基于DDPG算法的光伏充电站策略优化

由于引入了光伏新能源,光伏充电站既可以通过向车辆充电获取利润,也可以对电网供电获得效益,因此如何合理安排充电、获得最大收益是充电策略研究的重点。通过深度置信策略梯度(DDPG)算法,对充电策略进行优化,将利润作为回报函数,训练智能体,可在无先验的基础上实现在线优化。最后,通过仿真实验证明,该算法可以在不同车辆、不同光照情况下取得相对较高的利润。

Photovoltaic Charging Station Strategy Based on DDPG

Due to the introduction of new photovoltaic energy,photovoltaic charging stations can both obtain profits by charging vehicles and benefits from grid power supply.Therefore,how to reasonably arrange charging to obtain the maximum benefit is the focus of charging strategy research.A deep confidence strategy gradient algorithm is used to optimize the charging strategy,using profit as a return function,and training agents,which can achieve online optimization without a priori.Finally,simulation experiments prove that the algorithm can achieve relatively high profits under different vehicles and different lighting conditions.