Contraction Analysis of Continuation Method for Reliability Evaluation of Suboptimal Model Predictive Control

A study conducted by Ryotaro Shima et al. was selected for presentation at The 2025 American Control Conference (ACC).

Model predictive control (MPC) determines control inputs by solving an optimization problem while predicting the future behavior of a system. In many industrial applications, such as automotive control, computational resources are limited. As a result, MPC is often implemented in a suboptimal form, where control inputs are applied before the optimization problem is fully solved. Consequently, different types of MPC algorithms have been used in theory and in practical applications.

In this study, we analyze the differences between suboptimal and conventional MPC using contraction analysis. Focusing on the continuation method, which is one of the representative suboptimal MPC algorithms, it quantitatively evaluates the impact of suboptimality on control system reliability. The results enable reliable MPC design even in implementation environments with limited computational resources, including automotive control systems.

This research was published in IEEE Control Systems Letters prior to its presentation at ACC (https://doi.org/10.1109/LCSYS.2024.3523584).

Title: Contraction Analysis of Continuation Method for Suboptimal Model Predictive Control
Authors: Shima, R., Ito, Y., Miyano, T.
Appears in: The 2025 American Control Conference (ACC)
Presented: July 8, 2025