Optimization in Model Predictive Control Using Evolutionary-Gradient Algorithm

摘要

The term predictive control designates a class of control methods which are suitable for control of various kinds of systems. The basic principle of model based predictive control (MPC) is the calculating of the future behaviour of a system by means of a model. The best control values are determined by solving of an optimization problem which must be solved in each sampling period. One of the major advantages of MPC is its ability to take into account constraints within a controller. However, these constraints may cause that the optimization problem is computationally complex. There are many algorithms which may be used for solving this problem. The crucial issue of optimization is computational time. Hence, there must be applied algorithms which are computationally effective. This contribution deals with the optimization using a mixed evolutionary-gradient algorithm.