Migration, Selection Pressure, and Superlinear Speedups

Abstract

The choice of migrants and the replacement of individuals are not often considered important parameters of parallel GAs. However, this chapter used two different methods to show that choosing the migrants or replacements according to their fitness increases the selection pressure. Some migration policies may cause the algorithm to converge significantly faster. The migration policy that accelerates convergence the most is to choose both the migrants and the replacements according to their fitness, which is also the most common policy.

The faster convergence may explain some of the claims of superlinear speedups in parallel GAs. This chapter showed an example where serial and parallel algorithms reached the same solution and used the same number of individuals, but the additional selection pressure resulted in superlinear speedups.

The chapter also included calculations of the higher moments of the distribution of fitness. These calculations showed that different combinations of the degree of the topology and the migration rate affect the population in different ways, even if they result in the same selection intensity.