论文部分内容阅读
针对目前电价预测算法的局限性,提出一种基于自适应动态规划方法的自学习、自适应智能算法。按照Bellman最优化基本原理,使用Agent逐步与环境的交互作用来寻求预测电价和实际电价的误差最小值,得到系统边际电价的最优解。采用美国加州电力市场的数据进行电价预测仿真。与常规方法相比,该方法的拟合精度和平均绝对百分误差均有很大提高。
Aiming at the limitation of current price forecasting algorithm, a self-learning and self-adaptive intelligent algorithm based on adaptive dynamic programming is proposed. According to the basic principle of Bellman optimization, the minimum error of the forecast electricity price and the actual electricity price is obtained by gradually interacting with the environment and the optimal solution of the system marginal price is obtained. Electricity Price Prediction Simulation Based on Data from California Electricity Market in the United States. Compared with the conventional method, the fitting accuracy and the average absolute percentage error of the method are greatly improved.