Information Technology Reference
In-Depth Information
e) Calculate
J
(
i
,
j
+
1
k
,
l
),
i
J
(
i
,
j
+
1
k
,
l
)
=
J
(
i
,
j
+
1
k
,
l
)
+
J
(
am
(
j
+
1
k
,
l
),
p
(
j
+
1
k
,
l
)).
cc
f) swim: let
m
=0 (counter for swim length)
m
<
N
While
s
J
(
i
,
j
+
1
k
,
l
)
<
J
If
(if doing better), let
fit
J
fit
=
J
(
i
,
j
+
1
k
,
l
)
and let
i
i
am
(
j
+
1
k
,
l
)
=
am
(
j
+
1
k
,
l
)
+
C
(
i
)
ˆ
(
i
)
am
i
(
j
+
1
k
,
l
)
And use this
to calculate the
J
(
i
,
j
+
1
k
,
l
)
new
as step e) did
Else, let
m
=
N
, this is the end of this while
s
statement.
g) Scheduling vector update: randomly choose a
scheduling vector from the population of
bacterium; generate a new scheduling vector
through crossover operation. Calculate the new
objective function
J
new
(
j
,
k
,
l
)
, if
J
(
j
,
k
,
l
)
<
J
(
j
,
k
,
l
)
,
new
last
update the scheduling vector.
j
<
N
8) if
, go to step 6, continue chemotaxis step.
c
9) Reproduction:
h) For the given
k
and
l
, and for each
i
∈
1
2
,...,
S
,
N
+
1
i
health
be the health of bacterium
i
(a
J
c
let
=
∑
J
(
i
,
j
,
k
,
l
)
j
=
1
measure of how many nutrient it got over its
lifetime and how successful it was at avoiding
noxious substances). Here in workflow scheduling
problem
healt
J
denotes the accumulated makespan
value during the bacterium lifetime. Sort bacteria
in order of ascending cost
health
J
.
i) The half bacteria with higher
healt
J
values die
and the other half bacteria are split (and the
copies that are made are placed at the same
location as their parent).
