Computer Engineering and Applications ›› 2009, Vol. 45 ›› Issue (3): 1-6.DOI: 10.3778/j.issn.1002-8331.2009.03.001

• 博士论坛 • Previous Articles     Next Articles

Comparison of neural network and evolutionary algorithm on engineering optimization

ZHANG Yu-dong1,WU Le-nan1,WU Han-qian2   

  1. 1.School of Information Science & Engineering,Southeast University,Nanjing 210096,China
    2.School of Software,Southeast University,Nanjing 210096,China
  • Received:2008-10-06 Revised:2008-12-10 Online:2009-01-21 Published:2009-01-21
  • Contact: ZHANG Yu-dong

工程优化问题中神经网络与进化算法的比较

张煜东1,吴乐南1,吴含前2   

  1. 1.东南大学 信息科学与工程学院,南京 210096
    2.东南大学 软件学院,南京 210096
  • 通讯作者: 张煜东

Abstract: There are various types engineering optimization problems with different models and methods.This paper divides all the problems into two sorts,black-box optimization and white-box optimization,along with the deduction of their models.For the black-box optimization,the superiority of Feed-forward Neural Network(FNN) on system approximation,and the merits and demerits of Evolutionary Algorithm (EA) and Back Propagation (BP) on the weight values problems are discussed.For the white-box optimization,the merits and demerits of EA and Recurrent Neural Network(RNN),and the survey on popular EAs and their common improvements are discussed.This article provides a more comprehensive cognition on current engineering optimization problem,and the function by neural network and evolutionary algorithm.

Key words: engineering optimization, Feed-forward Neural Network(FNN), Recurrent Neural Network(RNN), Evolutionary Algorithm (EA)

摘要: 目前工程优化问题不仅种类繁多,而且各自采用的模型与方法迥异。从方法论的高度,将现有工程优化问题分为黑箱优化与白箱优化,然后推出各自的优化模型。对于黑箱优化问题,阐述了前向神经网络在系统逼近上的优势,以及进化算法与BP算法在求解神经网络权值上的优劣;对于白箱优化问题,阐述了进化算法与反馈神经网络的优缺点和目前流行的进化算法及其通用改进策略。通过分析,可以对目前的优化问题,以及神经网络与进化算法在其中的作用,有更加全面的认识。

关键词: 工程优化问题, 前向神经网络, 反馈神经网络, 进化算法