A blending function method with interior contraints (Q1195960)

When large dimensional systems are solved iteratively, it is usual to put all the unknowns interior to the domain equal to the same constant and then to assign the boundary unknowns according to the boundary conditions. Some years ago some techniques based on the use of blending functions have been discovered in order to obtain a good initial approximation. In this paper the author applies these methods to obtain an initial approximation in a particular case: the Poisson problem with Dirichlet boundary conditions on a rectangular domain. The approach can be considered for a parallel implementation.