Neuro-Monte Carlo solution of electrostatic problems (Q1387903)

The use of artificial neural networks in solving electrostatic boundary value problems is presented. The authors propose to combine one artificial neural network with one Monte Carlo method. They use two separate neuro-Monte Carlo systems to solve the Laplace and Poisson equations for homogeneous and inhomogeneous problems. Numerical examples are also presented. Two-dimensional problems of electrostatic are considered.