Multi-population Genetic Algorithms with Variational Search Areas

GONG Dun-wei,SUN Xiao-yan
DOI: https://doi.org/10.3969/j.issn.1000-8152.2006.02.018
2006-01-01
Abstract:A multi-population genetic algorithm with variational search areas is proposed to cope with the limitation of island multi-population genetic algorithms.Firstly,the condition and strategy for changing search areas dynamically are presented according to the distribution of the best individuals of all populations.Secondly,an adaptive adjustment method of the population size is given based on the search areas' measure and the search granularity.Thirdly,the algorithm's performance is analyzed quantitatively from the views of the search areas' measure and the population size.Finally,the algorithm's efficiency is validated through optimization of two benchmark functions.
What problem does this paper attempt to address?