A Methodology for Tuning and Optimization of a Quadruple Tank Control System Using Firefly Algorithm in Wavelet Space

Using more efficient tuning techniques becomes imperative, due to the increasing competitiveness in the industry. With this propose, meta-heuristics, such as Firefly Algorithm (FA), can be used to obtain the parameters of the controller according to a cost function, which should encode how good a controller is, adequately expressing the desired specifications, so that the metaheuristic employed can find the desired controller that is able to reach the response wanted. The methods traditionally used for automatic tuning of controlers present difficulties in expressing the desired specifications, being able to mapping the desired search space and allowing that the algorithm finds the proper answer. These difficulties is more evident when more complex controllers are required, as for Multiple Input Multiple Output (MIMO) problems. Aiming to solve these difficulties, a methodology using wavelet transform to describe the behavior of a controller response and its use for obtain better performance of the optimization algorithm. A case study will be done using the quadruple tank system, showing the efficiency of the methodology proposed.