77

2.4 SUCCESSIVE OVERRELAXATION (SOR) METHOD

We shall now consider SOR method for the system

Ax = y ..(I)

We take a scalar parameter 0 and multiply both sides of (I) to get an equivalent
system,

Ax = y (II)

Now as before we split the given matrx as

( ) U L D A + + =

We write (II) as

(D + L + U)x = y,

i.e.

(D + L) = - Ux + y

i.e.

(D + L)x + (-1) Dx = - Ux +y

i.e.

(D + L)x = - [( 1)D + U]x + y

i.e.
x = - (D + L)
-1
[(-1)D + U]x + [D + L]
-1
y.

We thus get the SOR scheme as

( ) ( )
( )
1
0

zero vector; initial guess

k k
x M x y
x

+
= +
=
(III)

where,

( ) ( ) [ ] u D L D M

+ + =

1
1



78
and
( ) y L D y
1

+ =

M

is the SOR matrix for the system.

Notice that if = 1 we get the Gauss Seidel scheme. The strategy is to choose such
that , 1 < is M
sp

and is as small as possible so that the scheme converges as rapidly as

possible. This is easier said than achieved. How does one choose ? It can be shown
that convergence cannot be achieved if 2. (We assume > 0). Usually is chosen
between 1 and 2. Of course, one must analyse
sp
M

as a function of and find that

value
0
of for which this is minimum and work with this value of
0
.

Let us consider an example of this aspect.

Example 6:

Consider the system given in example 5 in section 2.3.

For that system,

M

= - (D + L)
-1
[(-1) D +U]


|
|
|
|
|
|
|
|
|
.
|

\
|
+ + +
+ +
+

=
2 3 2 4 3 2 4 3
2 3 2 3 2
2 2
4
1
1
8
1
2
1
2
1
16
1
4
1
4
1
8
1
8
1
2
1
4
1
1
8
1
2
1
2
1
4
1
4
1
0
2
1
4
1
1
2
1
2
1
0 0
2
1
1






and the characteristic equation is

( ) ( ) ( )

M
C ....... .......... 0 1 12 1 16
2 4
2
2
4
= + + +

Thus the eigenvalues of M

are roots of the above equation. Now when is = 0 a root?

If = 0 we get, from (C
M
),

16(-1)
4
= 0 = 1, i.e. as in the Gauss Seidel case. So let us take 1; so
= 0 is not a root. So we can divide the above equation (C
M
) by
4

2
to get


79
( ) ( )
0 1
1
12
1
16
2
2
2
2
2
= +
+

+






Setting

( )

2
2
2
1+
=
we get

0 1 12 16
2 4
= +

which is the same as (C
J
). Thus

. 8090 . 0 ; 3090 . 0 =

Now

( )
2
2
2
1

=
+
= 0.0955 or 0.6545 .(*)

Thus, this can be simplified as

( ) ( )
2
1
2 2 2 2
1
4
1
1
2
1
)
`

=

as the eigenvalues of M

.

With = 1.2 and using the two values of
2
in (*) we get,

= 0.4545, 0.0880, -0.1312 i (0.1509).

as the eigenvalues. The modulus of the complex roots is 0.2

Thus
sp
M

when = 1.2 is 0.4545

which is less that
sp
J = 0.8090 and
sp
G = 0.6545 computed in Example 5 in section
2.3. Thus for this system, SOR with = 1.2 is faster than Jacobi and Gauss Seidel
scheme.


80
We can show that in this example when =
0
= 1.2596, the spectral radius
0
M is
smaller than

M for any other . We have

2596 . 1
M = 0.2596

Thus the SOR scheme with = 1.2596 will be the method which converges fastest.

Note:
We had
1 . 2
0 . 4 5 4 5
s p
M =
and

1.2596
0.2596
sp
M =

Thus a small change in the value of brings about a significant change in the spectral
radius
sp
M

.