摘要： 标准BP学习算法是多层感知器的一种训练学习算法，是基于无约束极值问题的梯度法而设计的。针对标准算法存在的收敛速度慢、目标函数易陷入局部极小等缺点，该文提出了一种基于粒群优化的全新学习算法——粒群学习算法。该算法采用并行全局寻优策略，使网络以更快的速度收敛至全局最优解，且更易于编程实现。仿真实例证明，该算法是一种简洁高效的BP神经网络学习算法，有着极为广泛的应用前景。

关键词: 多层感知器, BP算法, 粒群优化, 粒群学习算法

Abstract:

Standard BP algorithm is a kind of learning algorithm of multilayer perceptrons. It is designed based of gradient method. To overcome its defects, the paper proposes a new learning algorithm of BP network —— particle swarm learning algorithm. The algorithm adopts parallel technology to quicken the speed, and it is simply achieved by programming. Simulation results indicate that the particle swarm learning algorithm is a simple and efficient learning algorithm, and it has widely application prospect.

Key words: Multilayer perceptrons(MLP), BP algorithm, Particle swarm optimization, Particle swarm learning algorithm(PSLA)