Combination of evolutionary and gradient optimization techniques in model predictive control

Model predictive control (MPC) designates a control method based on the model. This method is suitable for controlling of various kinds of systems. The basic principle is to calculate the future behaviour of a system and to use this prediction for the optimization of a control process. The optimization problem must be then solved in each sampling period. One of the advantages of MPC is its ability to do online constraints handling systematically. These constraints may, however, cause that the optimization problem is more complex. In this case, some iterative algorithms must be applied in order to solve this problem effectively. This paper is focus on the combination of the optimization techniques. The basic idea is to combine the advantages of gradient and evolutionary algorithms. © 2016, North Atlantic University Union NAUN. All rights reserved