Abstract In this paper, a novel multi-population parallel co-evolutionary differential evolution, named MPPCEDE, is proposed to optimize parameters of photovoltaic (PV) models and enhance conversion efficiency of solar energy. In the MPPCEDE, the reverse learning mechanism is employed to generate the initial several subpopulations to enhance the convergence velocity and keep the population diversity. A new multi-population parallel control strategy is developed to maintain the search efficiency in subpopulations. The co-evolutionary mutation strategy with elite population and three mutation strategies is proposed to reduce computing resources and balance the exploration and exploration capability through the cooperative mechanism, improve the convergence speed, realize the information exchange. Then the MPPCEDE is employed to effectively optimize parameters of PV models under various conditions and environments to obtain a parameter values of PV models. Finally, the effectiveness of the proposed method is tested by different PV models and manufacturer's datasheet. The experimental and comparative results demonstrate that the MPPCEDE exhibits higher accuracy and reliability, and has fast convergence speed by comparing with several methods in extracting parameters of PV models.

中文翻译：

---

MPPCEDE：用于参数优化的多种群并行协同进化差分进化

摘要 在本文中，提出了一种新的多种群并行协同进化差分进化，命名为MPPCEDE，用于优化光伏（PV）模型的参数并提高太阳能的转换效率。在MPPCEDE中，采用反向学习机制生成初始的几个子种群，以提高收敛速度并保持种群多样性。开发了一种新的多种群并行控制策略，以保持子种群中的搜索效率。提出了精英种群协同进化变异策略和三变异策略，通过协同机制减少计算资源，平衡探索与探索能力，提高收敛速度，实现信息交换。然后采用MPPCEDE对各种条件和环境下的光伏模型参数进行有效优化，得到光伏模型的参数值。最后，通过不同的光伏模型和制造商的数据表测试了所提出方法的有效性。实验和对比结果表明，MPPCEDE与几种提取光伏模型参数的方法相比，具有更高的精度和可靠性，收敛速度快。