.png){kind=icon}
.png){kind=icon}
Return
Reinforcement learning-enhanced genetic algorithm for wind farm layout optimization
.png){kind=icon}
PRE
AI
DOI：10.1016/j.renene.2025.125093.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
A reinforcement learning-enhanced genetic algorithm (RLGA) is proposed for wind farm layout optimization (WFLO). While genetic algorithms (GAs) are among the most effective and accessible methods for WFLO, their performance and convergence are highly sensitive to parameter selections. To address the issue, RL is introduced to dynamically select optimal parameters throughout the GA process. The proposed RLGA is validated by evaluating the WFLO for four physics-informed layouts (aligned, staggered, sunflower, and unstructured) in a classical 50D x 50D wind farm, with a spacing of 5D between neighboring turbines. The advantage of the RLGA becomes particularly evident in optimizing complex layouts. While it achieves similar results to the GA for aligned and staggered layouts, it outperforms the GA for sunflower and unstructured layouts. To further validate and explore its capabilities, we investigate larger wind farms of 150D x 150D with four layouts under three wind conditions: unidirectional uniform, omnidirectional uniform, and spread nonuniform, which are expected to impose greater computational challenges compared to the classical 50D x 50D wind farms. The proposed RLGA is more efficient than the GA, especially for large-scale WFLO problems. This improvement stems from RL's ability to adjust parameters, avoiding local optima and accelerating convergence.
Keywords:
Reinforcement learning
Genetic algorithm
Wind farm layout optimization
Journal
IF:
9.1
Papers:
2.6W
Citations:
12.1W