Genetic Programming - based Controller Design
Ivan Sekaj, J. Perkacz · 2007
Two genetic programming-based approaches are proposed for continuous-time process control design. The first approach uses a network of interconnected continuous-time elementary dynamic building blocks with a specific table-based representation of individuals. In the second approach the control algorithm is represented as a recurrent function of discrete-time variables with the tree representation of individuals. Both approaches are demonstrated on controller design examples, which are compared with the genetic algorithm-based PID controller design.