基于遗传算法的RBF-PLS方法在辐射源识别中的应用

RBF-PLS是一种有效的径向基网络构造方法，较好地解决了隐单元数和各中心的取值问题，但宽度系数和PLS成分数难以选定。为此，该文提出采用混合编码遗传算法，以径向基网络的拟合性能和泛化能力为目标，优选宽度系数和PLS成分数，以此建立RBF-PLS-GA模型。将该方法用于雷达辐射源识别，效果良好，明显优于其他网络模型。

The RBF-PLS Approach Based on Genetic Algorithm and Its Application in Radar Model Recognition

Radial Basis Function-Partial Least Square regression(RBF-PLS)approach is a rapid and efficient method in constructing Radial Basis Function Network(RBFN),and it has put forward a solution to the problem about the choice of the number and the centers of the radial basis functions.But it is difficult to optimize the spread parameter of the radial basis functions and the number of PLS components extracted.A hybrid coding genetic algorithm,which uses different coding methods for different type of variables is proposed to get the optimal solution for the spread parameter and the number of PLS components.The object function of GA is the performance of fitting and predicting of the model.The approach is successfully applied to radar model recognition.