Overview on simulation optimization methods

evaluate the configurations performances (i. e. perform a number of simulation runs) 

determine non-dominated configurations of the population and assign fitness values 

create a new population by applying genetic operators (mutation and cross-over) on the subset of non-dominated configurations with best fitness scores 

repeat step 2 and 4 until the set of non-dominated configurations is sufficiently large. 

In contrast to single-objective optimization, in multi-objective optimization two goals have to be achieved  

---