基于多智能体深度强化学习的海上风电传感器节点能效优化

打开文本图片集

关键词：海上风电；无线传感网络；能效优化；多智能体深度强化学习；自适应噪声策略 中图分类号：TP399；TN926 文献标志码：A 文章编号：1001-3695（2025）08-032-2490-07 doi：10.19734/j.issn.1001-3695.2024.12.0520

Energy efficiency optimization of sensor nodes in offshore wind farm based on multi-agent deep reinforcement learning

JiaLinpengl，WangXiaol†，HeZhiqin¹，WuQinmul，YinYaohua² （1.ColegeofElectricalEnginerin，GuzhouUniersityGuang5O，hina；2.Guangngneeingopraoid POWERCHINA，Guiyang550081，China）

Abstract：Theeficientoperationofofshorewindfarmsreliesonmonitoringdatafrom wirelessensornetworks.Thisstudy identifiedthelackofefectiveenergyeficiencyoptimizationfornodedeploymentandcommunicationinofshorewindfarmscenarios.To address this，this paper proposed anoptimizationscheme usingamulti-agentdepdeterministic policygradient algorithm.Considering limitednodeenergyandofshorecommunicationchalenges，theschemeoptimized sensingandcommunicationstrategiesthroughmulti-agentcollaboration，reducingenergyconsumptionandimprovingnetworkcoverageanddatatrasmision.Adaptive noise，priorized experiencereplay，andatailoredreward function further enhanced learning eficiencyand energy performance.Experiments show that the proposed scheme increases node energy efciency by 26% over DDPG and achieves training speeds 33% and 48% faster than DDQN and SAC algorithms.

Key words：offshore wind farms；wirelesssensor network（WSN））；energy eficiencyoptimization；multi-agent deepreinorcement learning；adaptive noise strategy

0 引言

随着全球可再生能源需求的不断增长，海上风电场作为一种清洁、可再生的能源形式，已经成为世界能源结构转型的关键。（剩余19905字）