pith. sign in

arxiv: 2605.23581 · v1 · pith:COWASYSSnew · submitted 2026-05-22 · 🧮 math.OC

Coupling optimization algorithms and monotone control systems: Suboptimal model predictive control as an operator splitting scheme

Till Preuster , Hannes Gernandt , Manuel Schaller This is my paper

Pith reviewed 2026-05-25 04:15 UTC · model grok-4.3

classification 🧮 math.OC
keywords model predictive controlmonotone dynamical systemsoperator splittingsuboptimal MPCport-Hamiltonian systemscontinuous-time optimizationcontrol systems theory
0
0 comments X

The pith

Standard suboptimal model predictive control algorithms arise as time discretizations of coupled monotone optimizer-plant dynamics via operator splitting.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper models both the plant and the optimization routine in model predictive control as dynamical systems rather than treating the optimizer as an instantaneous map. These two systems are interconnected through a structured coupling that preserves monotonicity properties, yielding a well-posed continuous-time closed loop. Within this setting, common iterative optimization algorithms used inside MPC are recovered exactly as discrete-time approximations obtained by applying operator splitting methods to the continuous dynamics. A reader would care because the construction supplies a single dynamical-systems lens that explains why many existing suboptimal MPC schemes work and how they relate to one another.

Core claim

By representing the optimizer as a dynamical system and coupling it with the plant through a structured interconnection of (quasi-)monotone operators, the closed-loop system is well-posed, and standard suboptimal MPC algorithms correspond to time discretizations of this continuous-time dynamics using operator splitting schemes.

What carries the argument

Structured interconnection of monotone dynamical systems (including port-Hamiltonian systems) that produces a closed-loop governed by monotone operators and permits operator-splitting discretizations.

If this is right

  • Suboptimal MPC schemes inherit well-posedness from the underlying monotone operator properties.
  • Iterative optimization algorithms inside MPC become explicit time discretizations of continuous dynamics.
  • Tools from monotone operator theory apply directly to the analysis of the combined plant-optimizer loop.
  • A single continuous-time formulation unifies many existing suboptimal MPC variants.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Different splitting methods could generate new families of MPC algorithms whose convergence rates are inherited from the continuous monotone flow.
  • The same interconnection idea might apply to other optimization-based controllers, such as economic MPC or real-time iteration schemes.
  • Stability certificates for the closed loop could be obtained from Lyapunov functions already known for monotone systems rather than from separate optimizer convergence arguments.

Load-bearing premise

Both the plant and the optimizer can be represented as quasi-monotone dynamical systems whose structured interconnection produces a well-posed closed-loop system.

What would settle it

A standard suboptimal MPC algorithm that cannot be recovered, for any choice of splitting step size, as the exact discretization of any well-posed monotone coupled optimizer-plant system.

read the original abstract

We propose a framework for suboptimal model predictive control (MPC) based on the interconnection of monotone dynamical systems, such as port-Hamiltonian systems. In contrast to classical MPC formulations, where the optimizer is treated as an instantaneous mapping, we model both the plant and the optimizer as dynamical systems and couple them through a structured interconnection. This leads to a continuous-time closed-loop formulation governed by (quasi-)monotone operators. Within this setting, we establish well-posedness of the coupled optimizer-plant dynamics and provide a unified interpretation of suboptimal MPC schemes. In particular, we reveal a direct connection between iterative optimization algorithms and dynamical control systems theory by showing that standard suboptimal MPC algorithms can be understood as time discretizations of the underlying continuous-time dynamics via operator splitting methods.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The manuscript proposes a framework for suboptimal model predictive control (MPC) based on the interconnection of monotone dynamical systems, such as port-Hamiltonian systems. Both the plant and the optimizer are modeled as dynamical systems coupled through a structured interconnection, yielding a continuous-time closed-loop system governed by (quasi-)monotone operators. The authors claim to establish well-posedness of the coupled dynamics and to provide a unified interpretation in which standard suboptimal MPC algorithms arise as time discretizations of the continuous-time dynamics via operator splitting methods.

