Application of deep reinforcement learning in power system scheduling strategy simulation

With the increasing scale and complexity of power system, the current scheduling strategy methods often have the problems of limited adaptability and unsatisfactory results. Therefore, a power system scheduling method using DQN (Deep Q Network) is introduced in this study. This method uses neural network model to learn and optimize scheduling strategy, and has strong adaptability and enhanced global optimization ability. Comparative evaluations with genetic algorithms and particle swarm algorithms on IEEE-118 and SG-126 datasets show that the proposed method has significant advantages in terms of cumulative reward and stable runtime.