Multi-objective optimization for the tuning of the control system of a quadruple tank structure

This paper presents the application of multi-criteria optimization techniques for the adjustment and tuning of the PI controllers used for the control of the quadruple tank system. Specifically, it is proposed to use the technique of decoupling the control loops to tune each of them independently. An evolutionary algorithm was used as optimizer, specifically of differential evolution, for its practicality and effectiveness. For the analysis phase of the possible solutions, the Pareto front representation is used in conjunction with the calculation of the area as a function of the Nadir to discriminate a subset of optimal solutions. The work does not focus on analyzing different controller structures for the quadruple tank system, but rather, given a control configuration, analyzing how to maximize the possibilities of this structure to meet objectives that are contrary and were proposed in the design parameters. One of the advantages of using multi-criteria techniques is the ease of analyzing the conflicts between the objectives of the problem, making it possible to select an appropriate response solution according to the designer's criteria, taking the whole solution set as a reference.