A Hierarchical Actor‐Critic Reinforcement Learning‐Based Method for Dynamic Coverage Optimization in Distribution Network Communications
Year: 2025
Type: article
Abstract: ABSTRACT As the energy internet and next‐generation power systems evolve, distribution network communications face significant challenges due to the integration of distributed energy and the need for millisecond‐level load control. This paper introduces a dynamic network coverage optimization algorithm based on Hierarchical Actor‐Critic (HAC) reinforcement learning to address issues like inadequate signal coverage and poor adaptability in changing environments. The algorithm features a collaborative framework with a global policy layer and a local execution layer, utilizing a hybrid action space and a multi‐objective reward function to enhance signal coverage. Experiments using real‐world cellular data from a metropolitan area show that this method outperforms traditional approaches, achieving a 14.59% improvement in coverage performance, along with better stability against dynamic interference and increased energy efficiency. This work presents an effective solution for optimizing communications in modern distribution networks.
Source: Engineering Reports
DOI: 10.1002/eng2.70522
Authors: Yulong Liu, Yuanchao Zhou, Zhouzhou Wu, Yi Li
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $2,270
OpenAlex ID: W4416931729