Neuro-Monte Carlo solution of electrostatic problems

This paper presents the use of artificial neural networks (ANNs) in solving electrostatic boundary? value problems. In forming a solution system, ?one ANN is combined !vith one Monte Cai lo method (namely, the floating random walk or the Markov chain method). Two, separate Neuro-Monte Carlo systems are used to solve Laplace's and Poisson's equations Sol homogeneous and inhomogeneous problems. Neuro-Monte Car lo solutions reflect higher computation speed without loss of accuracy ou significant "down time" with respect to the time needed to ti ain the ANN with the data obtained from the partial solution using a Monte Carlo method. (C) 1997 The Franklin Institute. Published by Elsevier Science Ltd.