Cooperative Resource Allocation for Multi-BS Integrated Sensing, Communication and Computing in Power Internet of Things

For the Power Internet of Things (PIoT) system, terminal equipment (TE) with integrated sensing, communication, and computing (ISCC) capabilities faces constrained resources and requires task offloading to mobile edge computing (MEC) servers at the edge layer. As the demands for sensing data processing and distributed computation escalate, a standalone base station (BS) equipped with MEC server (MECS) confronts queue congestion and energy inefficiency due to resource contention. To overcome the mentioned difficulties, a multiBS cooperative structure for PIoT which dynamically allocates resources based on real-time requirements is proposed. A long-term delay minimization problem is formulated under resource constraints and transformed into a tractable per-slot framework by applying Lyapunov optimization. We provide an algorithm founded on MultiAgent Deep Deterministic Policy Gradient (MADDPG) that jointly optimizes queue stability and energy efficiency by effectually managing ISCC workloads through distributed resource allocation. The suggested algorithm performs better than baseline methods with regard to delay as well as energy efficiency, according to simulations.