Least-Squares Multi-Step Koopman Operator Learning for Model Predictive Control

TL;DR

This paper introduces a multi-step EDMD approach for Koopman-based MPC that improves long-horizon prediction accuracy by avoiding error accumulation and enabling convex optimization.

Contribution

It proposes a convex least-squares multi-step learning framework that bypasses explicit lifted system identification, allowing efficient and accurate long-horizon predictions in Koopman-MPC.

Findings

Enhanced long-horizon prediction accuracy

Convex formulation enables efficient computation

Improved closed-loop control performance

Abstract

MPC is widely used in real-time applications, but practical implementations are typically restricted to convex QP formulations to ensure fast and certified execution. Koopman-based MPC enables QP-based control of nonlinear systems by lifting the dynamics to a higher-dimensional linear representation. However, existing approaches rely on single-step EDMD. Consequently, prediction errors may accumulate over long horizons when the EDMD operator is applied recursively. Moreover, the multi-step prediction loss is nonconvex with respect to the single-step EDMD operator, making long-horizon model identification particularly challenging. This paper proposes a multi-step EDMD framework that directly learns the condensed multi-step state-control mapping required for Koopman-MPC, thereby bypassing explicit identification of the lifted system matrices and subsequent model condensation. The resulting identification problem admits a convex least-squares formulation. We further show that the problem decomposes across prediction horizons and state coordinates, enabling parallel computation and row-wise $\ell_1$-regularization for automatic dictionary pruning. A non-asymptotic finite-sample analysis demonstrates that, unlike one-step EDMD, the proposed method avoids error compounding and yields error bounds that depend only on the target multi-step mapping. Numerical examples validate improved long-horizon prediction accuracy and closed-loop performance.