We propose and investigate two Markov chain models in order to provide recommendations on the selection of genotype redundant encoding schemes in genetic algorithms (GA). It was deduced from these Markov chains why classic encoding methods lead to a insufficiently broad search of GA’s. Then, modifications, concerning additional redundancy in genotype encoding algorithms, are proposed and investigated regarding its influence on mutations and recombinations. Finally, the results of extensive simulation studies are reported that indicate classes of goal functions for which these modifications provide better search results and/or smaller variances of them.
CITATION STYLE
Rafajłowicz, W. (2020). A Markov Process Approach to Redundancy in Genetic Algorithms. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 12415 LNAI, pp. 445–453). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-61401-0_41
Mendeley helps you to discover research relevant for your work.