Immigrant schemes for evolutionary algorithms in dynamic environments: Adapting the replacement rate

Abstract

One approach for evolutionary algorithms (EAs) to address dynamic optimization problems (DOPs) is to maintain diversity of the population via introducing immigrants. So far all immigrant schemes developed for EAs have used fixed replacement rates. This paper examines the impact of the replacement rate on the performance of EAs with immigrant schemes in dynamic environments, and proposes a self-adaptive mechanism for EAs with immigrant schemes to address DOPs. Our experimental study showed that the new approach could avoid the tedious work of fine-tuning the parameter and outperformed other immigrant schemes using a fixed replacement rate with traditionally suggested values in most cases.