confidence bound (UCB) is employed for device access, ensuring fairness by incorporating access rate awareness into utility evaluation. Second, a dual-constraint foresight twin delayed deep deterministic policy gradient (TD3) algorithm optimizes time slicing by embedding energy and data constraints into Q-value estimation. Third, convex optimization is used for computing resource allocation. Simulation results validate that the proposed algorithm significantly improves performance in delay reduction, energy efficiency, data acquisition, and access fairness, effectively enhancing resource utilization under multiple QoS constraints in distribution network dispatch.