SOR

Is a variant of the Gauss–Seidel method for solving a linear system of equations, resulting in faster convergence.

​

 SOR is focused on improving the convergence of Gauss seidel by adding a new parameter in the formulations. This parameter is w and there are some instructions in order to determine the response of the algorithm depending of the value taken by w:

1. If w=1 the method is equivalent to Gauss seidel. There is no relaxation

2. If 0<w<, SOR applies a sub-relaxation and it is used for obtaining convergence in some situations in which Gauss Seidel diverges

3. If 1<w<2, SOR applies an over-relaxation and it is used for accelerating the convergence of Gauss Seidel.