The problem of aggregating multiple outcomes to form overall objective functions is of considerable importance in many applications. The ordered weighted averaging (OWA) aggregation uses the weights assigned to the ordered values (i.e., to the largest value, the second largest and so on) rather than to the specific coordinates. It allows to evaluate solutions impartially, when distribution of outcomes is more important than assignments these outcomes to the specific criteria. This applies to systems with multiple independent users or agents, whose objectives correspond to the criteria. The ordering operator causes that the OWA optimization problem is nonlinear. Several MILP models have been developed for the OWA optimization. They are built with different numbers of binary variables and auxiliary constraints. In this paper we analyze and compare computational performances of the different MILP model formulations.