A New Adaptive Genetic Neural Network Based Active Evolution

摘要：

Neural networks constructions and weights are one aspect of the basic questions. A kind of artificial neural network method based on an active evolution genetic algorithm is proposed. Introduce the algorithms basic idea. Active evolution genetic algorithm is combined the active evolution algorithm which is advantaged both overcoming the local optimized value and keeping rapidly convergence. Save time and space for the construction of new network, improve the outputs error precision and find the better way to solve how to build the networks weights and structures at the beginning. The experiment results show that the algorithm is superior to simple genetic neural network algorithm with higher convergent speed, optimization and practical value of structures and weights, and improves networks forecasting accuracy.

关键词： genetic algorithm active evolution neural network

作者: Yan Ying-fu Wen Hui

作者单位: Key Laboratory of Nondestructive Test Ministry of Education, Nanchang HangKong University Nanchang,J School of Computing, Nanchang HangKong University Nanchang,JiangXi,330063,P.R.Chin

会议类型: 国际会议

会议名称: Second International Symposium on Electronic Commerce and Security(第二届电子商务与安全国际研究大会)(ISECS 2009)

会议地点: 南昌

会议语种:英文

页码: 444-447

在线出版日期: 2009-05-22（万方平台首次上网日期，不代表论文的发表时间）