next up previous contents
Next: Fitness Up: The Genetic Algorithm Previous: The Individual   Contents

Subsections

The Population

The genetic algorithm doesn't work on a single individual but on a whole population ${\bf P}$ of $p$ individuals which undergoes an evolutionary process starting with the initial Population ${\bf P}_0$.

The Initial Population

The simplest way to create ${\bf P}_0$, is by simply generating $p$ random strings of length $l$. However, it is also possible to generate the phenotypes of the individual and then translating them into their corresponding genotypes. If certain parameter-combinations of the phenotype tend to be more successful, this knowledge can be used to improve the quality of ${\bf P}_0$ and lead to a faster convergence of the algorithm. However this will cut down the probability of finding totally different combination which might perform even better.


Decimation

A better method of improving the quality of the initial population is decimation. A population $\bar{\bf P}_0$ of the size $p d$ is created at random of which the best (i.e. the fittest) $p$ individuals are selected into ${\bf P}_0$. The factor $d$ is called decimation factor.


next up previous contents
Next: Fitness Up: The Genetic Algorithm Previous: The Individual   Contents

(c) Bernhard Ömer - oemer@tph.tuwien.ac.at - http://tph.tuwien.ac.at/~oemer/