Information Technology Reference
In-Depth Information
Table 2.
Comparison of Memorizing Capacity of
GMACA
and Hopfield Net
Size of Memorizing Capacity of Network
Pattern (
n
)
GMACA
Hopfield Net
10
4
2
20
5
3
30
7
5
40
10
6
50
12
8
60
13
9
70
15
11
80
18
12
90
20
14
100
23
15
Table 3.
Evolution time for Synthesize GMACA
Size of No of Initial Evolution
Pattern (
n
) Patterns (
k
) Temp (
T
) Time (min)
10
4
15
0.43
20
5
15
1.06
30
7
15
1.55
40
10
20
3.01
50
12
25
3.35
60
13
25
4.52
70
15
30
7.03
80
18
35
7.45
90
20
30
9.21
100
23
40
15.08
The
Table 2
demonstrates the potential of
CAM
as an associative memory
model.
Column II
of
Table 2
depicts the maximum number of patterns that an
n
-cell
GMACA
can memorize. The results of Hopfield Net on the same data set
are provided in
Column III
for the sake of comparison. The experimental result
clearly indicates that :
(i)
the memorizing capacity of
GMACA
is found to be more than
20% of its lattice size
; and
(ii)
it is superior to conventional Hopfield Net by 33%
.
4.2 Evolution Time
The
Table 3
represents the evolution time to evolve
GMACA
by
Simulated An-
nealing
.
Column I
and
II
of
Table 3
represent different
CA
size (
n
) and number of
attractors (
k
) respectively; while the
Column III
depicts the
Initial Temperature
required to find out the best possible
GMACA
configuration by
SA
.In
Column
IV
, we provide the evolution time required to synthesize
GMACA
. Growth of
time, as the results indicate, is super linear in nature.
