Application of Reinforcement Learning Algorithm for Scheduling of Microgrid

The increased penetration of renewable energy sources poses challenges in the control and operation of micro grids. Machine learning algorithms have emerged as a powerful tool in solving these issues. This paper presents an attempt to employ Q learning algorithm which is a variant of Reinforcement learning for conducting economic scheduling of a micro grid in the presence of photovoltaic cells and battery. The optimal scheduling reduces the net transaction cost with the grid and gives the battery operating schedule over a period of one day. The numerical results show the effectiveness of the method employed in comparison with the traditional methods.