pith. sign in

arxiv: 2606.09609 · v1 · pith:L7V2FOWEnew · submitted 2026-06-08 · 🧮 math.OC

Complete Trajectory Tracking for Polynomial Differential Variational Inequalities: A Moment-SOS Based Structure-Aware Approach

Zhang Huang , Wei Li , Jie Wang This is my paper

Pith reviewed 2026-06-27 15:32 UTC · model grok-4.3

classification 🧮 math.OC
keywords polynomial differential variational inequalitiesparametric variational inequalitiessemi-algebraic setsmoment sum of squarestrajectory trackingRunge-Kutta methodsnon-smooth systems
0
0 comments X

The pith

The parameter space of parametric polynomial variational inequalities stratifies into finitely many semi-algebraic regions with distinct solution structures.

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

This paper shows that under mild assumptions the parameter space of a parametric polynomial variational inequality can be partitioned into a finite collection of disjoint semi-algebraic sets. Each set corresponds to a qualitatively different way the variational inequality can be solved. The authors use this partition to prove that solutions to the associated polynomial differential variational inequality exist and have only finitely many switches for almost all starting points. They then build an algorithm that follows the geometry of these regions to track entire trajectories even when the solution map jumps at the boundaries.

Core claim

Under mild assumptions, the parameter space of the parametric polynomial variational inequality admits a stratification into finitely many disjoint semi-algebraic regions, each corresponding to a qualitatively distinct solution structure. Maximal regular solutions to the PDVI exist for almost every initial condition in the parametric feasible set, each having finitely many switches and evolving according to continuous semi-algebraic selection laws between switching times. The structure-aware SOS-PDVIRK4 algorithm integrates this geometric information with a fourth-order Runge-Kutta method and the Moment-SOS hierarchy to manage solution multiplicity and transitions.

What carries the argument

Stratification of the parameter space into semi-algebraic regions corresponding to distinct solution structures of the parametric polynomial variational inequality.

If this is right

  • Maximal regular solutions to the PDVI exist almost everywhere.
  • These solutions switch only finitely many times.
  • Between switches the solution follows continuous semi-algebraic selection laws.
  • The structure-aware algorithm can track complete trajectories despite discontinuities in the solution map.

Where Pith is reading between the lines

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

  • The finite stratification may allow computational enumeration of all possible long-term behaviors in such systems.
  • Similar geometric partitions could apply to other polynomial-defined non-smooth dynamics if the mild assumptions hold.
  • Divergence of trajectories from different initial controls points to the need for careful initialization in control applications.

Load-bearing premise

The polynomials and feasible set satisfy mild assumptions that permit a finite stratification of the parameter space by solution structure.

What would settle it

A concrete polynomial variational inequality whose parameter space requires infinitely many distinct semi-algebraic regions to capture all solution structures would disprove the stratification claim.

read the original abstract

This paper investigates Polynomial Differential Variational Inequalities (PDVIs), a class of non-smooth dynamical systems in which an ordinary differential equation is coupled with a time-dependent variational inequality (VI) defined by polynomials. As a foundation for polynomial differential variational inequalities, we investigate the parametric polynomial variational inequality and establish that, under mild assumptions, its parameter space admits a stratification into finitely many disjoint semi-algebraic regions, each corresponding to a qualitatively distinct solution structure. Furthermore, we establish the existence of maximal regular solutions to the PDVI for almost every initial condition in the parametric feasible set, where each solution has finitely many switches and evolves according to continuous semi-algebraic selection laws between switching times. Leveraging the structured partition of the parameter space, in which region boundaries correspond to discontinuities in the solution map of the parametric VI, we propose the structure-aware SOS-PDVIRK4 algorithm to handle solution multiplicity and structural transitions. This method integrates a fourth-order Runge--Kutta integrator with the Moment--SOS hierarchy and features a novel hybrid strategy that is explicitly guided by the underlying parameter space geometry. The efficacy of our structure-aware framework is demonstrated through numerical experiments on a constrained cyber-physical control system. Our numerical experiments demonstrate that, from identical initial states but different initial controls, trajectories can diverge significantly, revealing the discontinuous structure of the solution map of the parametric polynomial VI. The experiments also include numerical convergence studies and highlight the potential of our algorithm to explore the long-term behavior of PDVI systems.

