Design of a Discrete Deadbeat Controller Based on Block Diagram Oriented Genetic Programming

This paper proposes a Genetic Programming representation and algorithm that can be used for control system design. The proposed Genetic Programming algorithm named Block Diagram Oriented Genetic Programming BDOGP. The characteristics of the tree structure and the used genetic operations of the proposed algorithm are illustrated here in detail. A new numeric constant mutation operation is added to the algorithm to strengthen the search for optimal parameters of the BDOGP solutions. The proposed BDOGP is used as an automated method to synthesize the block diagram of a deadbeat controller for discrete control system. The important BDOGP steps and remarks, which perform the solution, are presented. The simulation results show that the comparison between the GP solution and the conventional state-space one indicates the validity and accuracy of the solution obtained by BDOGP.