Fuzzy Logic Augmentation of the Multiverse Optimizer Applied to Fuzzy Controllers Design
Lucio Amézquita, Oscar Castillo, José Soria and Prometeo Cortes-Antonio
In this work, we propose the use of fuzzy logic for dynamic parameter adaptation in the Multiverse Optimizer algorithm. For this study, we use the typical unimodal and multimodal benchmark functions for comparing results with respect to the original algorithm, and for the control problems, we use some benchmark fuzzy controllers. In the systems that were studied, first we use a problem that involves system identification, this is the case of the tipper problem, which is compared with the original algorithm; then, we proceed to use some control problems, one is cruise control, which focuses on achieving a desired speed on a vehicle with a certain weight; from here, we use the temperature control in a shower, which controls the temperature of the water by adjusting the water valves, another control problem is the inverted pendulum, which is one of the most common test control problems used for optimization in fuzzy logic controllers. The last control problem is the autonomous mobile robot, which consists in controlling the direction of the robot by adjusting the torque on its wheels. In all of these problems we optimize the membership functions for fuzzy controllers. The main goal of this study is to analyze whether Fuzzy Logic can improve the Multiverse Optimizer performance in certain cases by dynamically adjusting some parameters of the algorithm so it can be more competitive on more complex problems.
Keywords: Multi-verse optimizer, fuzzy logic, optimization, dynamic parameter, cruise control, shower, inverted pendulum, robot