论文部分内容阅读
本文将Hopfield自联想神经网络和Kosko异联想神经网络推广到无穷维状态空间动态神经网络,即动态分布参数神经网络,并给出了它们的有界性和稳定性。尤其是还研究了带微分算子的多维分布参数神经网络的时空稳定性以及保证稳定情况下所应满足的边界条件。最后,还给出了一个应用实例。
In this paper, we generalize Hopfield self-associated neural networks and Kosko heterogeneous neural networks to infinite-dimensional state-space dynamic neural networks, that is, dynamic distributed parameter neural networks, and give their boundedness and stability. In particular, the space-time stability of multidimensional distributed parameter neural networks with differential operators and the boundary conditions that should be satisfied under stable conditions are also studied. Finally, an application example is given.