Accelerating the Linearized Quadrature Method for Nonlinear Diffusion Using a Neural Network Surrogate
Abstract
In this paper, we investigate the [Formula: see text]-dimensional non-linear porous medium equation under the [Formula: see text]-Hilfer derivative. A comprehensive qualitative analysis of the solution is presented, including existence, uniqueness, and Ulam-Hyers (UH) stability. Moreover, for the numerical simulations, a linearized quadrature method (LQM) is developed for the proposed model. MATLAB numerical simulations comprise two- and three-dimensional plots, which demonstrate how changes in the fractional order [Formula: see text] and type parameter [Formula: see text] influence the dynamics of the solutions. The neural network approach demonstrates excellent predictive capability, accurately capturing the behaviour of the model with high precision. This work advances both the theory and computation of the high-dimensional porous medium flows under the [Formula: see text]-Hilfer derivative.