pith. sign in

arxiv: 2605.15534 · v1 · pith:Y67QXTGRnew · submitted 2026-05-15 · 🧮 math.OC · cs.GT· cs.SY· eess.SY

Distributionally Robust Nash Equilibrium Seeking with Partial Observations and Distributed Communication

Nirabhra Mandal , Sonia Mart\'inez This is my paper

Pith reviewed 2026-05-19 14:51 UTC · model grok-4.3

classification 🧮 math.OC cs.GTcs.SYeess.SY
keywords distributionally robust Nash equilibriumstochastic gamesWasserstein ambiguity setsupergradient dynamicsdistributed consensuspartial observationsNash equilibrium seeking
0
0 comments X p. Extension
pith:Y67QXTGR Add to your LaTeX paper What is a Pith Number? 
\usepackage{pith}
\pithnumber{Y67QXTGR}

Prints a linked pith:Y67QXTGR badge after your title and writes the identifier into PDF metadata. Compiles on arXiv with no extra files. Learn more

The pith

Stochastic games admit a nonempty set of distributionally robust Nash equilibria close to standard ones, which inertial dynamics can seek when amicable supergradients exist.

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

The paper studies stochastic one-shot games where each agent's utility depends on the joint strategy profile and an unknown random variable, with agents given only finite i.i.d. samples rather than the true distribution. Agents hedge by solving a distributionally robust version that maximizes the worst-case expected utility inside a Wasserstein ball around the empirical distribution. The authors supply conditions ensuring the resulting game has a nonempty set of distributionally robust Nash equilibria and prove that this set lies close to the Nash equilibria of the original stochastic game. They introduce inertial supported better-response ascending supergradient dynamics that converge to the robust equilibria whenever the game satisfies the amicable-supergradients property and then construct a distributed version that uses dynamic consensus over directed networks to handle both private and shared samples.

Core claim

We provide conditions under which the game has a non-empty set of distributionally robust Nash equilibria (DRoNE) and then characterize the closeness of the DRoNE set to the Nash equilibria (NE) of the associated stochastic game. We then propose an inertial, supported, better response, ascending supergradient dynamics ISBRAG that seeks the DRoNE's when the distributionally robust game possesses amicable supergradients. This forms the basis of a distributed version (d-ISBRAG) where agents estimate others' strategies by means of a dynamic consensus subroutine over a directed communication network.

What carries the argument

The inertial, supported, better-response, ascending supergradient dynamics (ISBRAG) that uses supergradients of the worst-case utility to drive strategy updates toward the DRoNE set when amicable supergradients are present.

If this is right

  • The DRoNE set is nonempty under the conditions supplied in the paper.
  • The DRoNE set remains close to the Nash equilibria of the underlying stochastic game.
  • ISBRAG dynamics converge to the DRoNE set whenever amicable supergradients exist.
  • The distributed d-ISBRAG algorithm enables agents to seek the DRoNE set using only local communication over directed networks.
  • A tractable reformulation of the distributionally robust problem permits distributed computation of the required supergradients.

Where Pith is reading between the lines

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

  • The closeness characterization implies that DRoNE converge to stochastic NE as the sample size grows and the Wasserstein radius shrinks.
  • The consensus-based distributed implementation could be adapted to time-varying communication graphs common in mobile multi-agent systems.
  • The same supergradient framework might be tested on repeated or continuous-time versions of the one-shot game.

Load-bearing premise

The distributionally robust game must possess amicable supergradients for the proposed dynamics to be guaranteed to seek the DRoNE set.

What would settle it

A concrete two-player stochastic game that satisfies all stated conditions except the amicable-supergradients property, together with a simulation showing that ISBRAG fails to reach any DRoNE, would falsify the convergence claim.

Figures

Figures reproduced from arXiv: 2605.15534 by Nirabhra Mandal, Sonia Mart\'inez.

Figure 1
Figure 1. Figure 1: Effect of step-size parameters {αi}i∈A on conver￾gence of ISBRAG. The plots share a common legend. (Top) αi = 0.1, ∀i ∈ A. (Bottom) αi = 0.01, ∀i ∈ A. to a larger set around the NE. On the other hand, with a smaller αi value, the solutions converge closer to the NE, but take more time-steps to get there. Performance of d-ISBRAG: For this scenario, we test our algorithm on a non-monotone game and report (em… view at source ↗
Figure 2
Figure 2. Figure 2: Convergence properties of d-ISBRAG. The plots share [PITH_FULL_IMAGE:figures/full_fig_p019_2.png] view at source ↗
read the original abstract

