On the mean convergence time for simple genetic algorithms
T. Niwa, Masaru Tanaka · 2002
In this paper, we describe genetic drift by a Markov model arid give the mean couvergeiice time under genetic drift in tlie context of GAS. Our arialyses of GAS niakll use of tlie blarkov cliain formalisni based on the Wright-Fisher model which is a typical iiiodel in genetics. Tlie states of tlie Markov chain are giveii. Tlieii, we consider the mean coiivcrgeiicc time to the absorbing states of the Wright-Fisher model. Finally, we derive an optimal iiiutatioii rate for GAS and show that this rate works effectively.