Rational genetic alterations of a microorganism for a specific purpose are not possible in many situations where our knowledge of the relationship between phenotype and genotype is limited. In such cases evolutionary techniques must be applied. Evolutionary methods are usually time consuming; therefore, more efficient techniques are highly desirable. In this work we present the optimization of strain development in a cytostat. The time required for mutant strain isolation is dependent on the total cells present, the wild-type specific growth rate, the beneficial mutation probability, the mutant specific growth rate, and several bioreactor operating conditions. These parameters are highly related, and a theoretical model, as developed here, is needed to define the conditions that optimize the isolation. The model is based on a discrete, stochastic description of mutant formation and selection in the background of abundant wild-type cells. Using the model, we determined the optimal cytostat operating strategy for mutant isolation that varies according to the probability of beneficial mutations. It is also shown that mutants with as little as a 5% growth advantage can be isolated in less than 15 days which is significantly faster than in a chemostat. The described optimal mutant isolation procedure is expected to be particularly useful for the generation of industrial strains that are robust in challenging growth conditions.