论文部分内容阅读
本文提出一类用于求解偏微分方程的分布参数神经网络,并且在连续时空上研究了它的动态特性。最后还给出了两个模拟试验,用于检验这类神经网络的有效性。
In this paper, a class of distributed parameter neural networks for solving partial differential equations is proposed, and its dynamic characteristics are studied on a continuous space-time. Finally, two simulation tests are given to verify the validity of these neural networks.