Overview of Multiobjective Optimization - Multiobjective Optimization Methodology: A Jumping Gene Approach - page 24

START

No

Set the required
parameters and
iteration counter
g = 0

Selection Nominal convergence?

Yes

Crossover

Filter

Generate random population
for population memory

Mutation

Nonreplaceable Replaceable

External memory

Objective value
evaluation

Yes

g = g + 1

g = MAXIMUM
ITERATION?

Elitism

No

Output Pareto-optimal
solutions

Population initialization

New population

Objective value
evaluation

END

Figure 2.11
Flowchart.of.MICROGA.

2.2.6.1 Population Memory

The.population.memory.consists.of.two.parts,.replaceable.and.nonreplaceable..
The. replaceable. part. may. be. modiied. after. each. cycle.. In. contrast,. the. non-
replaceable.part.is.unaltered.during.the.run,.aiming.at.providing.the.required.
diversity.for.the.algorithm.
At.the.beginning.of.each.cycle,.the.population.is.taken.from.both.portions.
of.the.population.memory.so.there.is.a.mixture.of.randomly.generated.indi-
viduals.(nonreplaceable).and.evolved.individuals.(replaceable).

2.2.6.2 Adaptive Grid Algorithm

The.adaptive.grid.algorithm.employed.in.MICROGA.is.similar.to.that.in.the.
PAES..It.offers.diversity.to.nondominated.solutions..Once.the.archive.storing.
the.nondominated.solutions.reaches.its.limit,.the.objective.space.covered.by.
the.archive.is.divided.into.a.number.of.grids..Each.solution.in.the.archive.is.
then.assigned.a.set.of.coordinates.
When. a. new. nondominated. solution. is. generated,. it. is. accepted. if. it. is.
located.at.a.grid.where.the.number.of.stored.nondominated.individuals.is.
smaller.than.that.of.the.most.crowded.grid.or.located.outside.the.previously.
speciied. boundaries.. Note. that. this. adaptive. grid. algorithm. requires. two.
parameters:.the.expected.size.of.the.Pareto.front.and.the.number.of.positions.
in.which.the.solution.space.will.be.divided.for.each.objective.