Parameters optimization applying Monte Carlo methods and Evolutionary Algorithms. Enforcement to a trajectory tracking controller in non-linear systems

In this work, a closed-loop control strategy is proposed. It allows tracking optimal profiles for a fed-batch bioprocess. The main advantage of this approach is that the control actions are computed from a linear equations system without linearizing the mathematical model, which allows working in any range. In addition, three techniques are developed to tune the controller. First, a completely probabilistic method, Monte Carlo. Second, a methodology based on Genetic Algorithms, an evolutionary optimization technique. Third, a Hybrid Algorithm, combining above algorithms advantages. Here, the objective function is to find the controller parameters that minimize the trajectory tracking total error. The controller performance is evaluated through simulations under normal operations conditions and parametric uncertainty, using the obtained controller parameters.