The SLOR equations can be solved as they are a set of NNAS tridiagonal equations soluble in 5NNAS operations |
One can now use the usual over relaxation procedure |
Note that upto a constant SLOR has the same complexity as original SOR |
For Lapace's equation, one can show that if you choose the appropriate optimal w for both that SLOR is twice as fast as SOR -- as a SLOR calculation is somewhat more computation, one will not gain this full factor |