基于进化规划的自适应高斯神经网络

自适应高斯神经网络能够对目标信号的功率谱有效识别特征进行自动提取和分类，但此网络使用BP算法，其误差能量函数是一个不规则的超曲面，容易陷入局部极小值。因此，提出了一种使用进化规则来设计和训练自适应高斯神经网络的新方法，该方法能够自动的确定网络的最优结构和联结权值，同时避免网络的局部优化。将该方法用于被动声呐晶目标的分类识别，实验结果表明基于进化规则的自适应高斯神经网络能够有效的克服局部最小问题，具有更好的识别率。

Adaptive Gauss Neural Network Based on Evolutionary Programming

Adaptive Gauss neural network can pick up recognition characteristics from ship noise spectrum and classify the ship noise effectively,but error energy function of Back-Propagation algorithm used in the Adaptive Gauss neural network is an irregular hyper camber,and the network always seeks in the local extreme value.Then evolutionary programming based adaptive Gauss neural network is proposed in this paper,which can optimize the structure and adjust the weight of the network automatically.This kind of adaptive Gauss neural network is used to classify passive sonar target,the result of experiment shows that the adaptive Gauss neural network based on evolutionary programming can solve the local extreme value problem and can be more effective in classification.