Approximation of functions and their derivatives: A neural network implementation with applications (Q1960876)

The authors apply neural network techniques for the numerical approximation of functions of several variables and their first- and second-order partial derivatives. In order to eliminate the need of domain discretization, they propose to use a neural network based approximation of functions of several variables and their first- and second-order partial derivatives combined with a nonlinear boundary element method. Numerical results for functions of one, two and three variables are presented. Examples of viscoelastic fluid flows of a polymeric liquid (planar Couette and Poiseuille flows) are included.