mutation from +1 or 1 to 0 is very small since it is almost impossible you forget someone
that you have known previously in your group. With the presence of mutation or the
change of sentiment relation sign, we obtain the new balance index every mutation. The
new value of adjacency matrix’s component (Rij = Rji) that gives higher balance index will
be adopted and mutation will re-implement from this new relation pattern, until it
reaches configuration that will not change (balance configuration), while every mutation
resulting lower global balance index will not be adopted.
3.2 Simulations & Discussions
We do several simulations based on the model above. In the first simulation, we
use the model described above with 8 agents (N=8) and randomize an adjacency matrix
(8 x 8) as initial state. After 350 iterations, we found out that the group has been in its
balance state; the index balance is higher now, from 0.0714 to 1. The interpersonal
relations are not changed any more since they have gained maximum balance index.
Table 6
The change of edges string (dyadic relation) after 350 iterations
1 ÷ 2 |
0 |
÷ |
1 |
1 ÷ 3 |
1 |
÷ |
1 |
1 ÷ 4 |
1 |
÷ |
-1 |
1 ÷ 5 |
1 |
÷ |
1 |
1 ÷ 6 |
-1 |
÷ |
1 |
1 ÷ 7 |
1 |
÷ |
1 |
1 ÷ 8 |
0 |
÷ |
-1 |
2 ÷ 3 |
1 |
÷ |
1 |
2 ÷ 4 |
-1 |
÷ |
1 |
2 ÷ 5 |
0 |
÷ |
1 |
2 ÷ 6 |
1 |
÷ |
1 |
2 ÷ 7 |
1 |
÷ |
1 |
2 ÷ 8 |
-1 |
÷ |
-1 |
3 ÷ 4 |
0 |
÷ |
1 |
3 ÷ 5 |
0 |
÷ |
1 |
3 ^ 6 |
1 |
÷ |
1 |
3 ÷ 7 |
0 |
÷ |
1 |
3 ÷ 8 |
0 |
÷ |
-1 |
4 ÷ 5 |
0 |
÷ |
1 |
4 ÷ 6 |
0 |
÷ |
1 |
4 ÷ 7 |
-1 |
÷ |
1 |
4 ÷ 8 |
0 |
÷ |
-1 |
5 ÷ 6 |
1 |
÷ |
1 |
5 ÷ 7 |
-1 |
÷ |
1 |
5 ÷ 8 |
1 |
÷ |
-1 |
6 ÷ 7 |
1 |
÷ |
1 |
6 ÷ 8 |
-1 |
÷ |
-1 |
7 ÷ 8 |
0 |
÷ |
-1 |