Graphics Reference
In-Depth Information
weaving of the global and local search phases allows the two to influence each
other; i.e. SSGA chooses good starting points, and local search provides an accu-
rate representation of that region of the domain. A brief pseudocode of the SSMA
is shown as follows:
Initialize population
While (not termination-condition) do
Use binary tournament to select two parents
Apply crossover operator to create
offspring (
Of f
1
,
Of f
2
)
Apply mutation to
Of f
1
and
Of f
2
Evaluate
Of f
1
and
Of f
2
For each
Of f
i
Invoke Adaptive-
P
LS
-mechanism to
obtain
P
LS
i
for
Of f
i
If
υ(
)
<
P
LS
i
then
Perform meme optimization for
Of f
i
End if
End for
Employ standard replacement for
Of f
1
and
Of f
2
End while
Return the best chromosome
0
,
1
•
COoperative COevolutionary Instance Selection (CoCoIS)
[
72
]—The cooper-
ative algorithm presents the following steps:
Initialize population of combinations
Initialize subpopulations of selectors
While Stopping criterion not met do
For N iterations do
Evaluate population of combinations
Select two individuals by roulette selection
and perform two point crossover
Offspring substitutes two worst individuals
Perform mutation with probability
P
mutation
For M iterations do
Foreach subpopulation i do
Evaluate selectors of subpopulation i
Copy Elitism% to new subpopulation i
Fill (1-Elitism)% of subpopulation
i by HUX crossover
Apply random mutation with probability
P
random
Apply RNN mutation with probability
P
rnn
Evaluate population of combinations
Where
N
and
M
are the number of iterations.
