论文部分内容阅读
首次利用变办法研究运算放大器增益趋于无穷时Hopfield连续神经网络的稳定性。通过研究能量函数的变分,文中无需求解非线性微分方程便可得到A.N.Michel等人所得到的稳定性定理。文中的结果表明:当运算放大器增益趋于无穷时,Hopfield连续神经网络和离散神经网络具有非常相似的稳定特性,特别是,这两种神经网络具有相同的稳态集合。
For the first time, the stability of Hopfield continuous neural network with infinite gain is studied. By studying the variation of energy function, the stability theorem of A.N. Michel et al. Can be obtained without solving the nonlinear differential equation. The results show that the Hopfield continuous neural network and the discrete neural network have very similar stability characteristics when the gain of the operational amplifier tends to infinity. In particular, the two neural networks have the same steady state set.