摘要
提出了一种基于蚁群聚类算法和裁剪方法的RBF神经网络优化算法。利用蚁群算法的并行寻优特征和一种自适应调整挥发系数的方法,提出一种新的聚类算法来确定RBF神经网络中基函数的位置;通过一种裁减的方法,除去对整个网络的输出贡献不是很重要的隐层单元来约简隐含层的神经元,以达到简化RBF神经网络结构的目的。对非线性函数进行逼近仿真,结果表明:优化算法有比较好的优化效果,而且,优化后的RBF神经网络的结构小,RBFNN的泛化能力得到了提高。
An optimization algorithm of RBF Neural Networks based on ant colony clustering and a pruning method is proposed. Based on the feature of parallel search optimum of the ant colony algorithm and a dynamic method to adjust the parameter of evaporation coefficient, the center of each basis function of RBF can be defined by using a new proposed clustering algorithm in order to simplify the structure of RBF network, we use a pruning method to remove those hidden units which make insignificant contribution to the overall network output. Then, the approach is used in the approximation of nonlinear function. The resuits indicate that the optimization algorithm has a good optimization performance, furthermore, the RBFNN optimized by the optimization algorithm has a smaller structure, and the generalization ability of RBFNN is improved.
出处
《青岛大学学报(工程技术版)》
CAS
2008年第3期35-39,共5页
Journal of Qingdao University(Engineering & Technology Edition)
关键词
RBF神经网络
蚁群聚类算法
泛化能力
radial basis function neural network ant colony clustering generalization ability