论文部分内容阅读
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 neural network based on gradient projection matrix is proposed to solve the linear constrained programming problem. Aiming at the stability of solution, the relationship between the parameters of the network is derived. It can be seen from the definition in this paper that this network is not only suitable for solving linear or non-quadratic programming problems under linear constraints, but also for solving linear or nonlinear equations. It is more general than other neural network methods for programming problems.