摘要
根据求解非线性规划的连续Hopfield网络技术的发展,利用网络模型所依据的最优化方法,对Hopfield神经网络(HNN)模型进行了分类:求解无约束化的HNN和求解约束化的HNN.后者又可分为基于罚函数的HNN和基于Lagrange乘子法的HNN.阐述了神经最优化技术所涉及的前沿问题,并指出今后的发展方向为开发智能优化求解系统.
In order to develop neural networks techniques, the history and current research of solving nonlinear programming problems of artificial neural networks techniques are reviewed briefly. Hopfield neural networks (HNN) models can be classified according to different optimization methods of the models. It can classified into HNN of optimization with or without restraining. The latter again is classified into HNN based on penalty function and on Lagrange methods. Advanced problems of 'neural' optimization technique are discussed, and the direction in further research of intelligent optimization systems is pointed out.
出处
《大庆石油学院学报》
CAS
北大核心
2006年第2期90-92,152-153,共3页
Journal of Daqing Petroleum Institute
基金
黑龙江省自然科学基金项目(TF2005-26)
