简化的分类微粒群算法及其在风电场建模中的应用

提出一种简化的分类微粒群算法.首先将微粒按适应值的差异划分成较好、普通和较差3类;然后对这3类微粒分别采用3种对应的没有速度项的简化模型进行动态制整,有效地增加了种群的多样性.通过对4种典型测试函数的仿真实验,并与经典PSO和2个目前较为流行的改进PSO进行比较,实验结果表明了所提出的改进算法具有更好的优化性能.将改进算法用于风电场风速概率模型优化的实验结果表明,与传统最小二乘法相比,该方法拟合的Weibull参数精度更高,更具实际参考价值.

Simpilified Classification Particle Swarm Optimization Algorithm and Its Apllication In Wind Farm Modeling

A simplified classification particle swarm optimization algorithm(PSO) is proposed. At first, particles are divided
into three categories, such as the better, ordinary and the worse according to their fitness. Then, three types of simplified
models without velocity part in classical particle swarm optimization algorithm are used to adjust these three kinds of
classified paticles respectively. The diversity of algorithm is enhanced effectively. Through the simulation experiments with
four test functions, compared with the basic PSO and another improved PSO currently, the improved algorithm proposed has
better optimization performance. Finally, the improved algorithm is applied to optimize wind probability modeling, and the
results show that this method has more accuracy and more practical reference than least-squares method.