Comparative study of diploid and haploid binary genetic algorithms
Genetic Algorithms are generally used to solve a large variety of problems in the real world. As the problems become more difficult some types of algorithm don’t seem to be able to solve them with such ease, or even at all. Problems that contain local optima, or vary with time, are more complex to solve, and resemble real world problems. The ultimate real world problem would be to evolve a species to survive on this planet, one which nature seems to have coped with pretty well. One way to look for guidance when designing a type of genetic algorithm is towards nature, and thus many GA structures exist which have been inspired by the observation of complex biological systems and biochemical interactions. One of these inspirations is the use of a genotype consisting of more than one chromosome which could hold redundant information in abeyance. Yet, if we are trying to create an efficient problem solver, simplified models of biological and biochemical systems might not always prove to be the best. This work will compare current models of genetic algorithms on a variety of problems and will look into ways of extending current models in order to improve their performance.