一种跨邻域学习的改进粒子群优化算法

An Improved Particle Swarm Optimization Algorithm by Learning Across Neighborhoods

为了改善传统粒子群优化算法过早陷入局部最优解的缺点,进一步增强算法收敛性,通过使用一定范围内邻域最好位置lBest代替自身历史最好位置pBest进行速度与位置更新,以增强粒子跨邻域学习能力。使用整个群体中最好位置gBest进行速度与位置更新,可增强算法收敛性,且具有较好的全局搜索能力。在8个不同的单峰和多峰函数上系统地对3种算法进行测试与比较,实验结果表明,提出的跨邻域学习改进粒子群优化算法可避免粒子群陷入局部最优解,求解精度与算法收敛性都提升了15%以上。

In order to improve the shortcoming of the traditional particle swarm optimization algorithm,the convergence of the algo?rithm is further enhanced.By using the best position lBest in a certain range of neighborhood to replace the best position pBest in its own history for speed and position update,the ability of particles to learn across the neighborhood is enhanced,and the best position gBest in the whole population is used for speed and position update,which enhances the convergence of the algorithm and has a good global search ability.Three algorithms are tested and compared systematically on eight different single-peak and multi-peak functions.The experimental results show that the improved particle swarm optimization algorithm with cross-neighborhood learning improves the solution accuracy by more than 15%and the convergence by more than 15%.The improved particle swarm optimization algorithm based on cross-neighborhood learning avoids the particle swarm falling into the local optimal solution and enhances the convergence of the algorithm.