Abstract
This paper aims to present a multilayer neural network model for a (3 + 1)-dimensional p-gBLMP equation. The generalized bilinear p-gBLMP equation is constructed, on the basis of the generalized bilinear operators. Through selecting different values in each layer, novel types of tensor functions can be furnished. We set the hidden neurons to some specific functions in some cases, and compute four types of new exact network model solutions for the p-gBLMP equation. The novelty and advantage of the proposed model are illustrated by applying to this model. Some plots of those presented new solutions are made to exhibit wave characteristics.