Differential evolution, diversity of population, adaptation of population size.

Differential evolution (DE) has become one of the most used optimization techniques. DE is population-based algorithm with several control parameters. There are many adaptive DE variants working with mutation and crossover parameters proposed since the DE algorithm has been introduced. In this work, we study population-size adaptive mechanism proposed recently which is based on the population diversity. In this mechanism, the population size is increased by one (a newly randomly generated point is added) or decreased by one (the worst point is erased) dependently on actual diversity of population. This mechanism is investigated in the original version of DE algorithm by applying to benchmark set developed for CEC2013 competition.