将遗传算法与正交优选法结合 ,用来训练径向基函数 ( RBF)神经网络 ,并对基函数宽度进行自动地调整 ,得到了一种训练 RBF神经网络的新方法 .将其应用于连续流体搅拌反应槽 ( CFSTR)生化反应器的建模中 ,得到了令人满意的结果 .该算法提高了径向基函数神经网络的泛化能力和鲁棒性 ,研究表明是一种有效的“黑箱”动态建模方法 The genetic algorithm is combined with orthogonal optimization method to train radial basis function (RBF) neural network and adjust the width of the basis function automatically, and a new method of training RBF neural network is obtained.It is applied to continuous flow Satisfactory results have been obtained in the modeling of stirred bioreactor (CFSTR) bioreactor.The algorithm improves the generalization ability and robustness of radial basis function neural networks, and studies show that it is an effective “black box ”Dynamic modeling method