计算机与现代化

• 网络与通信 • 上一篇    下一篇

基于迭代参数优化的自适应小波网络均衡算法

  

  1. (中山大学新华学院信息科学系,广东广州510520)
  • 收稿日期:2015-02-03 出版日期:2015-06-16 发布日期:2015-06-18
  • 作者简介:邱泽敏(1983-),女,广东潮州人,中山大学新华学院信息科学系工程师,硕士,研究方向:软件工程,计算机网络技术。

Adaptive Wavelet Network Equalization Algorithm Based on Iterative Parameter Optimization

  1. (Information Science Department, Xinhua College of Sun Yatsen University, Guangzhou 510520, China)
  • Received:2015-02-03 Online:2015-06-16 Published:2015-06-18

摘要: 针对现有网络均衡算法中存在的收敛速度慢、计算冗余等问题,通过对传统的自适应均衡算法与小波变换进行相关研究分析,提出一种基于小波变换的网络均衡算法。小波变换的良好鲁棒性弥补了传统自适应均衡算法中收敛速度慢的缺陷,通过分析算法的收敛性,重新设置迭代中的参数。实验结果表明,该算法的实验结果与预期效果基本相符,具有良好收敛效果的同时并保持了较低的误码率。

关键词: 神经网络, 均衡算法, 小波变换, 收敛, 迭代

Abstract: Existing network balancing algorithm has the problems of the slow convergence and the computational redundancy. In response to this phenomenon, through the analysis of traditional adaptive equalization algorithms and wavelet transform, we proposed a network equalization algorithm based on wavelet transform. The robustness of the wavelet transform compensates the slow convergence of the adaptive equalization algorithm. Through analyzing the convergence of the algorithm, the iteration parameters are reset. Experimental results show that the experimental results are basically consistent with the expected results. It is of a good convergence effect and a lower error rate.

Key words: neural networks, balancing algorithm, wavelet transform, convergence, iteration

中图分类号: