基于Q学习算法的发电商报价策略模型  

Power supplier bieding strategies based on Q-learning algorithm

作　　者：高瞻[1] 宋依群[1] 

机构地区：[1]上海交通大学电气工程系,上海200240

出　　处：《华东电力》2008年第4期20-22,共3页East China Electric Power

摘　　要：针对日前电力市场发电商利益最大化问题,利用Q学习算法构造了发电商策略报价模型。以发电商即时收益和市场相对占有率组成奖赏函数,使发电商策略收益最大化并同时达到提高市场占有率的目的。通过算例验证了模型的有效性,发现如果发电商试图提高市场占有率将选择低报价策略;考虑爬坡限制后使得24 h独立的Q学习联立,引起发电商报价策略变化。Based on the Q-leaning algorithm, a model of power supplier bidding strategy to maximize the supplier＇s profit in the day-ahead electricity market is presented. The reward function is composed of the instant revenue and the relative market share, which enables the supplier to achieve the goal of raising its market share. The model is proved effective by simulation. It is found that the supplier has to choose a low-bidding strategy to raise their market share, and the optimal bidding strategy changes accordingly when the Q-learning process separated among 24 hours are connected after the constraint of ramp rate is considered.

关 键 词：Q学习 报价策略 市场相对占有率Agent 

分 类 号：TM76[电气工程—电力系统及自动化]