Abstract
In this paper, a prolonged exploration ability of distance based parameter adaptation is subject to a test via clustering analysis of the population in Success-History based Adaptive Differential Evolution (SHADE). The comparative study is done on the CEC 2015 benchmark set in two dimensional settings – 10D and 30D. It is shown, that the exploration phase of distance based adaptation in SHADE (Db_SHADE) lasts for more generations and therefore avoids the premature convergence into local optima.
Author supplied keywords
Cite
CITATION STYLE
Viktorin, A., Senkerik, R., Pluhacek, M., & Kadavy, T. (2018). On the prolonged exploration of distance based parameter adaptation in SHADE. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10841 LNAI, pp. 561–571). Springer Verlag. https://doi.org/10.1007/978-3-319-91253-0_52
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
