Parameter estimation of partial differential equations using artificial neural network

The work presented in this paper aims at developing a novel meshless parameter estimation framework for system of partial differential equations (PDEs) using artificial neural network (ANN) approximations. PDE models to be treated consist linear and nonlinear PDEs, with Dirichlet Neumann boundary conditions, considering both regular irregular boundaries. This focuses on testing the applicability networks estimating process model parameters while simultaneously computing predictions state variables PDEs representing process. capability proposed methodology is demonstrated five numerical problems, showing that ANN-based approach very efficient by providing accurate solutions reasonable times.