Information Technology Reference
In-Depth Information
for SISO FLB1, and
R
1
:
v
,
IF
i
IS N THEN
r
e
2
IS PS
R
2
:
IF
i
v
IS PS THEN
r
e
2
IS PB
,
(8.30)
R
3
:
IF
i
v
IS PB THEN
r
e
2
IS PS
,
for SISO FLB2. The parameter settings suggested for SISO FLB1 and SISO FLB2
to ensure a trade-off to the convergence speed and to the number objective function
evaluations are (David et al.
2012
)
B
i
v
PS
=
0
.
3
,
B
i
v
PB
=
0
.
8
,
B
e
1
PS
=
0
.
15
,
B
e
1
PB
=
0
.
2
,
B
e
2
PS
=
0
.
45
,
B
e
2
PB
=
0
.
5
.
(8.31)
Mamdani's MAX-MIN compositional rule of inference is used in the inference
engines of SISO FLB1 and SISO FLB2. The defuzzification for both SISO FLB1
and SISO FLB2 is carried out by the center of gravity method for singletons.
8.4 Adaptive Charged System Search Algorithms
The modeling of standard CSS algorithms is conducted in the framework of a multi-
agent approach where agents are represented by charged particles (CPs) with the
magnitude (Kaveh and Talatahari
2010a
,
b
,
c
)
q
i
=
(
f
i
−
f
best
)/(
f
best
−
f
),
i
=
1
...
N
,
(8.32)
w
or st
where
f
best
and
f
or st
are the best and worst fitness of all CPs. The separation
distance between two CPs is
w
X
j
)/
X
best
+
ϕ
,
R
N
,
(8.33)
where
X
best
is the position vector of the best current CP, and the relative small
parameter
r
ij
=
(
X
i
−
0
.
5
(
X
i
+
X
j
)
−
X
i
,
X
j
,
X
best
∈
0 is introduced to avoid singularities. The value of the resultant
electrical force acting on
j
th CP is
ϕ>
N
q
i
c
ij
(
a
3
r
ij
)(
F
j
=
q
j
r
ij
i
1
/
+
i
2
/
X
i
−
X
j
)
,
i
=
1
,
i
=
j
i
1
=
0
,
i
2
=
1
⇔
r
ij
≥
a
,
i
1
=
1
,
i
2
=
0
⇔
r
ij
<
a
,
j
=
1
...
N
,
(8.34)
−
1if
f
i
<
f
j
,
c
ij
=
1
otherwise
,
Search WWH ::
Custom Search