Runtime Analysis of Evolutionary Algorithms via Symmetry Arguments
share
We use an elementary argument building on group actions to prove that the selection-free steady state genetic algorithm analyzed by Sutton and Witt (GECCO 2019) takes an expected number of Ω(2^n / √(n)) iterations to find any particular target search point, regardless of the population size μ. This improves over the previous best lower bound of Ω((n^δ/2)) valid for population sizes μ = O(n^1/2 - δ), 0 < δ < 1/2.
READ FULL TEXT