Adaptive strategy Q-learning differential evolution for microgrid scheduling optimization

doi:10.1016/j.swevo.2026.102509

Adaptive strategy Q-learning differential evolution for microgrid scheduling optimization

2026-08-20

0

PRE

AI

DOI：10.1016/j.swevo.2026.102509

Original

Original request for help

Save

Abstract

En 中文

Optimal scheduling of grid-connected microgrids is challenging due to the nonlinear coupling among distributed energy resources, multiple operational constraints, and the uncertainty of renewable generation. To address this problem, this study develops a comprehensive microgrid scheduling model that incorporates photovoltaic units, wind turbines, microturbines, diesel generators, fuel cells, and battery energy storage systems. Because the resulting optimization problem is high-dimensional, nonlinear, and strongly constrained, conventional methods often struggle to achieve high-quality solutions. Therefore, an Adaptive Strategy Q-learning Differential Evolution (ASQDE) algorithm is proposed. The algorithm integrates a reward-driven Q-learning mechanism for adaptive mutation-strategy selection and a stochastic parameter pool for dynamic adjustment of control parameters, thereby improving the balance between exploration and exploitation. The proposed method is evaluated on both 24-hour and 15-minute scheduling scenarios. Experimental results show that ASQDE consistently outperforms several state-of-the-art algorithms in terms of convergence speed, solution quality, and robustness. These findings demonstrate that ASQDE is an effective approach for solving complex microgrid scheduling problems.

Adaptive strategy Q-learning differential evolution for microgrid scheduling optimization

Abstract

Journal

Organization