Research on Cooperative Control Algorithm Based on Distributed Multi-region Integrated Energy System

In this paper, an OQL-SA (opposite Q learning-simulated annealing) cooperative control algorithm for distributed multi-region integrated energy systems is proposed from the perspective of automatic generation control. The proposed algorithm combines the SA algorithm with the OQL algorithm. The SA algorithm effectively avoids the damage of too much exploration on the process of learning algorithm and even the learning result, while the OQL algorithm improves the learning speed by updating two Q values at the same time. Through the simulation analysis of the multi-region load frequency control model based on the Southwest Power Grid, the results show that the proposed algorithm can obtain the multi-region optimal cooperative control, and has better control performance and faster convergence rate compared with other reinforcement learning algorithms.