Cascade PID Controller Optimization Using Bison Algorithm

Abstract

Meta-heuristic algorithms are reliable tools for modern optimization. Yet their amount is so immense that it is hard to pick just one to solve a specific problem. Therefore many researchers hold on known, approved algorithms. But is it always beneficial? In this paper, we use the meta-heuristics for the design of cascade PID controllers and compare the performance of the newly developed Bison Algorithm with well-known algorithms like the Differential Evolution, the Genetics Algorithm, the Particle Swarm Optimization, and the Cuckoo Search. Also, in the proposed approach, the controller parameters were encoded to increase the chance of reducing the controller structure, and thus facilitate the automatic selection of its configuration. The simulations were performed for three different control problems and checked whether the use of cascade structures could bring significant benefits in comparison to the use of classic PID controllers.