In this work, we study stochastic one-shot games where agents' utilities depend on the collective strategy profiles of other agents as well as on some well-behaved randomness. While each decision-maker is agnostic to the random variable's underlying distribution, they have access to finitely many i.i.d. samples generated from it. We consider two cases: one where samples are shared; and another, more special one, where samples are individually accessible. To hedge against the unknown uncertainty, each agent plays a distributionally robust game and aims to maximize the worst-case expected utility over a Wasserstein ball around the sample average distribution. In this setting, we provide conditions under which the game has a non-empty set of distributionally robust Nash equilibria (DRoNE) and then characterize the closeness of the DRoNE set to the Nash equilibria (NE) of the associated stochastic game. We then propose an inertial, supported, better response, ascending supergradient dynamics ISBRAG that seeks the DRoNE's when the distributionally robust game possesses what we term as amicable supergradients. This forms the basis of a distributed version (d-ISBRAG) where agents estimate others' strategies by means of a dynamic consensus subroutine over a directed communication network. While initially the distributed algorithm works in the case where agents have individual samples, we later extend this to the case of shared observations under certain simplifying assumptions. This involves analyzing a tractable reformulation of the distributionally robust optimization problem and solving it in a distributed manner to compute the required supergradients. Simulations illustrate our results.

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

3 major / 2 minor

Summary. The manuscript investigates stochastic one-shot games with utilities depending on strategy profiles and randomness. Agents, having access to finite i.i.d. samples but agnostic to the true distribution, formulate distributionally robust games using Wasserstein balls centered at the empirical distribution. The authors establish conditions for the existence of a non-empty set of distributionally robust Nash equilibria (DRoNE) and characterize the proximity of this set to the Nash equilibria of the underlying stochastic game. They introduce inertial supported better-response ascending supergradient (ISBRAG) dynamics that converge to DRoNE under the assumption of amicable supergradients, and extend this to a distributed version (d-ISBRAG) over directed networks using dynamic consensus, with extensions to shared samples under simplifying assumptions. Simulations are provided to illustrate the results.

Significance. If the central claims hold, this paper makes a valuable contribution to the intersection of distributionally robust optimization and game-theoretic equilibrium seeking in multi-agent systems with uncertainty. The provision of existence conditions and closeness characterization for DRoNE sets offers theoretical insights into robustness in stochastic games. The proposed dynamics, particularly the distributed variant, could have implications for practical implementation in networked systems. However, the significance is tempered by the reliance on the 'amicable supergradients' property, whose compatibility with the existence conditions is not explicitly verified.

major comments (3)
  1. [Section on ISBRAG dynamics proposal] The convergence of the ISBRAG dynamics to the DRoNE set is established only when the distributionally robust game possesses amicable supergradients. However, it is not shown whether the conditions provided for the non-emptiness of the DRoNE set guarantee or are consistent with this property. This is a load-bearing assumption for the algorithmic contribution.
  2. [Distributed d-ISBRAG extension] For the case of shared samples, the extension to d-ISBRAG relies on 'certain simplifying assumptions' for the tractable reformulation of the DRO problem and distributed supergradient computation. These assumptions are not stated explicitly enough to assess their restrictiveness or compatibility with the earlier existence and closeness results.
  3. [Existence and characterization results] The abstract asserts existence conditions for non-empty DRoNE sets and a closeness characterization to stochastic NE, but the manuscript provides no derivation details or verification for these technical properties (such as properties of worst-case expected utilities under Wasserstein ambiguity). This undermines assessment of the foundational claims.
minor comments (2)
  1. [ISBRAG dynamics] The term 'amicable supergradients' is introduced without a highlighted definition or equation number, making it difficult to cross-reference in the convergence analysis.
  2. [Numerical experiments] Simulation figures would benefit from explicit labeling of how parameter choices satisfy the theoretical conditions for DRoNE existence and amicable supergradients.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their detailed and constructive comments on our manuscript. We address each major comment point by point below. Where revisions are needed to improve clarity or add discussion, we will incorporate them in the revised version.

read point-by-point responses

  1. Referee: The convergence of the ISBRAG dynamics to the DRoNE set is established only when the distributionally robust game possesses amicable supergradients. However, it is not shown whether the conditions provided for the non-emptiness of the DRoNE set guarantee or are consistent with this property. This is a load-bearing assumption for the algorithmic contribution.

    Authors: We acknowledge that the amicable supergradients property is an additional assumption required specifically for the convergence analysis of ISBRAG and is not automatically implied by the existence conditions for non-empty DRoNE sets. The existence results rely on continuity and quasi-concavity of the worst-case utilities, while amicable supergradients ensure alignment for the ascending dynamics. These are compatible under standard smoothness assumptions on the utilities, which we will now explicitly state and verify with a brief discussion and example in the revised manuscript to clarify their relationship. revision: yes

  2. Referee: For the case of shared samples, the extension to d-ISBRAG relies on 'certain simplifying assumptions' for the tractable reformulation of the DRO problem and distributed supergradient computation. These assumptions are not stated explicitly enough to assess their restrictiveness or compatibility with the earlier existence and closeness results.

