Discrete bi-population differential evolution for optimizing complex wind farm layouts in diverse terrains

Wind energy is a widely promoted renewable energy technology in many countries. Traditional wind farm layouts often adopt regular geometric patterns, resulting in significant wake effects within the wind farm. Reducing these wake effects through the design of optimal turbine layouts to enhance wind farm efficiency is a critical challenge. While there are existing studies that optimize wind farm layouts using evolutionary algorithms, issues such as instability in optimization performance and difficulty in optimizing complex terrains persist. To address these challenges, we propose an effective discrete bi-population differential evolution (BDE) algorithm for wind farm layout optimization (WFLO). Compared to traditional evolutionary algorithms, our algorithm offers greater interpretability and a broader search space, making it highly suitable for WFLO. Extensive experiments show that our proposed algorithm can identify optimal layouts faster in both simple and complex terrains, compared to other evolutionary algorithms specifically designed for WFLO and the champion algorithms from competitions. This evidence confirms that our algorithm achieves state-of-the-art performance in WFLO. The code is available at the following repository: https://toyamaailab.github.io/sourcedata.html.