| 发电商基于Q—Learning算法的日前市场竞价策略 |
| |
| 引用本文： | 王帅.发电商基于Q—Learning算法的日前市场竞价策略[J].电力技术经济,2010,22(3):34-39. |
| |
| 作者姓名： | 王帅 |
| |
| 作者单位： | 国网北京经济技术研究院,北京,100052 |
| |
| 摘 要： | 电力市场仿真可以研究市场规则、市场结构对价格形成的影响和市场参与者的动态行为。初步建立了用智能多代理模拟日前市场发电商竞价策略的模型，采用Q-Learning算法优化自身策略。改进了增强学习算法中探索参数的选取。使程序在开始阶段以较大的概率进行新的搜索，避免过早陷入局部最优。改进了阶梯形报价曲线的构造方法，减小了计算量，提高了计算速度。
|
| 关 键 词： | 电力市场 定价 智能代理 Q—Learning算法 竞价策略 |
Generators'Bidding Strategies in the Day-ahead Market Based on Q-Learning Algorithm |
| |
| Wang Shuai.Generators'Bidding Strategies in the Day-ahead Market Based on Q-Learning Algorithm[J].Electric Power Technologic Economics,2010,22(3):34-39. |
| |
| Authors: | Wang Shuai |
| |
| Institution: | WANG Shuai(State Power Economic Research Institute,Beijing 100052,China) |
| |
| Abstract: | Electricity market simulation can be used to study the influence of the market mechanism and structure on pricing and to analyze the dynamic behaviors of the market participants.This paper sets up an agent-based day-ahead market simulation model for the generator bidding strategy,in which the Q-Learning algorithm is used to optimize bidding strategies.The algorithm is modified to improve the convergence performance of test parameters so that the program can start new searches with a larger probability at th... |
| |
| Keywords: | power market pricing agent-based system Q-Learning bidding strategy |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
