基于进化规划的自适应高斯神经网络.pdfVIP

  • 1
  • 0
  • 约1.3万字
  • 约 4页
  • 2017-08-06 发布于浙江
  • 举报

基于进化规划的自适应高斯神经网络.pdf

基于进化规划的自适应高斯神经网络

139 Vol. 24 No. 1 2004 1 Ship Electronic Engineering 62 ( 430033) : , BP , , , , , , , : ; ; : U666.7 Adaptive Gauss Neural Network Based on Evolutionary Programming Wang Xiangjun Lin Chunsheng Gong Shenguang ( Dept. of Weaponry Eng. , Naval Univ. of Engineering, W han 430033) A stract:Adapt ive Ga ss ne ral netw ork can pick p recognition characteristics from ship noise spectr m and classify the ship noise effectively, b t error energy f nction of Back- Propagation algorithm sed in the Adaptive Ga ss ne ral netw ork is an irreg lar hyper camber, and the network always seeks in the local extreme val e. Then evol tionary programming based adaptive Ga ss ne ral network is proposed in this paper,w hich can opt imize t he str ct re and adj st the weight of the network a tomati cally. This kind of adaptive Ga ss ne ral network is sed to classify passive sonar target, the res lt of experiment shows that the adaptive Ga ss ne ral netw ork based on evol tionary programming can solve t he local extreme val e problem and can be more ef fective in classification. Key words: Evol tionary Programming; recognition; ne ral network Class num er:U666. 7 : a ; b , BP a, b, x( n) , , [ 1,2] , BP , , BP , , , BP [ 4] , , , [3] ,

您可能关注的文档

文档评论(0)

1亿VIP精品文档

相关文档