In this section, we discuss how to develop a working fixed point algorithm for (3.1) that not only is much faster than the explicit time- marching methods reviewed in section 3, but also has proved to be always convergent as a stand-alone algorithm in all of the simulations we have carried out. Our aim, however, goes further and is oriented to developing an even faster MG algorithm; with this in mind we will not use our fixed point algorithm as a stand-alone method but as a smoother for such a nonlinear MG algorithm