MALE AND FEMALE CHROMOSOMES IN GENETIC ALGORITHMS

Ghodrat Moghadampour

2010

Abstract

Evolutionary algorithms work on randomly generated populations, which are converged over runs toward the desired optima. Randomly generated populations are of different qualities based on their average fitness values. In many cases switching all bits of a randomly generated binary individual to their opposite values might quickly produce a better individual. This technique increases diversity among individuals in the population and allows exploring the search space in a more rigorous way. In this research the effect of such operation during the initialization of the population and crossover operator has been investigated. Experimentation with 44 test problems in 2200 runs showed that this technique can facilitate producing better individuals on average in around 32% of cases.

Paper Citation

in Harvard Style

Moghadampour G. (2010). MALE AND FEMALE CHROMOSOMES IN GENETIC ALGORITHMS . In Proceedings of the 12th International Conference on Enterprise Information Systems - Volume 2: ICEIS, ISBN 978-989-8425-05-8, pages 220-225. DOI: 10.5220/0002897702200225

in Bibtex Style

@conference{iceis10,
author={Ghodrat Moghadampour},
title={MALE AND FEMALE CHROMOSOMES IN GENETIC ALGORITHMS},
booktitle={Proceedings of the 12th International Conference on Enterprise Information Systems - Volume 2: ICEIS,},
year={2010},
pages={220-225},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0002897702200225},
isbn={978-989-8425-05-8},
}

in EndNote Style

TY - CONF
JO - Proceedings of the 12th International Conference on Enterprise Information Systems - Volume 2: ICEIS,
TI - MALE AND FEMALE CHROMOSOMES IN GENETIC ALGORITHMS
SN - 978-989-8425-05-8
AU - Moghadampour G.
PY - 2010
SP - 220
EP - 225
DO - 10.5220/0002897702200225