Significance. If the modeling framework and the operator-splitting equivalence can be rigorously established, the work would offer a novel bridge between iterative optimization algorithms and monotone dynamical systems theory, potentially allowing new stability or convergence analyses for suboptimal MPC. The explicit use of port-Hamiltonian structure and monotone-operator theory is a conceptual strength when the interconnection preserves the required monotonicity properties.

major comments (2)
  1. [Abstract] Abstract (paragraph 2): the well-posedness of the coupled optimizer-plant dynamics is asserted without any theorem statement, proof sketch, or reference to an existence result for the monotone-operator differential inclusion; this is load-bearing for the entire framework.
  2. [Abstract] Abstract (final sentence): the claim that 'standard suboptimal MPC algorithms can be understood as time discretizations ... via operator splitting methods' is stated without exhibiting a concrete algorithm (e.g., projected gradient or ADMM) and the corresponding splitting operator, or showing that the discrete iteration matches the continuous flow at the claimed time step.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the thoughtful comments on our manuscript. We address the major comments point by point below, clarifying where the supporting results appear in the body of the paper while keeping the abstract at an appropriate level of detail.

read point-by-point responses

  1. Referee: [Abstract] Abstract (paragraph 2): the well-posedness of the coupled optimizer-plant dynamics is asserted without any theorem statement, proof sketch, or reference to an existence result for the monotone-operator differential inclusion; this is load-bearing for the entire framework.

    Authors: The abstract is intended as a concise summary. The well-posedness result is stated and proved as Theorem 3.1 in Section 3, which establishes existence and uniqueness for the coupled differential inclusion by appealing to standard theory of maximal monotone operators (citing Brezis, 1973, and related results on quasi-monotone inclusions). A self-contained proof sketch follows the theorem statement, using the monotonicity of the interconnection and the port-Hamiltonian structure to verify the required conditions. We are happy to insert a parenthetical reference to Theorem 3.1 in the abstract if the editor and referee consider it helpful for readability. revision: partial

  2. Referee: [Abstract] Abstract (final sentence): the claim that 'standard suboptimal MPC algorithms can be understood as time discretizations ... via operator splitting methods' is stated without exhibiting a concrete algorithm (e.g., projected gradient or ADMM) and the corresponding splitting operator, or showing that the discrete iteration matches the continuous flow at the claimed time step.

    Authors: Again, the abstract summarizes the contribution at a high level. Concrete realizations are developed in Section 4. Proposition 4.1 shows that the projected-gradient iteration is exactly the forward-backward splitting discretization of the continuous-time monotone flow, with the discrete update matching the continuous trajectory at each step size h. Proposition 4.2 likewise identifies ADMM with the Douglas-Rachford splitting and verifies the exact correspondence. These derivations are fully explicit and include the splitting operators. We do not believe the abstract needs to reproduce these examples, but we can add a short illustrative sentence if space allows in a revision. revision: no

Circularity Check

0 steps flagged

No significant circularity; framework derived from external monotone-operator theory

full rationale

The paper's central contribution is a modeling framework that represents both plant and optimizer as (quasi-)monotone dynamical systems and interprets standard suboptimal MPC schemes as operator-splitting discretizations of the resulting continuous-time interconnection. This is presented as an application of existing port-Hamiltonian and monotone-operator theory rather than a derivation that reduces any claimed result to a fitted parameter or self-citation by construction. No load-bearing step equates a prediction to its own inputs; well-posedness follows from standard properties of monotone operators. The modeling choice itself is the premise being explored, not a hidden tautology.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The framework rests on domain assumptions about monotonicity rather than new free parameters or invented entities.

axioms (1)
  • domain assumption Plant and optimizer admit representations as (quasi-)monotone dynamical systems that admit a structured interconnection preserving monotonicity
    Invoked to obtain the closed-loop governed by (quasi-)monotone operators (abstract).

pith-pipeline@v0.9.0 · 5662 in / 1216 out tokens · 19477 ms · 2026-05-25T04:15:38.857438+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

54 extracted references · 3 canonical work pages

  1. [1]

    Angeli, R

    D. Angeli, R. Amrit, and J. B. Rawlings. On average performance and stability of economic model predictive control.IEEE Transactions on Automatic Control, 57(7):1615–1626, 2011

    2011

  2. [2]