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 / 2 minor

Summary. The paper claims that, under mild assumptions, the parameter space of a parametric polynomial variational inequality admits a stratification into finitely many disjoint semi-algebraic regions each corresponding to a qualitatively distinct solution structure; that maximal regular solutions to the associated polynomial differential variational inequality (PDVI) exist for almost every initial condition, each having finitely many switches and evolving via continuous semi-algebraic selection laws; and that a structure-aware SOS-PDVIRK4 algorithm, which integrates a fourth-order Runge-Kutta integrator with the Moment-SOS hierarchy and is guided by the parameter-space geometry, can handle solution multiplicity and structural transitions. These claims are supported by numerical experiments on a constrained cyber-physical control system that illustrate trajectory divergence from identical initial states but different initial controls.

Significance. If the stratification and existence results hold with the stated properties, the work would supply a geometrically structured foundation for analyzing and numerically tracking non-smooth polynomial dynamical systems that couple ODEs with time-dependent polynomial VIs. The explicit use of semi-algebraic geometry to partition the parameter space and the hybrid Moment-SOS / Runge-Kutta scheme represent a concrete algorithmic contribution that could be useful for control and optimization problems exhibiting solution-map discontinuities.

major comments (2)
  1. [Abstract] Abstract: the central existence and stratification claims are asserted only 'under mild assumptions,' yet no explicit list of these assumptions (e.g., compactness of the feasible set, monotonicity or convexity of the VI mapping, degree bounds, or non-degeneracy conditions ensuring the solution map is semi-algebraic) is supplied anywhere in the manuscript. Without such a list the finiteness of the stratification and the 'almost every' measure-zero claim cannot be verified against standard results in real algebraic geometry.
  2. [Abstract] Abstract and numerical-experiments section: the manuscript states that 'maximal regular solutions' exist with 'finitely many switches' and 'continuous semi-algebraic selection laws,' but supplies neither the proof of these statements nor any quantitative error metrics (e.g., residual norms, convergence rates of the Moment-SOS hierarchy, or trajectory-tracking errors) for the reported experiments. This absence renders the mathematical support for the central claims unverifiable from the given text.
minor comments (2)
  1. The algorithm is named SOS-PDVIRK4; a brief expansion of the acronym on first use would improve readability.
  2. The experiments mention 'numerical convergence studies' but do not report tables or figures with explicit convergence orders or residual values; adding such quantitative data would strengthen the presentation.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments. We address each major comment below and indicate the revisions we will undertake.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central existence and stratification claims are asserted only 'under mild assumptions,' yet no explicit list of these assumptions (e.g., compactness of the feasible set, monotonicity or convexity of the VI mapping, degree bounds, or non-degeneracy conditions ensuring the solution map is semi-algebraic) is supplied anywhere in the manuscript. Without such a list the finiteness of the stratification and the 'almost every' measure-zero claim cannot be verified against standard results in real algebraic geometry.

    Authors: We agree that an explicit enumeration of the assumptions is necessary for verifiability. The assumptions invoked are those standard for applying semi-algebraic geometry to parametric polynomial VIs: compactness of the feasible set, bounded polynomial degree, monotonicity of the VI mapping, and non-degeneracy conditions ensuring the solution map is semi-algebraic. These are used throughout Sections 2--4 but were not collected in one place. In the revised manuscript we will add a concise, numbered list of these assumptions immediately after the abstract and in the introduction. revision: yes

  2. Referee: [Abstract] Abstract and numerical-experiments section: the manuscript states that 'maximal regular solutions' exist with 'finitely many switches' and 'continuous semi-algebraic selection laws,' but supplies neither the proof of these statements nor any quantitative error metrics (e.g., residual norms, convergence rates of the Moment-SOS hierarchy, or trajectory-tracking errors) for the reported experiments. This absence renders the mathematical support for the central claims unverifiable from the given text.

