基于粒子群算法的RBF神经网络的优化方法

Optimization Approach Based on Particle Swarm for RBF Neural Network

本文用粒子群算法来优化RBF神经网络的中心值和连接权值,使之具有更强的非线性逼近能力,并将优化后的RBF神经网络和未经优化的RBF神经网络用于非线性函数的逼近,实例证明优化后的RBF神经网络比未经优化的RBF神经网络具有更强的非线性函数的逼近能力。

In this paper, particle swarm is used to optimize the centers and widths of RBF, so that it has a strong nonlinear approximation ability. Example proves that the optimized RBF neural network has stronger non-linear function approximation ability than without the optimization of RBF neural network.