2.2 Evaluative Crossover
For evaluation purpose, this study applies the crowding distance as an evaluation of
chromosome’s quality in the evaluative crossover. The crowding distance proposed in
NSGA-II is used to estimate the density quantity of a particular solution in the population
by calculating the average distance between other surrounding solutions with respect to
each objective (Deb et al., 2002). After two parents have been selected from population, let
the parent with larger crowding distance be named as the better parent ( b
x
) and the other
one is the worse parent (
w
x
). Their crossover child is denoted as y
.
The proposed evaluative crossover imitates the gene-therapy process at the forefront of
medicine, which inserts genes into an individual's cells to treat a disease by replacing a
defective mutant allele by a functional one. Therefore, the evaluative crossover integrates a
gene-evaluation method with a gene-therapy approach in the traditional uniform crossover
scheme. By randomly generating a therapeutic mask with the same length as chromosomes,
each parity bit in the therapeutic mask indicates whether the gene locus should be cured or
not. For each gene locus, a random number in the interval [0, 1] is generated and compared
to a pre-defined crossover rate pc. If the random number is larger than the crossover rate, the
parity bit in the therapeutic mask is assigned as 0 and no crossover occurs at this locus
(iGc). Otherwise, the parity bit in the therapeutic mask is assigned as 1 and the child’s gene
is generated by the gene-therapy approach (iGc).
Firstly, the gene-evaluation method mutually exchanges two parity genes between two
mating parents and then compares their fitness variance as a measurement of these genes’
merit. The exclusive features of the gene-evaluation method include that 1) the contribution
www.intechopen.com