    Authors: We agree that the simplifying assumptions for the shared-samples case of d-ISBRAG require more explicit statement. These include identical ambiguity sets across agents and a structure on the worst-case distributions permitting closed-form supergradient expressions. We will revise the manuscript to list these assumptions clearly in the relevant section, prove that they preserve the non-emptiness of DRoNE and the closeness characterization, and discuss their restrictiveness relative to the general individual-samples case. revision: yes

  3. Referee: The abstract asserts existence conditions for non-empty DRoNE sets and a closeness characterization to stochastic NE, but the manuscript provides no derivation details or verification for these technical properties (such as properties of worst-case expected utilities under Wasserstein ambiguity). This undermines assessment of the foundational claims.

    Authors: The existence conditions and closeness characterization are derived in Section 3 using duality for Wasserstein DRO and fixed-point arguments for the resulting game. However, to address the concern that these details may not be sufficiently prominent, we will expand the main text with additional proof sketches, highlight the key properties of the worst-case utilities (e.g., continuity and monotonicity with respect to the radius), and move relevant technical lemmas from the appendix into the body for easier verification. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation remains self-contained

full rationale

The paper first states conditions for non-empty DRoNE sets and their distance to stochastic NE using Wasserstein DRO and standard stochastic-game arguments. It then introduces ISBRAG dynamics whose convergence is explicitly conditioned on the separate technical property of amicable supergradients (defined to guarantee ascent compatibility). The distributed d-ISBRAG version adds a consensus subroutine under simplifying assumptions. None of these steps reduce a claimed result to a fitted parameter, self-citation chain, or definitional tautology; each layer adds independent content (existence, proximity, and conditional convergence) without the output being forced by construction from the inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract supplies no explicit free parameters, axioms, or invented entities; the work relies on standard assumptions from stochastic games, Wasserstein DRO, and consensus algorithms.

pith-pipeline@v0.9.0 · 5824 in / 1173 out tokens · 92043 ms · 2026-05-19T14:51:25.531934+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

56 extracted references · 56 canonical work pages

  1. [1]

    T. Basar. Lecture notes on non-cooperative game theory. Game Theory Module of the Graduate Program in Network Mathematics, pages 3–6, 2010

    work page 2010

  2. [2]

    Narahari.Game theory and mechanism design, volume 4

    Y. Narahari.Game theory and mechanism design, volume 4. World Scientific, 2014

    work page 2014

  3. [3]

    J. R. Marden, G. Arslan, and J. S. Shamma. Cooperative control and potential games.IEEE Transactions on Systems, Man, & Cybernetics. Part B: Cybernetics, 39(6):1393–1407, 2009

    work page 2009

  4. [4]

    Zhu and S

    M. Zhu and S. Martínez. Distributed coverage games for energy-aware mobile sensor networks.SIAM Journal on Control and Optimization, 51(1):1–27, 2013

    work page 2013

  5. [5]

    Pan and L

    Y. Pan and L. Pavel. Games with coupled propagated constraints in optical networks with multi-link topologies. Automatica, 45(4):871–880, 2009

    work page 2009

  6. [6]

    W. Saad, Z. Han, H. V. Poor, and T. Başar. Game- theoretic methods for the smart grid: An overview of microgrid systems, demand-side management, and smart grid communications.IEEE Signal Processing Magazine, 29(5):86–105, 2012

    work page 2012

  7. [7]

    Ghaderi and R

    J. Ghaderi and R. Srikant. Opinion dynamics in social networks with stubborn agents: Equilibrium and convergence rate.Automatica, 50(12):3209–3215, 2014. 20

    work page 2014

  8. [8]

    Wankhede, N

    P. Wankhede, N. Mandal, S. Martínez, and P. Tallapragada. Opinion dynamics for utility maximizing agents: exploring the impact of resource penalty.IEEE Transactions on Control of Network Systems, 12(1):5–17, 2025

    work page 2025

  9. [9]

    Zhu and T

    Q. Zhu and T. Başar. Revisiting game-theoretic control in socio-technical networks: Emerging design frameworks and contemporary applications.IEEE Control Systems Letters, 2025

    work page 2025

  10. [10]

    Roughgarden and I

    T. Roughgarden and I. Talgam-Cohen. Approximately optimal mechanism design.Annual Review of Economics, 11(1):355–381, 2019

    work page 2019

  11. [11]

