Process optimization is a very important point in modern industry. There are many classical optimization methods, which can be applied when some mathematical conditions are verified. Real situations are not very simple so that classical methods may not succeed in optimizing; as in cases when the optimization has several contradictory objectives (Collette, 2002).
The purpose of this work is to propose an optimization method for industrial processes with multiple inputs and multiple outputs (MIMO), for which the optimization objectives are generally contradictory and for which some objectives are not maximum or minimum but performance criteria.
The first step of this method is modeling each process response by a quadratic model. After
establishing the model, we use a simplified numerical optimization algorithm in order to determine
values of the parameters allowing optimizing the different responses, for MIMO processes. This method will also allow finding optimum target values for multiple inputs single output processes.