    Authors: The existence, finite-switch, and semi-algebraic selection results are proved in Theorems 3.1--3.3 and the supporting lemmas of Section 3, which rely on the stratification constructed in Section 2 together with standard results on semi-algebraic sets and differential inclusions. The numerical-experiments section does contain convergence studies, yet we acknowledge that explicit quantitative metrics (residual norms, SOS hierarchy convergence rates, and trajectory errors) are not tabulated. In the revision we will add these metrics for the cyber-physical example. revision: partial

Circularity Check

0 steps flagged

No circularity; claims invoke external mild assumptions without reduction to self-definition or fitted inputs.

full rationale

The abstract states results on stratification and maximal regular PDVI solutions hold 'under mild assumptions' on polynomials and the feasible set, with the structure-aware algorithm then leveraging the resulting partition. No equations, self-citations, or derivations are exhibited that reduce the claimed stratification or existence statements to fitted quantities, prior self-referential definitions, or ansatzes smuggled via citation. The derivation chain remains self-contained against external semi-algebraic geometry benchmarks, yielding a normal non-finding.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claims rest on unspecified mild assumptions about the polynomial data and the feasible set; no free parameters, invented entities, or additional axioms are stated in the abstract.

axioms (1)
  • domain assumption Mild assumptions on the polynomials and feasible set that guarantee the stratification into finitely many semi-algebraic regions and the existence of maximal regular solutions
    Invoked for both the parametric VI stratification and the PDVI existence result.

pith-pipeline@v0.9.1-grok · 5806 in / 1347 out tokens · 26367 ms · 2026-06-27T15:32:16.963897+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

43 extracted references

  1. [1]

    Pang, J.S., Stewart, D.E.: Differential variational inequalities.Mathematical Programming113, 345–424 (2008)

    2008

  2. [2]

    Nonlinear Analysis: Hybrid Systems2(3), 812–818 (2008)

    Stewart, D.E.: Uniqueness for index-one differential variational inequalities. Nonlinear Analysis: Hybrid Systems2(3), 812–818 (2008)

    2008

  3. [3]

    Springer, New York (2010)

    Friesz, T.L.:Dynamic Optimization and Differential Games. Springer, New York (2010)

    2010

  4. [4]

    Springer, Berlin (2008)

    Acary, V., Brogliato, B.:Numerical Methods for Nonsmooth Dynamical Systems: Applications in Mechanics and Electronics. Springer, Berlin (2008)

    2008

  5. [5]

    Li, X.S., Huang, N.J., O’Regan, D.: Differential mixed variational inequalities in finite dimensional spaces.Nonlinear Analysis72, 3875–3886 (2010) 31

    2010

  6. [6]

    Chen, X.J., Wang, Z.F.: Convergence of regularized time-stepping methods for differential variational inequalities.SIAM Journal on Optimization23(3), 1647– 1671 (2013)

    2013

  7. [7]

    Han, L.S., Pang, J.S.: Non-zeroness of a class of differential quasi-variational inequalities.Mathematical Programming121(1), 171–199 (2010)

    2010

  8. [8]

    Zhang, Y., Chen, T., Huang, N.J., Li, X.S.: Euler scheme for solving a class of stochastic differential variational inequalities with some applications.Communi- cations in Nonlinear Science and Numerical Simulation127, 107577 (2023)

    2023

  9. [9]

    Zhang, Y., Chen, T., Huang, N.J., Li, X.S.: Penalty method for solving a class of stochastic differential variational inequalities with an application.Nonlinear Analysis: Real World Applications73, 103889 (2023)

