利用遗传算法对RBF（径向基）神经网络的中心值和宽度进行优化。实现效果比较好，能够高精度的实现非线性逼近。(Using genetic algorithm to optimize the center value and width of RBF (radial basis) neural network. The realization effect is better, and the nonlinear approximation can be achieved with high accuracy.)