An analysis of genetic drift with selection

摘要

In this paper we use a Markov Chain model for analyzing the effect of drift in the presence of selection. Specifically, we have considered the behavior of the binary integer function under the action of a factorized distribution algorithm and truncation selection. The analysis enables us to estimate the critical population size needed for convergence to the optimum solution, under certain simplifying assumptions. The estimate of the critical population size has been used to formulate an empirical law regarding the optimal truncation threshold.