Abstract

We approach the problem of finding the zeros of a continuous and monotone operator through a second-order dynamical system with a damping term of the form 1/tr1/t^{r}, where r∈[0,1]r\in [0, 1]. The system features the time derivative of the operator evaluated along the trajectory, which is a Hessian-driven type damping term when the governing operator comes from a potential. Also entering the system is a time rescaling parameter β(t)\beta(t) which satisfies a certain growth condition. We derive o(1/(t2rβ(t)))o(1/(t^{2r}\beta(t))) convergence rates for the norm of the operator evaluated along the generated trajectories as well as for a gap function which serves as a measure of optimality for the associated variational inequality. The parameter rr enters the growth condition for β(t)\beta(t): when r<1r < 1, the damping 1/tr1/t^{r} approaches zero at a slower speed than Nesterov's 1/t1/t damping; in this case, we are allowed to choose β(t)\beta(t) to be an exponential function, thus having linear convergence rates for the involved quantities. We also show weak convergence of the trajectories towards zeros of the governing operator. Through a particular choice for the operator, we establish a connection with the problem of minimizing a smooth and convex function with linear constraints. The convergence rates we derived in the operator case are inherited by the objective function evaluated at the trajectories and for the feasibility gap. We also prove weak convergence of the trajectories towards primal-dual solutions of the problem.
A discretization of the dynamical system yields an implicit algorithm that exhibits analogous convergence properties to its continuous counterpart.
We complement our theoretical findings with two numerical experiments.

Author information

Contact details are reproduced from the original publication and may be historical.

Dang-Khoa Nguyen

(1) Faculty of Mathematics, University of Science, Ho Chi Minh City, Vietnam
(2) Vietnam National University, Ho Chi Minh City, Vietnam

ndkhoa@hcmus.edu.vn

Keywords

  • Monotone equation
  • variational inequality
  • slow asymptotically vanishing damping
  • sublinear convergence rates
  • linear convergence rates
  • convergence of trajectories

Mathematics Subject Classification

47H05, 47J20, 65K10, 65K15

References

  1. F. Álvarez: On the minimizing property of a second-order dissipative system in Hilbert spaces, SIAM J. Control Optimization 38/4 (2000) 1102–1119.https://doi.org/10.1137/S0363012998335802Google Scholar
  2. F. Álvarez, H. Attouch: An inertial proximal method for maximal monotone operators via discretization of a nonlinear oscillator with damping, Set-Valued Analysis 9/1-2 (2001) 3–11.https://doi.org/10.1023/A:1011253113155Google Scholar
  3. H. Attouch, Z. Chbani, J. Fadili, H. Riahi: Fast convergence of dynamical ADMM via time scaling of damped inertial dynamics, J. Optimization Theory Appl. 193 (2002) 704–736.https://doi.org/10.1007/s10957-021-01859-2Google Scholar
  4. H. Attouch, S. C. László: Continuous Newton-like inertial dynamics for monotone inclusions, Set-Valued Var. Analysis 29 (2021) 555–581.https://doi.org/10.1007/s11228-020-00564-yGoogle Scholar
  5. H. Attouch, P. E. Maingé: Asymptotic behavior of second order dissipative evolution equations combining potential with non-potential effects, ESAIM Control Optim. Calc. Var. 17/3 (2011) 836–857.https://doi.org/10.1051/cocv/2010024Google Scholar
  6. H. Attouch, J. Peypouquet: Convergence of inertial dynamics and proximal algorithms governed by maximal monotone operators, Math. Programming, Ser. B 174/1-2 (2019) 391–432.https://doi.org/10.1007/s10107-018-1252-xGoogle Scholar
  7. H. Attouch, J. Peypouquet, P. Redont: Fast convex minimization via inertial dynamics with Hessian driven damping, J. Diff. Equations 210/10 (2016) 5734–5783.https://doi.org/10.1016/j.jde.2016.08.020Google Scholar
  8. H. Attouch, B. F. Svaiter: A continuous dynamical Newton-like approach to solving monotone inclusions, SIAM J. Control Optimization 49/2 (2011) 574–598.https://doi.org/10.1137/100784114Google Scholar
  9. A. Böhm, M. Sedlmayer, E. R. Csetnek, R. I. Boţ: Two steps at a time-taking GAN training in stride with Tseng’s method, SIAM J. Math. Data Sci. 4/2 (2022) 750–771.https://doi.org/10.1137/21M1420939Google Scholar
  10. R. I. Boţ, E. R. Csetnek: Second order forward-backward dynamical systems for monotone inclusion problems, SIAM J. Control Optimization 54 (2016) 1423–1443.https://doi.org/10.1137/15M1012657Google Scholar
  11. R. I. Boţ, E. R. Csetnek, D.-K. Nguyen: Fast optimistic gradient descent ascent (OGDA) method in continuous and discrete time, Found. Comp. Math. 25/1 (2025) 163–222.https://doi.org/10.1007/s10208-023-09636-5Google Scholar
  12. R. I. Boţ, D.-K. Nguyen: Improved convergence rates and trajectory convergence for primal-dual dynamical systems with vanishing damping, J. Diff. Equations 303 (2021) 369–406.https://doi.org/10.1016/j.jde.2021.09.021Google Scholar
  13. I. J. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, Y. Bengio: Generative adversarial networks, NIPS 2014: Adv. Neural Inform. Processing Syst. 27 (2014) 2672–2680.Google Scholar
  14. X. He, R. Hu, Y.-P. Fang: Convergence rates of inertial primal-dual dynamical methods for separable convex optimization problems, SIAM J. Control Optimization 59/5 (2021) 3278–3301.https://doi.org/10.1137/20M1355379Google Scholar
  15. X. He, R. Hu, Y.-P. Fang: Inertial primal-dual dynamics with damping and scaling for linearly constrained convex optimization problems, Appl. Analysis 102/15 (2023) 4114–4139.https://doi.org/10.1080/00036811.2022.2104260Google Scholar
  16. A. Madry, A. Makelov, L. Schmidt, D. Tsipras, A. Vladu: Towards deep learning models resistant to adversarial attacks, in: ICLR 2018: International Conference on Learning Representations (2018).Google Scholar
  17. S. Omidshafiei, J. Pazis, C. Amato, J. P. How, J. Vian: Deep decentralized multi-task multi-agent reinforcement learning under partial observability, in: Proceedings of the 34th International Conference on Machine Learning, PLMR 70 (2017) 2681–2690.Google Scholar
  18. Z. Opial: Weak convergence of the sequence of successive approximations for nonexpansive mappings, Bull. Amer. Math. Soc. 73/4 (1967) 591–597.https://doi.org/10.1090/S0002-9904-1967-11761-0Google Scholar
  19. W. Su, S. Boyd, E. Candès: A differential equation for modeling Nesterov’s accelerated gradient method: theory and insights, J. Machine Learning Res. 17/153 (2016) 1–43.Google Scholar
  20. X. Zeng, J. Lei, J. Chen: Dynamical primal-dual accelerated method with applications to network optimization, IEEE Trans. Auto. Control 68/3 (2023) 1760–1767.https://doi.org/10.1109/TAC.2022.3152720Google Scholar

Suggested citation

R. I. Boţ, D. A. Hulett, D.-K. Nguyen. “Fast Second-Order Dynamics with Slow Vanishing Damping Approaching the Zeros of a Monotone and Continuous Operator.” Journal of Convex Analysis 33 (2026), No. 3&4, 637–690. https://doi.org/10.68381/jca33040

Download BibTeXPublished by Heldermann Verlag.

Published by Heldermann Verlag, 2026. Rights now held by Banach Press LLC.