•Starting with a problem on a fine grid ( b(i), x(i) ), we want to damp the high frequency error by coarsening the grid, and improve the low frequency error by a smoothing operator.
•First a cut at the high frequency error by smoothing/approximate solution:
• x(i) = S(i)( b(i), x(i) ).
•Coarsen the grid: R(i)( b(i), x(i) ).
•Solve the coarser problem recursively
• MGV( R(i)( b(i),
x(i) ) ).