说明：  PSO的RBFNN优化程序 算法步骤 1.样本数据归一化处理，即将输入输出归一化到[-1,1]区间； 2.确定RBF网络的中心和宽度； 3.以拟合误差的均方根作为性能指标，使用PSO算法优化RBF网络输出层到隐层的连接权值矩阵
(PSO-RBFNN algorithm optimization procedures Step 1. Sample data normalization treatment, about input and output normalized to [-1,1] interval 2. To determine the center and width of the RBF network 3. To the fitting error of the mean square roots as a performance index, using the PSO algorithm to optimize RBF network output layer to hidden layer connection weight matrices)