Abstract:
Within the framework of the linear theory of accuracy the paper leads to a system of equations for errors in analog computer solution of nonlinear finite equations by the gradient method asto estimate probabilistic characteristics of the errors. Expressions for these estimates are given.