Databases Reference
In-Depth Information
Initialising
Evaluation
Selection
Cross Over
Mutation
Service
Procedures
Stopping
Conditions
(a)
(b)
Fig. 10.4.
(a) Block diagram of GA. (b) Pseudo code of GA.
stopping condition is reached, e.g., number of generation or time. The terms
such as selection, cross over, mutation used in both Figs. 10.4(a) and (b)
are discussed comprhensively in Mitchell.
6
However, we briefly present them
here for completeness and ready reference.
•
Selection: The selection operator is used to choose cromosomes from a
population for mating. This mechanism defines how these cromosomes
will be selected, and how many offsprings each will create. The
expectation is that, like in the natural process, cromosomes with higher
fitness will produce better offsprings. The selection has to be balanced:
too strong selection means that best cromosome will take over the
population reducing its diversity needed for exploration; too weak
selection will result in a slow evolution. Some of the classic selection
methods are Roulette-wheel, Rank based, Tournament, Uniform, and
Elitism.
6
•
Crossover: The crossover operator is practically a method for sharing
information between two cromosomes; it defines the procedure for
generating an offspring from two parents. The crossover operator is
considered the most important feature in GA, especially where building
blocks exchange is necessary. One of the most common crossover operator
is Single-point crossover: a single crossover position is chosen at random
