基于免疫进化算法的径向基函数网络

Radial Basis Function Network Based on Immune Evolutionary Algorithm

基于免疫进化算法,提出了一种设计径向基函数(RBF)网络的新算法——免疫径向基函数网络(IRBF)训练算法.该算法通过提取RBF网络核函数宽度的先验知识作为疫苗构成免疫算子,缩小了标准进化算法搜索空间的范围,提高了算法的收敛速度.计算机仿真表明,采用这种算法训练的RBF网络达到了较好的性能.

Based on immune evolutionary algorithm, a novel radial basis function (RBF) designing algorithm, the immune RBF (IRBF) network, was proposed. Through extracting the preliminary knowledge about the width of the basis function as the vaccine to form an immune operator, the IRBF algorithm reduces the searching space of the evolutionary algorithm and improves the convergence speed. The computer simulations demonstrate that the RBF network designed in this method has good performances.