欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
在重复博弈的电力市场中，发电商的竟价策略和个体学习行为是个复杂的动态问题。疗析发电商竟价策略这一热点问题的最新研究思路和仿真方法。为模拟发电商的自主学习和市场的动态演化过程，引入博弈学习理论，建立基于模仿学习的发电商竞价策略模型，分析不同信息条件下的学习动态过程，及重复竞标时发电商模仿学习对博弈演变及市场均衡的影响。仿真结果表明模仿学习具有较好的收敛性。另外，完全信息条件下，发电商全局学习能引导市场趋于瓦尔拉斯均衡和古诺一纳什均衡二者之间的一个均衡状态；非完全信息条件下，发电商全局学习的结果是市场总体行为收敛到古诺一纳什均衡，而发电商的个体行为则与古诺一纳什均衡状态存在偏差。
英文摘要：
In electricity markets with repeated games, GenCos＇ bidding strategies and learning behaviors are complex dynamic problems. The research thinking and simulation methods of the above hot issues are reviewed. To simulate and study GenCos＇ learning behaviors and the dynamic evolution process of electricity markets, a GenCo＇ bidding model based on imitation learning is presented. By importing the idea of learning in games and applying the proposed bidding model, the dynamic learning process, and its influence on the evolution process of games and market equilibrium in repeated bidding are analyzed under various information conditions. Simulation results show the convergency and efficiency of imitation learning. Meanwhile, GenCos＇ imitation learning in global view will produce quite different outcomes under full and partial information conditions. Under full information condition, the general market behavior converges at an intermediate equilibrium state between Walrasian equilibrium and Cournot-Nash equilibrium; under partial information condition, the general market behavior converges at Cournot- Nash equilibrium while the individual behavior of a GenCo diverges.
同期刊论文项目
同项目期刊论文
期刊信息
位置：