Abstract
AbstractWe introduce the concept of Semantic Neutral Drift (SND) for genetic programming (GP), where we exploit equivalence laws to design semantics preserving mutations guaranteed to preserve individuals’ fitness scores. A number of digital circuit benchmark problems have been implemented with rule-based graph programs and empirically evaluated, demonstrating quantitative improvements in evolutionary performance. Analysis reveals that the benefits of the designed SND reside in more complex processes than simple growth of individuals, and that there are circumstances where it is beneficial to choose otherwise detrimental parameters for a GP system if that facilitates the inclusion of SND.
Funder
Engineering and Physical Sciences Research Council
Publisher
Springer Science and Business Media LLC
Subject
Computer Science Applications
Reference40 articles.
1. Atkinson T, Plump D, Stepney S (2018a) Evolving graphs by graph programming. In: Proceedings of 21st European conference on genetic programming (EuroGP 2018), LNCS, vol. 10781. Springer, pp 35–51. https://doi.org/10.1007/978-3-319-77553-1_3
2. Atkinson T, Plump D, Stepney S (2018b) Probabilistic graph programs for randomised and evolutionary algorithms. In: Proceedings of 10th international conference on graph transformation, LNCS, vol. 10887. Springer, pp 63–78. https://doi.org/10.1007/978-3-319-92991-0_5
3. Banzhaf W (1994) Genotype–phenotype-mapping and neutral variation—a case study in genetic programming. In: Proceedings of 3rd international conference on parallel problem solving from nature, LNCS, vol. 866. Springer, pp 322–332. https://doi.org/10.1007/3-540-58484-6_276
4. Barnett L (1998) Ruggedness and neutrality; the NKp family of fitness landscapes. In: Proceedings of 6th international conference on artificial life. MIT Press, pp 18–27. http://dl.acm.org/citation.cfm?id=286139.286143
5. Chickering DM (2002) Learning equivalence classes of Bayesian-network structures. J Mach Learn Res 2:445–498
Cited by
5 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献