Performance Measures on Jumping Gene - Multiobjective Optimization Methodology: A Jumping Gene Approach - page 98

Table 5.2
Correspondence.between.the.Parameter.
Index.and.the.Associated.Total.Number.and.
Length.of.Transposons.of.the.Binary-Coded.
JG.for.Test.Functions.POL,.SRIN,.and.TAN

Parameter
Index

Parameter
Index N L N L

1 1 1 15 4 1

2 1 2 16 4 2

3 1 4 17 4 4

4 1 6 18 4 6

5 1 8 19 6 1

6 1 10 20 6 2

7 1 20 21 6 4

8 1 40 22 8 1

9 2 1 23 8 2

10 2 2 24 10 1

11 2 4 25 10 2

12 2 6 26 20 1

13 2 8 27 40 1

14 2 10

both.the.number.and.length.of.transposons.less.than.10..This.verifies.the.man-
ifestation.of.low.transposition.rates.at.which.the.host.remains.unharmed.[11]..
However,. when. both. the. jumping. rate. and. number. and. length. of. transpo-
sons.were.increased.to.a.certain.higher.value,.poor.convergence.was.evident.
On. the. other. hand,. variations. of. the. jumping. rate. and. the. number. and.
length. of. transposons. did. not. affect. the. diversity. performance. greatly.. In.
conclusion,.value.selection.of.the.three.parameters.based.on.the.mentioned.
ranges.is.recommended.for.JG.users.to.acquire.good.performance.

5.7.2 Comparisons with Other MOeas

This.section.compares.the.JG.with.other.commonly.used.MOEAs.described.
in. Chapter. 2 .. To. perform. statistical. comparisons,. 50. simulation. runs. were.
performed.for.each.MOEA.in.each.test.function. . Table 5.3. shows.the.param-
eters.of.MOEAs.utilized.in.the.simulations.
In. addition,. the. number. of. fitness. function. evaluations. (denoted. by.
E = P × G ,.where. P is the.population.size.and. G .is.the.maximum.number.of.
generations. or. iterations). was. varied. for. different. test. functions.. However,.
for. a. particular. test. function,. the. number. of. itness. function. evaluations. E .
was.fixed.for.all.algorithms.to.allow.fair.comparisons..For.example,.referring.
to. Table 5.3 ,. the.number.of.itness.function.evaluations.for.the.unconstrained.
test.function.SCH.was. E = 100 × 20 = 2000.