An Analysis of Island Models in Evolutionary Computation

Island Models (IM) are a class of distributed evolutionary algorithms (EAs) in which the population is split into multiple sub-populations called islands. Separate EAs run independently on each island, but they interact by means of migrating individuals. Therefore, IMs are different both from a single-population standard EA, as well as from a set of multiple isolated EAs.

IMs are interesting for several reasons. They have been reported to yield better results than standard EAs. IMs are also advantageous when computational tasks must be distributed across multiple machines because their structure is easy to parallelize. However, despite many studies, no comprehensive theory describing their behavior has been developed. Due to the lack of theory and a complex architecture with many control parameters, setting up IMs has been a trial-and-error process, guided mostly by "rules of thumb."

In this thesis, I adopt a two-level (intra- and inter-island) view of IMs and show how this approach simplifies our understanding of their dynamics. They behave very differently than standard EAs, and in order to take full advantage of this, I propose shifting balance towards the inter-population level of evolution. Consequently, IMs support compositional evolution and reduce random fixation of genes, for which I give examples.

The two levels of evolution influence each other, and I analyze this interaction more deeply. Migrations profoundly change the local dynamics and stimulate evolution, which often ultimately results in better performance. I study the role of genetic operators in this behavior and also create mathematical models of after-migration dynamics. This analysis gives us better understanding of mixing and survival of genes locally, and these processes in turn determine the type and level of interaction between islands globally. Further, using island heterogeneity enhances the inter-island evolution. Following the study, I analyze IM behavior on a range of test problems, including two complex domains.

This thesis improves our understanding of the dynamics of IMs and suggests a qualitative change in the way we think about them. This perspective results in new guidelines for configuring IM parameters and opens new directions for future work.