Hybrid predictive control: Mono-objective and multi-objective design

Abstract

The design of hybrid predictive control (HPC), which is based on both piecewise affine (PWA) models and hybrid fuzzy models, is described. For solving the nonlinear optimization problems resulting from these controllers, evolutionary algorithms are proposed, thereby demonstrating a reduction of computation effort in comparison with classical methods. Moreover, this chapter presents a new approach for dealing with HPC formulations using evolutionary multi-objective optimization. Different criteria are proposed to obtain an optimal control action from the Pareto front.