A Scalable Learned Iterative Solver for Nonlinear MPC via Temporal Decomposition
Abstract
Nonlinear model predictive control (NMPC) can be applied to large nonlinear systems with complex objectives while enforcing constraints, but the cost of solving the underlying optimization problems limits real-time application for systems with fast dynamics or many states. Neural networks can reduce this online cost by approximating the control law, but may violate constraints and typically require many precomputed optimizer solutions for training. The learning-based iterative solver for constrained optimization (LISCO) instead learns to predict primal-dual steps toward satisfaction of the Karush-Kuhn-Tucker (KKT) conditions through self-supervised training without precomputed solutions, but its dense, monolithic architecture grows with the prediction horizon. We remove this horizon dependence by exploiting the temporal structure of the problem in the solver architecture, which makes the network size independent of the horizon while leaving the training loss and the solver iteration unchanged. A closed-loop resampling strategy further aligns the training distribution with closed-loop operation at no additional sampling cost. On an economic NMPC benchmark of a CSTR cascade, the proposed solver scales to larger problems than monolithic LISCO and achieves median speedups of up to a factor of four over IPOPT. In closed loop, it achieves comparable control performance to IPOPT with no observed constraint violations.
Related
- Combined High-Dimensional and Reduced-Order Modeling Based on an Iterative Domain Decomposition Method
- Parameters Robust Parallelizable Iterative Decoupled Algorithms for Thermo-Poroelasticity with Nonlinear Convective Transport
- Two-level domain-decomposition AdaGrad method for scalable training of graph neural networks
- A Class of Higher-Order Iterative Methods for Simultaneous Solutions of Nonlinear Equations with Applications to Biological Problems
- Hybrid accelerated iterative scheme based on the Picard–Ishikawa approach for systems of nonlinear equations
BibTeX
@misc{luken2026scalable,
title = {{A Scalable Learned Iterative Solver for Nonlinear MPC via Temporal Decomposition}},
author = {Lukas Lüken and Sergio Lucia},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.13480},
eprint = {2610.13480},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.13480},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - UNPB
TI - A Scalable Learned Iterative Solver for Nonlinear MPC via Temporal Decomposition
AU - Lukas Lüken
AU - Sergio Lucia
PY - 2026
DA - 2026/10/09
PB - arXiv
DO - 10.48550/arxiv.2610.13480
UR - https://arxiv.org/abs/2610.13480
AB - Nonlinear model predictive control (NMPC) can be applied to large nonlinear systems with complex objectives while enforcing constraints, but the cost of solving the underlying optimization problems limits real-time application for systems with fast dynamics or many states. Neural networks can reduce this online cost by approximating the control law, but may violate constraints and typically require many precomputed optimizer solutions for training. The learning-based iterative solver for constrained optimization (LISCO) instead learns to predict primal-dual steps toward satisfaction of the Karush-Kuhn-Tucker (KKT) conditions through self-supervised training without precomputed solutions, but its dense, monolithic architecture grows with the prediction horizon. We remove this horizon dependence by exploiting the temporal structure of the problem in the solver architecture, which makes the network size independent of the horizon while leaving the training loss and the solver iteration unchanged. A closed-loop resampling strategy further aligns the training distribution with closed-loop operation at no additional sampling cost. On an economic NMPC benchmark of a CSTR cascade, the proposed solver scales to larger problems than monolithic LISCO and achieves median speedups of up to a factor of four over IPOPT. In closed loop, it achieves comparable control performance to IPOPT with no observed constraint violations.
ER -