Bugfix Multigrid

[TimoH447]

Check convergence

[TimoH447]

Possible Areas of mistake:

• Smoothing: Jacobi on its own converges, thus seems correct
• Prolongation
• Restriction
• Gaussian elimination: Probably right, tested by doing multigrid with only one level for small N = > solution is exact in this case (no error)

[TimoH447]

Prolongation and Restriction had index error which are now corrected. But the multigrid is still not converges as fast as expected

[TimoH447]

Issue will be resolved with "debug" branch.

At the end what needed to be changed was:

• missing braces which led to running multigrid with grid size N which did not work for the number of levels
• changed the h=1/(N+1) from "being" on the right hand side to the matrix of the poisson problem
• main problem: bad convergence of jacobi -> fixed by using weighted jacobi

Due to chaotic debugging we added a lot bugs which we had to again "refix". But now we get the expected fast convergence with usually <20 v-cycles.