A genetic algorithm is a form of optimization algorithm based on biological evolution. A candidate solution is represented as a string of parameters called a "chromosome". The algorithm works by starting with an initial chromosome, or pool of chromosomes. It runs iteratively, and in each iteration it:

1) Introduces random mutations in the population.
2) Creates new chromosomes by combining existing ones as in sexual reproduction (called "crossover").
3) Performs natural selection by eliminating those chromosomes that do not satisfy a certain threshold according to a predefined "fitness function".

After a given number of iterations, the fittest chromosome is selected to be the optimized solution.
Source(s):
Master's degree in computer science.

Permalink | Report