摘要
提出基于QPSO算法优化的RBF神经网络。此网络的核心算法是将RBF神经网络的参数组成1个向量,构造成QPSO算法中的粒子,由此在可行范围内搜索一组使网络均方误差最小的最优解。实例仿真部分分别用优化前后的网络对Legendre函数进行函数逼近。研究结果表明:经过优化的RBF网络与传统RBF网络相比具有计算精度高、收敛速度快的优点。
The radial basis function neural network was designed based on QPSO algorithm.This algorithm builds a vector composed of parameters of network which is the quantum in the QPSO algorithm,so it is possible to search a set of parameters minimizing the mean squared error.In the simulation,the Legendre function was approximated by the new neural network.The result shows that the new neural network has high accuracy and fast convergence rate compared with the traditional RBF.
出处
《中南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2013年第S1期27-30,共4页
Journal of Central South University:Science and Technology
