Large population sizes and crossover help in dynamic environments

Abstract Dynamic linear functions on the boolean hypercube are which assign to each bit a positive weight, but weights change over time. Throughout optimization, these maintain same global optimum, and never have defecting local optima. Nevertheless, it was recently shown [Lengler, Schaller, FOCI 2019] that $$(1+1)$$ ( 1 + ) -Evolutionary Algorithm needs exponential time find or approximate optimum for some algorithm configurations. In this experimental paper, we study effect of larger population sizes dynamic binval , extreme form functions. We moderately increased extend range efficient configurations, crossover boosts substantially. Remarkably, similar static setting monotone in Zou, FOGA 2019], hardest region optimization $$(\mu +1)$$ μ -EA is not close far away from it. contrast, -GA, around all studied cases.Kindly check confirm inserted city name correctly identified.Correct.