文章摘要
刘洪,李吉峰,葛少云,等.基于多主体博弈与强化学习的并网型综合能源微网协调调度[J].电力系统自动化,2019,43(1):40-48. DOI: 10.7500/AEPS20180627008.
LIU Hong,LI Jifeng,GE Shaoyun, et al.Coordinated Scheduling of Grid-connected Integrated Energy Microgrid Based on Multi-agent Game and Reinforcement Learning[J].Automation of Electric Power Systems,2019,43(1):40-48. DOI: 10.7500/AEPS20180627008.
基于多主体博弈与强化学习的并网型综合能源微网协调调度
Coordinated Scheduling of Grid-connected Integrated Energy Microgrid Based on Multi-agent Game and Reinforcement Learning
DOI:10.7500/AEPS20180627008
关键词: 综合能源微网  协调调度  多智能体  博弈理论  Q学习
KeyWords: integrated energy microgrid  coordinated scheduling  multi-agent  game theory  Q learning
上网日期:2018-11-28
基金项目:国家重点研发计划资助项目(2017YFB0903400,2017YFB0903401);国家自然科学基金资助项目(51777133)
作者单位E-mail
刘洪 智能电网教育部重点实验室(天津大学), 天津市300072  
李吉峰 智能电网教育部重点实验室(天津大学), 天津市300072  
葛少云 智能电网教育部重点实验室(天津大学), 天津市300072  
张鹏 智能电网教育部重点实验室(天津大学), 天津市300072 zhangpeng1984@tju.edu.cn 
陈星屹 智能电网教育部重点实验室(天津大学), 天津市300072  
摘要:
      针对传统集中式优化调度方法难以全面反映综合能源微网内不同智能体的利益诉求,以及人工智能技术在综合能源调度方面的应用亟待进一步挖掘等问题,提出了基于多主体博弈与强化学习的并网型综合能源微网协调调度模型和方法。首先,针对并网型综合能源微网中横向电气热冷各子系统及纵向源网荷储等各环节的不同投资与运营主体,开展了多智能体划分;其次,针对可再生能源服务商、微网系统能源服务商、电动汽车用户等智能体,分别构建了各自的决策模型,并建立了以多智能体间利益均衡为目标的联合博弈决策模型;再次,针对多主体博弈这一高维决策难题,引入人工智能求解方法,提出了基于Nash博弈和强化学习算法的综合能源微网协调调度方法;最后,通过实例验证了所提模型和方法的有效性与实用性。
Abstract:
      Considering that the traditional centralized optimized scheduling methods cannot comprehensively reflect the interests of different agents in the integrated energy microgrid and the application of artificial intelligence techniques in integrated energy scheduling is deepened, the coordinated scheduling model and method for grid-connected integrated energy microgrid based on multi-agent game and reinforcement learning are proposed in this paper. Firstly, the multiple investment and operation agents are divided from the perspectives of electrical/heating/cooling subsystems and source/grid/load/storage links, respectively. Secondly, the decision-making models for renewable energy provider, microgrid energy provider and electric vehicle owner are constructed, and the joint game based decision-making model is established with the objective of balancing the multi-agent interests. Thirdly, the artificial intelligence techniques are introduced to solve the highly-dimensional multi-agent game decision-making problem, and the coordinated scheduling method for integrated energy microgrid based on Nash game and Q learning algorithm is proposed. Finally, the effectiveness and applicability of the model and method proposed are verified by case study.
查看全文(Free!)   查看附录   查看/发表评论  下载PDF阅读器