应用科学学报 ›› 2000, Vol. 18 ›› Issue (1): 24-26.

• 论文 • 上一篇    下一篇

单参数快速搜索BP算法的研究与应用

陈晓东1, 冯英浚2   

  1. 1. 哈尔滨工业大学网络中心, 黑龙江哈尔滨 150001;
    2. 哈尔滨工业大学应用数学系, 黑龙江哈尔滨 150001
  • 收稿日期:1998-08-30 修回日期:1999-04-08 出版日期:2000-03-31 发布日期:2000-03-31
  • 作者简介:陈晓东(1969-),男,广东省兴宁县人,讲师,博士;冯英浚(1940-),男,上海人,教授,博导.
  • 基金资助:
    国家自然科学基金资助项目(69974013)

Study on BP Algorithm of Single Parameter Quick Search and Its Application

CHEN Xiao-dong1, FENG Ying-jun2   

  1. 1. Network Centre, Harbin Polytechnic University, Harbin 150001, China;
    2. Department of Applied Mathematics, Harbin Polytechnic University, Harbin 150001, China
  • Received:1998-08-30 Revised:1999-04-08 Online:2000-03-31 Published:2000-03-31

摘要: 提出了一种基于单参数坐标轮换法的改进BP算法.与其他算法相比,它不仅寻优方向性强,目标函数计算量小,而且使BP网络的收敛速度明显提高.仿真应用表明,该算法具有很好的鲁棒性.

关键词: 单参数, 神经网络, BP算法, 过热器系统

Abstract: Single parameter quick search method, an improved BP algorithm based on single parameter coordinate -rotated method, is presented in this paper. Compared with other BP algorithms, it not only has a definite orientation in optimization and little time-consuming in computation, but also makes the convergent speed of the BP network very fast. Simulation results show that the new BP algorithm has strong robustness.

Key words: neural network, single parameter, BP algorithm, superheater system

中图分类号: