Multi-Objective Interval Optimization Dispatch of Microgrid via Deep Reinforcement Learning
Chaoxu Mu, Yakun Shi, Na Xu, Xinying Wang, Zhuo Tang, Hongjie Jia, Hua Geng
Abstract
This paper presents an improved deep reinforcement learning (DRL) algorithm for solving the optimal dispatch of microgrids under uncertaintes. First, a multi-objective interval optimization dispatch (MIOD) model for microgrids is constructed, in which the uncertain power output of wind and photovoltaic (PV) is represented by interval variables. The economic cost, network loss, and branch stability index for microgrids are also optimized. The interval optimization is modeled as a Markov decision process (MDP). Then, an improved DRL algorithm called triplet-critics comprehensive experience replay soft actor-critic (TCSAC) is proposed to solve it. Finally, simulation results of the modified IEEE 118-bus microgrid validate the effectiveness of the proposed approach.