一种线性约束规划问题的神经网络模型及其应用

A Novel Neural Network Model for Solving Programming Problems Based on Linear Restraints

提出了一种基于梯度投影矩阵下的求解线性约束下规划问题的神经网络。针对解的稳定性问题,导出了该网络相关参数之间的关系。由文中定义可知,该网络不但适合于求解线性约束下线性或非二次规划问题,而且也用于求解线性或非线性方程组问题,比其它规划问题的神经网络方法更具有一般性。

A novel neural network is proposed,which solve the programming problems with linear restraints based on gradient projecting matrix.And its parameters relation stable solution algorithm is deduced.Compare it with other models which solve the programming problems with linear restraints,it is more universal,since it not only can use to solve linear or that more than quadratic programming problems ,but also can use to solve the solution to the linear and non-linear equation.