    2023

  10. [10]

    Online first

    Huang, Z., Li, W., Zhang, Q., Xu, J., Yang, C.: A System of Differential Inverse Mixed Variational Inequalities: Existence, Uniqueness, and Algorithm.Journal of Optimization Theory and Applications209, 37 (2026). Online first

    2026

  11. [11]

    Claeys, M., Daafouz, J., Henrion, D.: Modal occupation measures and LMI relaxations for nonlinear switched systems control.Automatica64, 143–154 (2016)

    2016

  12. [12]

    In:2013 IEEE 52nd Annual Conference on Decision and Control (CDC), pp

    Henrion, D., Daafouz, J., Claeys, M.: Optimal switching control design for poly- nomial systems: an LMI approach. In:2013 IEEE 52nd Annual Conference on Decision and Control (CDC), pp. 1349–1354. IEEE (2013)

    2013

  13. [13]

    Springer, New York (2003)

    Facchinei, F., Pang, J.S.:Finite-Dimensional Variational Inequalities and Com- plementarity Problems. Springer, New York (2003)

    2003

  14. [14]

    Gowda, M.S.: Polynomial complementarity problems.Pacific Journal of Opti- mization13, 227–241 (2017)

    2017

  15. [15]

    Journal of Optimization Theory and Applications183, 1–23 (2019)

    Huang, Z.H., Qi, L.Q.: Tensor complementarity problems—Part I: Basic theory. Journal of Optimization Theory and Applications183, 1–23 (2019)

    2019

  16. [16]

    Qi, L.Q., Huang, Z.H.: Tensor complementarity problems—Part II: Solution methods.Journal of Optimization Theory and Applications183, 365–385 (2019)

    2019

  17. [17]

    Hieu, V.T.: Solution maps of polynomial variational inequalities.Journal of Global Optimization77(4), 807–824 (2020)

    2020

  18. [18]

    Hieu, V.T.: On the numbers of connected components in the solution sets of polynomial vector variational inequalities.Journal of Optimization Theory and Applications181(1), 95–100 (2019)

    2019

  19. [19]

    Huong, N.T.T., Yao, J.C., Yen, N.D.: Polynomial vector variational inequalities under polynomial constraints and applications.SIAM Journal on Optimization 32 26(2), 1060–1071 (2016)

    2016

  20. [20]

    Dinh, V.D., Pham, T.S.: Error bounds of regularized gap functions for polynomial variational inequalities.Journal of Global Optimization81(4), 935–954 (2021)

    2021

  21. [21]

    Nie, J.: Tight relaxations for polynomial optimization and Lagrange multiplier expressions.Mathematical Programming178(1), 1–37 (2019)

    2019

  22. [22]

    Computational Optimization and Applications90, 361–394 (2025)

    Nie, J., Sun, D., Tang, X., Zhang, M.: Solving polynomial variational inequal- ity problems via Lagrange multiplier expressions and Moment-SOS relaxations. Computational Optimization and Applications90, 361–394 (2025)

    2025

  23. [23]

    Mathematical Programming142, 485–510 (2013)

    Nie, J.: Certifying convergence of Lasserre’s hierarchy via flat truncation. Mathematical Programming142, 485–510 (2013)

    2013

  24. [24]

    Magron, V., Forets, M., Henrion, D.: Semidefinite approximations of invariant measures for polynomial systems.Discrete and Continuous Dynamical Systems - Series B24(12), 6745–6770 (2019)

    2019

  25. [25]

    Dirkse, S.P., Ferris, M.C.: The PATH solver: a nonmonotone stabilization scheme for mixed complementarity problems.Optimization Methods and Software5(2), 123–156 (1995)

    1995

  26. [26]

    Ferris, M.C., Munson, T.S.: Interfaces to PATH 3.0: design, implementation and usage.Computational Optimization and Applications12(1), 207–227 (1999)

