武汉大学学报(理学版)
武漢大學學報(理學版)
무한대학학보(이학판)
JOURNAL OF WUHAN UNIVERSITY(Natural Science Edition)
2001年
3期
347-350
,共4页
二次规划%神经网络%Lyapunov函数
二次規劃%神經網絡%Lyapunov函數
이차규화%신경망락%Lyapunov함수
根据人工神经网络的基本优化机理,提出了一种基于Lagrange函数的适合于求解二次规划问题的神经网络模型,研究了该神经网络的稳定性和收敛性,探讨了提高网络优化计算效率的神经优化策略,仿真结果证明了该神经网络能有效地求解二次规划问题.
根據人工神經網絡的基本優化機理,提齣瞭一種基于Lagrange函數的適閤于求解二次規劃問題的神經網絡模型,研究瞭該神經網絡的穩定性和收斂性,探討瞭提高網絡優化計算效率的神經優化策略,倣真結果證明瞭該神經網絡能有效地求解二次規劃問題.
근거인공신경망락적기본우화궤리,제출료일충기우Lagrange함수적괄합우구해이차규화문제적신경망락모형,연구료해신경망락적은정성화수렴성,탐토료제고망락우화계산효솔적신경우화책략,방진결과증명료해신경망락능유효지구해이차규화문제.
According to the basic optimization principle of artificial neural networks, a Lagrange function based neural network model approach appropriate for quadratic programming problems is given. Stability and convergence of the neural network are investigated. Neural optimization strategy to improve the neural computing efficiency is studied. Simulation demonstrates that these neural networks can effectively solve quadratic programming problems.The new network is easy to construct with circuits, capable of achieving the exact solution.