Approximate explicit model predictive control via piecewise nonlinear system identification

TL;DR

This paper introduces a methodology for approximating explicit model predictive control using piecewise nonlinear system identification, enabling faster computation while maintaining high accuracy in constrained nonlinear systems.

Contribution

The paper proposes a novel identification approach that produces piecewise nonlinear approximations of MPC, reducing computational complexity compared to traditional implicit methods.

Findings

Achieves high approximation accuracy with few regions

Significantly reduces computation time in industrial applications

Demonstrates effectiveness on constrained nonlinear systems

Abstract

This article presents an identification methodology to capture general relationships, with application to piecewise nonlinear approximations of model predictive control for constrained (non)linear systems. The mathematical formulation takes, at each iteration, the form of a constrained linear (or quadratic) optimization problem that is mathematically feasible as well as numerically tractable. The efficiency of the devised methodology is demonstrated via two industrial applications. Results suggest the possibility to achieve high approximate precision with limited number of regions, leading to a significant reduction in computation time when compared to the state-of-the-art implicit model predictive control solvers.