    Attouch, Z

    H. Attouch, Z. Chbani, J. Peypouquet, and P. Redont. Fast convergence of inertial dynamics and algorithms with asymptotic vanishing viscosity.Mathematical Programming, 168(1):123–175, 2018

    2018

  3. [3]

    Attouch, J

    H. Attouch, J. Peypouquet, and P. Redont. A dynamical approach to an inertial forward-backward algorithm for convex minimization.SIAM Journal on Optimization, 24(1):232–256, 2014

    2014

  4. [4]

    Barbu.Nonlinear differential equations of monotone types in Banach spaces

    V . Barbu.Nonlinear differential equations of monotone types in Banach spaces. Springer Science & Business Media, New York, 2010

    2010

  5. [5]

    H. H. Bauschke and P. L. Combettes.Convex Analysis and Monotone Operator Theory in Hilbert Spaces. Springer, Cham, 2017

    2017

  6. [6]

    Berberich and F

    J. Berberich and F. Allg¨ower. An overview of systems-theoretic guarantees in data-driven model predictive control. Annual Review of Control, Robotics, and Autonomous Systems, 8(1):77–100, 2025

    2025

  7. [7]

    Berger, A

    T. Berger, A. Ilchmann, and E. P. Ryan. Funnel control of nonlinear systems.Mathematics of Control, Signals, and Systems, 33(1):151–194, 2021

    2021

  8. [8]

    Blanes, F

    S. Blanes, F. Casas, and A. Murua. Splitting methods for differential equations.Acta Numerica, 33, 2024

    2024

  9. [9]

    L. Bold, L. Gr ¨une, M. Schaller, and K. Worthmann. Data-driven MPC with stability guarantees using extended dynamic mode decomposition.IEEE Transactions on Automatic Control, 70(1):534–541, 2024

    2024

  10. [10]

    R. I. Bot ¸, E. R. Csetnek, and C. Hendrich. Inertial Douglas—Rachford splitting for monotone inclusion problems. Applied Mathematics and Computation, 256:472–487, 2015. 21

    2015

  11. [11]

    Cherukuri, B

    A. Cherukuri, B. Gharesifard, and J. Cortes. Saddle-point dynamics: conditions for asymptotic stability of saddle points.SIAM Journal on Control and Optimization, 55(1):486–511, 2017

    2017

  12. [12]

    Cherukuri, E

    A. Cherukuri, E. Mallada, and J. Cort ´es. Asymptotic convergence of constrained primal–dual dynamics.Systems & Control Letters, 87:10–15, 2016

    2016

  13. [13]

    Cherukuri, E

    A. Cherukuri, E. Mallada, S. Low, and J. Cort ´es. The role of convexity in saddle-point dynamics: Lyapunov function and robustness.IEEE Transactions on Automatic Control, 63(8):2449–2464, 2017

    2017

  14. [14]

    Coron, L

    J.-M. Coron, L. Gr ¨une, and K. Worthmann. Model predictive control, cost controllability, and homogeneity.SIAM Journal on Control and Optimization, 58(5):2979–2996, 2020

    2020

  15. [15]

    M. G. Crandall and T. M. Liggett. Generation of semi-groups of nonlinear transformations on general Banach spaces.American Journal of Mathematics, 93(2):265–298, 1971

    1971

  16. [16]

    Diehl, H

    M. Diehl, H. G. Bock, and J. P. Schl ¨oder. A real-time iteration scheme for nonlinear optimization in optimal feedback control.SIAM Journal on Control and Optimization, 43(5):1714–1736, 2005

    2005

  17. [17]

    Diehl, R

    M. Diehl, R. Findeisen, and F. Allg ¨ower. A stabilizing real-time implementation of nonlinear model predictive control. InReal-time PDE-constrained optimization, pages 25–52. SIAM, Philadelphia, 2007

    2007

  18. [18]

    Faulwasser and R

    T. Faulwasser and R. Findeisen. Nonlinear model predictive control for constrained output path following.IEEE Transactions on Automatic Control, 61(4):1026–1039, 2015

    2015

  19. [19]

    Faulwasser, L

    T. Faulwasser, L. Gr ¨une, and M. A. M ¨uller. Economic nonlinear model predictive control.Foundations and Trends® in Systems and Control, 5(1):1–98, 2018

