论文部分内容阅读
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.
According to the development of continuous Hopfield network technology for solving nonlinear programming, the Hopfield neural network (HNN) model is classified by using the optimization method based on the network model: solving the unconstrained HNN and the constrained HNN, the latter And can be divided into HNN based on penalty function and HNN based on Lagrange multiplier method.The frontier problems involved in neural optimization technology are expounded and the future development direction is pointed out as developing intelligent optimization and solving system.