Travis, > Sorry for the inconvenience but I still cannot figure it out.The > algorithm my professor is working on is an implementation of a multiscale > finite element method.The key step of the multiscale finite element method > is to construct the basis function in each element. To construct the basis > function, we need to solve a local problem in each element. That is to say, > when we carry on the finite element method, the basis function in each > element is obtained by solving the local problem. I have a triangulation > object and it has been refined.Now I only need to solve an equation on one > of its cells and the solution will be regarded as the basis function of an > element in that cell. Is there an easier way other than associating > FE_Nothing elements to other cells?I wonder if active cell_iterator could > be somhow transformed into a triangulation object such that I can solve the > equation using the > techniques step-6 provide with only changes in the Coeffeicient > function.
The question is whether you need to solve for different shape functions on each cell, or whether the local problem you need to solve is the same for all cells? In either case, the programs I know of where people have tried this before use two Triangulation/DoFHandler objects. One describes the reference cell and fine scale problem, and the other one describes the real problem and uses the shape functions computed on the reference cell. Best W. ------------------------------------------------------------------------- Wolfgang Bangerth email: [email protected] www: http://www.math.tamu.edu/~bangerth/ _______________________________________________ dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
