Alternative approach to optimization in model predictive control using hill climbing algorithm

Abstract

The term predictive control designates a class of control methods suitable for control of various kinds of systems. One of the major advantages of predictive control is its ability to do on-line constraints handling in a systematic way. The predictive control is based on the prediction of a system behavior using a model. Based on this prediction, it is possible to optimize the systems behavior by utilization of a cost function. Each of control variables may be limited thus creating a specific subspace within a cost function. This problem is computationally complex and must be solved in each sampling period by optimization algorithms. Various kinds of algorithms may be applied. This contribution is focused on an alternative approach to optimization by implementation of Hill Climbing algorithm. The motivation for this concept is an effort to find algorithms suitable for reduction of computational expenses. These algorithms might be applied for control of systems with faster dynamics.