In activated sludge waster water plants, the pollution (ammonium,phosphate) is removed by microorganisms. The switching on/off of theaeration controls the switching between different states (aerobic,anoxic, anaerobic) where different complementary processes take place. We are interested in the optimal control of the aeration with different antagonist targets: minimize the remaining pollutants and the operating cost (aeration duration). As a first step, an ODE model was established and the parameters were estimated based on data collected on-site. This parameter estimation is obtained using a variational data assimilation strategy. In a second step, a feedback control is defined and its efficiency is quantified by estimating the distance of the corresponding control to the Pareto set. This distance to the Pareto set is obtained 1) using an augmented Lagrangian method 2) using an on-shelf genetic algorithm.