| 引用本文: | 李 刚,刘继春,魏震波,等.含分布式电源接入的市场多主体博弈分析[J].电力系统保护与控制,2016,44(19):1-9.[点击复制] |
| LI Gang,LIU Jichun,WEI Zhenbo,et al.Analysis of game among multi-agents in electrical power market with integration of distributed generation[J].Power System Protection and Control,2016,44(19):1-9[点击复制] |
|
| 摘要: |
| 研究能源互联网发展背景下含有分布式电源(distributed generation, DG)接入的电力市场中的多主体博弈问题。首先，利用多代理(multi-agents)技术，在由发电商、供电商与多类型用户组成的市场中，搭建了多主体博弈框架。其次，针对该框架下的市场各主体特点，分别采用统一市场出清价格(market clearing price, MCP)和按报价支付(pay as bid, PAB)的市场机制，构建了最优供应函数决策模型、最优投标电价决策模型以及考虑投标风险的最优投标电量决策模型。并且，基于效用函数，考虑弹性负荷(具有分布式发电或可中断能力)用户的购售能力，建立了最优购电和最优DG发电量决策模型。最终，实现了市场各博弈主体的利益均衡化目的。仿真结果表明，基于多代理技术的博弈能实现市场各主体的合理收益，不同主体组合的市场博弈结果存在较明显差异。充分挖掘弹性负荷调节能力可有效提高分布式清洁能源发电渗透率。以上结果符合工程实际与设计需求，验证了所提模型的合理性与有效性。 |
| 关键词: 分布式电源 多代理(multi-agent) 博弈 强化学习算法 |
| DOI：10.7667/PSPC151729 |
| 投稿时间：2015-09-26修订日期：2016-03-28 |
| 基金项目:国家高科技发展计划项目(2014AA051901) |
|
| Analysis of game among multi-agents in electrical power market with integration of distributed generation |
| LI Gang,LIU Jichun,WEI Zhenbo,LIU Junyong,LIU Yang,LI Dan,TANG Hu |
| (School of Electrical Engineering and Information, Sichuan University, Chengdu 610065, China) |
| Abstract: |
| This paper researches the game among multi-agents in electric power market with integration of distributed generation under the background of energy internet. First, it builts a game framework which is consisted by generation companies, power suppliers and multiple types of users using multi-agent technique. And then, it adopts the uniform market clearing price and payment as bidding to create a series of models including optimal bidding electricity decision model. These models serve for generators and users who have abilities to electricity with grid according to their features. At the same time, optimal decision model is modeled for customers who can use distributed generation and elastic load to adjust their demands based on the utility function. The interest equalization of the market is successfully achieved through the game theory in our proposed simulations. The experimental results show that all market participants can get reasonable revenue, whilst the outcomes are various in different cooperation scenarios. The results also prove that increasing the flexibility of the load regulation can significantly improve penetration of distributed clean energy. |
| Key words: distributed generation multi-agent game reinforcement learning algorithm |
