摘要
通过一种新的对偶形式 ,得到一种新的易于实现的解线性规划问题的神经网络 ,证明了网络具有全局指数收敛性 ,使得线性规划问题的神经网络解法趋于完善 .
By means of new theorems on duality, a sort of recurrent neural network for solving linear programming problems is given, which can be realized easily by circuits. The algorithm's exponentially asymptotic stability in the whole is proven. It makes the neural computing approach for linear programming trend to perfect.
出处
《自动化学报》
EI
CSCD
北大核心
2003年第2期219-226,共8页
Acta Automatica Sinica
基金
theNational 86 3Plan(2 0 0 2AA4 14 310 )
关键词
偶性
线性规划
神经网络
解法
全局指数收敛性
Algorithms
Computer simulation
Convergence of numerical methods
Linear programming
Stability
