计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (22): 216-219.

• 工程与应用 • 上一篇    下一篇

并行GA_ANN预测模型研究

赵 宏1,2,张 洁1,2,侯鲁健3,王 恺1,2,白志鹏2,4   

  1. 1.南开大学 信息技术科学学院,天津 300071
    2.国家环境保护城市空气颗粒物污染防治重点试验室,天津 300071
    3.济南市环境保护科学研究所,济南 250014
    4.南开大学 环境科学与工程学院,天津 300071
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-08-01 发布日期:2011-08-01

Parallel GA_ANN predicting model research

ZHAO Hong1,2,ZHANG Jie1,2,HOU Lujian3,WANG Kai1,2,BAI Zhipeng2,4   

  1. 1.College of Information Technical Science,Nankai University,Tianjin 300071,China
    2.State Environmental Protection Key Laboratory of Urban Ambient Air Particulate Matter Pollution Prevention and Control,Tianjin 300071,China
    3.Environmental Science Institute of Jinan,Jinan 250014,China
    4.College of Environmental Science and Engineering,Nankai University,Tianjin 300071,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-08-01 Published:2011-08-01

摘要: 遗传算法GA与人工神经网络ANN相结合的GA_ANN预测模型,在解决大规模问题时,训练模型产生的巨大计算量会导致相当耗时。利用gprof工具剖析出GA_ANN模型的瓶颈所在,并基于OpenMP多线程技术设计出一种并行方案。实验结果表明随着种群规模、繁殖代数以及ANN训练次数的增加,粗粒度的策略结合一定数量的线程能够获得理想的加速比。

关键词: 遗传算法_人工神经网络(GA_ANN), 预测模型, OpenMP, gprof

Abstract: In solving large scale problems,the GA_ANN model which combined the genetic algorithm and artificial neural network would be very time-consuming on the training stage.The bottleneck of GA_ANN model is analyzed by using gprof,and a parallel scheme is designed based on OpenMP multi-threading technology.The experiment results indicate that coarse-grained threading parallelization strategy can get an ideal speedup as the increasing of the population size,number of breeding generation and ANN training.

Key words: Genetic Algorithm_Artifical Neural Network(GA_ANN), predicting model, OpenMP, gprof