Nonlinear method for solving the stationary quasi-diffusion difference equation

The two dimensional difference task is reduced to sequence of one dimensional ones by new manner. Nonlinear transformations of difference equations accelerates the iterations convergence. The choice of optimal iterational parameter provides linear dependence of the number of iterations on grid dimension. The self-adjointness of operator and information of its spectrum are not necessary. This method runs with strong heterogeneity and cavities.