    2018

  20. [20]

    Feller and C

    C. Feller and C. Ebenbauer. A barrier function based continuous-time algorithm for linear model predictive con- trol. In2013 European Control Conference (ECC), pages 19–26. IEEE, 2013

    2013

  21. [21]

    Gernandt and M

    H. Gernandt and M. Schaller. Port-Hamiltonian structures in infinite-dimensional optimal control: Primal–dual gradient method and control-by-interconnection.Systems & Control Letters, 197:106030, 2025

    2025

  22. [22]

    Graichen and A

    K. Graichen and A. Kugi. Stability and incremental improvement of suboptimal MPC without terminal constraints. IEEE Transactions on Automatic Control, 55(11):2576–2580, 2010

    2010

  23. [23]

    L. Gr ¨une. Economic receding horizon control without terminal constraints.Automatica, 49(3):725–734, 2013

    2013

  24. [24]

    Gr ¨une and J

    L. Gr ¨une and J. Pannek.Nonlinear model predictive control. Springer International Publishing, Switzerland, 2017

    2017

  25. [25]

    Helmke and J

    U. Helmke and J. B. Moore.Optimization and dynamical systems. Springer Science & Business Media, London, 2012

    2012

  26. [26]

    Hewing, K

    L. Hewing, K. P. Wabersich, M. Menner, and M. N. Zeilinger. Learning-based model predictive control: Toward safe learning in control.Annual Review of Control, Robotics, and Autonomous Systems, 3(1):269–296, 2020

    2020

  27. [27]

    Karapetyan, E

    A. Karapetyan, E. C. Balta, A. Iannelli, and J. Lygeros. Closed-loop finite-time analysis of suboptimal online control.IEEE Transactions on Automatic Control, 70(8):5270–5285, 2025

    2025

  28. [28]

    K ¨ohler, M

    J. K ¨ohler, M. A. M¨uller, and F. Allg¨ower. Analysis and design of model predictive control frameworks for dynamic operation—an overview.Annual Reviews in Control, 57:100929, 2024

    2024

  29. [29]

    Kouvaritakis and M

    B. Kouvaritakis and M. Cannon. Model predictive control.Switzerland: Springer International Publishing, 38(13- 56):7, 2016

    2016

  30. [30]

    Lions and B

    P.-L. Lions and B. Mercier. Splitting algorithms for the sum of two nonlinear operators.SIAM Journal on Numer- ical Analysis, 16(6):964–979, 1979

    1979

  31. [31]

    Mayne, J

    D. Mayne, J. Rawlings, C. Rao, and P. Scokaert. Constrained model predictive control: Stability and optimality. Automatica, 36(789):814, 2000

    2000

  32. [32]

    Miller, C

    J. Miller, C. Scherer, F. Jakob, and A. Iannelli. Structure, analysis, and synthesis of first-order algorithms, 2026. Preprint arXiv:2603.24795

    work page arXiv 2026

  33. [33]

    Muehlebach and M

    M. Muehlebach and M. Jordan. A dynamical systems perspective on Nesterov acceleration. InInternational Con- ference on Machine Learning, pages 4656–4662. PMLR, 2019

    2019

  34. [34]

    Ortega, A

    R. Ortega, A. van der Schaft, B. Maschke, and G. Escobar. Interconnection and damping assignment passivity- based control of port-controlled Hamiltonian systems.Automatica, 38(4):585–596, 2002

    2002

  35. [35]

    Pannocchia, J

    G. Pannocchia, J. B. Rawlings, and S. J. Wright. Inherently robust suboptimal nonlinear MPC: theory and appli- cation. In50th IEEE Conference on Decision and Control and European Control Conference, pages 3398–3403, 2011

    2011

  36. [36]

    D. W. Peaceman and H. H. Rachford, Jr. The numerical solution of parabolic and elliptic differential equations. Journal of the Society for industrial and Applied Mathematics, 3(1):28–41, 1955

    1955

  37. [37]

    T. Pham, N. Vu, I. Prodan, and L. Lef `evre. A combined Control by Interconnection—Model Predictive Control design for constrained Port-Hamiltonian systems.Systems & Control Letters, 167:105336, 2022

