摘要： 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模型，并导出了线性约束下规划问题的稳定解法，该网络模型既适合于求解线性约束下线性或非二次规划问题，又适合于求解线性或非线性方程组，与其它规划问题的神经网络相比，更具有一般性。

关键词: 神经网络, 线性约束, 规划问题

Abstract: A novel neural network is proposed, which solves the programming problems with linear restraints based on gradient projecting matrix. And its stable solution algorithm is deduced. Moreover, the parameters to be determined and their relation are deduced too. Compares it with other models, which solve the programming problems with linear restraints, it is an universal model, since it not only can be used to solve linear or that more than quadratic programming problems, but also can be used to solve the solution to the linear and non-linear equation.

中图分类号: 

• TP183