求解具有约束的l_1-范数问题的神经网络模型

A neural network for solving l_1-norm problems with constraints

提出了一个解约束最小l_1-范数问题的单层神经网络模型。与已有神经网络模型相比,提出的模型所需神经元数少且层数少。通过引入Lyapunov函数,证明了该模型的稳定性和收敛性。数值试验结果表明所提出的模型具有良好的性能。

This paper presents a one-layer neural network model for solving l1-norm problems with constraints. Compared with some existing neural network models,the proposed model needs fewer neurons and has a simpler structure. The stability and convergence of the proposed model are proved by introducing a Lyapunov function. Some simulation examples are used to illustrate its validity and transient behaviors.