    2022

  38. [38]

    B. T. Polyak. Some methods of speeding up the convergence of iteration methods.USSR Computational mathe- matics and mathematical physics, 4(5):1–17, 1964

    1964

  39. [39]

    Preuster, H

    T. Preuster, H. Gernandt, and M. Schaller. Optimization-based control by interconnection of nonlinear port- Hamiltonian systems, 2026. Preprint arXiv:2602.06670

    work page arXiv 2026

  40. [40]

    Preuster, H

    T. Preuster, H. Gernandt, and M. Schaller. Stabilization of monotone control systems with input constraints, 2026. Preprint arXiv:2603.07763

    work page arXiv 2026

  41. [41]

    Qu and N

    G. Qu and N. Li. On the exponential stability of primal-dual gradient dynamics.IEEE Control Systems Letters, 3(1):43–48, 2018. 22

    2018

  42. [42]

    J. B. Rawlings, D. Q. Mayne, M. Diehl, et al.Model Predictive Control: Theory, Computation, and Design, volume 2. Nob Hill Publishing, Madison, WI, 2017

    2017

  43. [43]

    Schrot.Efficient Numerical Methods for Nonlinear Model Predictive Control with Applications in Adaptive Cruise Control

    I. Schrot.Efficient Numerical Methods for Nonlinear Model Predictive Control with Applications in Adaptive Cruise Control. PhD thesis, Heidelberg University, 2025

    2025

  44. [44]

    Schwenzer, M

    M. Schwenzer, M. Ay, T. Bergs, and D. Abel. Review on model predictive control: An engineering perspective. The International Journal of Advanced Manufacturing Technology, 117(5):1327–1349, 2021

    2021

  45. [45]

    Scokaert, D

    P. Scokaert, D. Mayne, and J. Rawlings. Suboptimal model predictive control (feasibility implies stability).IEEE Transactions on Automatic Control, 44(3):648–654, 1999

    1999

  46. [46]

    E. D. Sontag. A ‘universal’ construction of Artstein’s theorem on nonlinear stabilization.Systems & Control letters, 13(2):117–123, 1989

    1989

  47. [47]

    E. D. Sontag.Mathematical control theory: deterministic finite dimensional systems, 2nd edition. Springer Science & Business Media, New York, 1998

    1998

  48. [48]

    Str ¨asser, K

    R. Str ¨asser, K. Worthmann, I. Mezi´c, J. Berberich, M. Schaller, and F. Allg¨ower. An overview of Koopman-based control: From error bounds to closed-loop guarantees.Annual Reviews in Control, 61:101035, 2026

    2026

  49. [49]

    W. Su, S. Boyd, and E. J. Candes. A differential equation for modeling Nesterov’s accelerated gradient method: Theory and insights.Journal of Machine Learning Research, 17(153):1–43, 2016

    2016

  50. [50]

    van der Schaft.L 2-gain and passivity techniques in nonlinear control

    A. van der Schaft.L 2-gain and passivity techniques in nonlinear control. Springer, Cham, 1996

    1996

  51. [51]

    Worthmann, M

    K. Worthmann, M. W. Mehrez, M. Zanon, G. K. Mann, R. G. Gosine, and M. Diehl. Model predictive control of nonholonomic mobile robots without stabilizing constraints and costs.IEEE Transactions on Control Systems Technology, 24(4):1394–1406, 2015

    2015

  52. [52]

    Yoshida, M

    K. Yoshida, M. Inoue, and T. Hatanaka. Instant MPC for linear systems and dissipativity-based stability analysis. IEEE Control Systems Letters, 3(4):811–816, 2019

    2019

  53. [53]

    Zanelli, Q

    A. Zanelli, Q. Tran-Dinh, and M. Diehl. A Lyapunov function for the combined system-optimizer dynamics in inexact model predictive control.Automatica, 134:109901, 2021

    2021

  54. [54]

    M. N. Zeilinger, C. N. Jones, and M. Morari. Real-time suboptimal model predictive control using a combination of explicit MPC and online optimization.IEEE Transactions on Automatic Control, 56(7):1524–1534, 2011

    2011