    1999

  27. [27]

    Springer-Verlag, Berlin (1998)

    Bochnak, J., Coste, M., Roy, M.-F.:Real Algebraic Geometry. Springer-Verlag, Berlin (1998)

    1998

  28. [28]

    Universit` a di Pisa, Pisa (2002)

    Coste, M.:An Introduction to Semialgebraic Geometry. Universit` a di Pisa, Pisa (2002)

    2002

  29. [29]

    Hardt, R.M.: Semi-algebraic local-triviality in semi-algebraic mappings.American Journal of Mathematics102(2), 291–302 (1980)

    1980

  30. [30]

    Springer, New York (2006)

    Nocedal, J., Wright, S.J.:Numerical Optimization, 2nd edn. Springer, New York (2006)

    2006

  31. [31]

    Lasserre, J.B.: Global optimization with polynomials and the problem of moments.SIAM Journal on Optimization11(3), 796–817 (2001)

    2001

  32. [32]

    Putinar, M.: Positive polynomials on compact semi-algebraic sets.Indiana University Mathematics Journal42(3), 969–984 (1993)

    1993

  33. [33]

    Lee, J.H., Lee, G.M., Pham, T.S.: Genericity and H¨ older stability in semi- algebraic variational inequalities.Journal of Optimization Theory and Applica- tions178(1), 56–77 (2018). 33

    2018

  34. [34]

    Birkh¨ auser, Boston (1990)

    Aubin, J.-P., Frankowska, H.:Set-Valued Analysis. Birkh¨ auser, Boston (1990)

    1990

  35. [35]

    In:Automata Theory and Formal Languages

    Collins, G.E.: Quantifier elimination for real closed fields by cylindrical algebraic decomposition. In:Automata Theory and Formal Languages. Lecture Notes in Computer Science, vol. 33, pp. 134–183. Springer, Berlin (1975)

    1975

  36. [36]

    Han, S., Pang, J.S.: Continuous selections of solutions to parametric variational inequalities.SIAM Journal on Optimization34(1), 870–892 (2024)

    2024

  37. [37]

    Springer-Verlag, Berlin (1993)

    Hairer, E., Nørsett, S.P., Wanner, G.:Solving Ordinary Differential Equations I: Nonstiff Problems, 2nd revised edn. Springer-Verlag, Berlin (1993)

    1993

  38. [38]

    In:Proceedings of the 2004 IEEE International Conference on Robotics and Automation, pp

    L¨ ofberg, J.: YALMIP: A toolbox for modeling and optimization in MATLAB. In:Proceedings of the 2004 IEEE International Conference on Robotics and Automation, pp. 284–289. Taipei, Taiwan (2004)

    2004

  39. [39]

    Version 9.3 (2021)

    MOSEK ApS:The MOSEK Optimization Toolbox for MATLAB Manual. Version 9.3 (2021)

    2021

  40. [40]

    Wang, J., Magron, V., Lasserre, J.-B.: TSSOS: A Moment-SOS Hierarchy That Exploits Term Sparsity.SIAM Journal on Optimization31(4), 30–58 (2021)

    2021

  41. [41]

    Wang, J., Schlosser, C., Korda, M., and Magron, V.: Exploiting Term Spar- sity in Moment-SOS Hierarchy for Dynamical Systems.IEEE Transactions on Automatic Control68(12), 8232-8237 (2023)

    2023

  42. [42]

    Graduate Texts in Mathematics, vol

    Srivastava, S.M.:A Course on Borel Sets. Graduate Texts in Mathematics, vol. 180. Springer-Verlag, New York (1998)

    1998

  43. [43]

    Actualit´ es Math´ ematiques

    Benedetti, R., Risler, J.-J.:Real Algebraic and Semi-algebraic Sets. Actualit´ es Math´ ematiques. Hermann, Paris (1990) 34

    1990