    Hutchinson, B

    S. Hutchinson, B. Turan, and M. Alizadeh. Safe pricing mechanisms for distributed resource allocation with bandit feedback.IEEE Transactions on Control of Network Systems, 11(4):2010–2021, 2024

    work page 2010

  12. [12]

    J. S. Shamma and G. Arslan. Dynamic fictitious play, dynamic gradient play, and distributed convergence to Nash equilibria.IEEE Transactions on Automatic Control, 50(3):312–327, 2005

    work page 2005

  13. [13]

    Frihauf, M

    P. Frihauf, M. Krstic, and T. Basar. Nash equilibrium seeking for games with non-quadratic payoffs. InIEEE Conf. on Decision and Control,pages881–886,Atlanta,USA, December 2010

    work page 2010

  14. [14]

    Ye and G

    M. Ye and G. Hu. Distributed Nash equilibrium seeking by a consensus based approach.IEEE Transactions on Automatic Control, 62(9):4811–4818, 2017

    work page 2017

  15. [15]

    M Ye, Q. Han, L. Ding, and S. Xu. Distributed Nash equilibrium seeking in games with partial decision information: A survey.Proceedings of the IEEE, 111(2):140– 157, 2023

    work page 2023

  16. [16]

    Koshal, A

    J. Koshal, A. Nedić, and U. V. Shanbhag. A gossip algorithm for aggregative games on graphs. InIEEE Conf. on Decision and Control, pages 4840–4845, 2012

    work page 2012

  17. [17]

    Salehisadaghiani and L

    F. Salehisadaghiani and L. Pavel. Distributed Nash equilibrium seeking: A gossip-based algorithm.Automatica, 72:209–216, 2016

    work page 2016

  18. [18]

    Nedić, A

    A. Nedić, A. Ozdaglar, and P. A. Parrilo. Constrained consensus and optimization in multi-agent networks.IEEE Transactions on Automatic Control, 55(4):922–938, 2010

    work page 2010

  19. [19]

    S.S. Kia, B. Van Scoy, J. Cortés, R.A. Freeman, K.M. Lynch, and S. Martínez. Tutorial on dynamic average consensus. The problem, its applications, the algorithms.IEEE Control Systems Magazine, 39(3):40 – 72, 2018

    work page 2018

  20. [20]

    Gadjov and L

    D. Gadjov and L. Pavel. A passivity-based approach to Nash equilibrium seeking over networks.IEEE Transactions on Automatic Control, 64(3):1077–1092, 2018

    work page 2018

  21. [21]

    C. D. Persis and S. Grammatico. Distributed averaging integral Nash equilibrium seeking on networks.Automatica, 110:108548, 2019

    work page 2019

  22. [22]

    A Kannan and U. V. Shanbhag. Distributed iterative regularization algorithms for monotone Nash games. InIEEE Conf. on Decision and Control, pages 1963–1968, 2010

    work page 1963

  23. [23]

    R. T. Rockafellar. Variational analysis of nash equilibrium. Vietnam Journal of Mathematics, 46(1):73–85, 2018

    work page 2018

  24. [24]

    Y.W. Chen, C. Kizilkale, and M. Arcak. Solving monotone variationalinequalitieswithbestresponsedynamics. InIEEE Conf. on Decision and Control, pages 1751–1756, 2024

    work page 2024

  25. [25]

    Cunis and I

    T. Cunis and I. Kolmanovsky. Input-to-state stability of a bilevel proximal gradient descent algorithm.IF AC Papers Online, 56(2):7474–7479, 2023

    work page 2023

  26. [26]

    Shapiro, D

    A. Shapiro, D. Dentcheva, and A. Ruszczyński.Lectures on Stochastic Programming: Modeling and Theory, volume 16. SIAM, Philadelphia, PA, 2014

    work page 2014

  27. [27]

    Ben-Tal, L

    A. Ben-Tal, L. El Ghaoui, and A. Nemirovski.Robust optimization. Princeton University Press, 2009

    work page 2009

  28. [28]

    S. Kim, R. Pasupathy, and S. G. Henderson. A guide to sample average approximation.Handbook of simulation optimization, pages 207–243, 2014

    work page 2014

  29. [29]

    H. E. Scarf, K.J. Arrow, and S. Karlin. A min-max solution of an inventory problem. Technical report, Rand Corporation Santa Monica, 1957

    work page 1957

  30. [30]

    Bertsimas and M

    D. Bertsimas and M. Sim. The price of robustness. Operations Research, 52(1):35–53, 2004

    work page 2004

  31. [31]

    Mohajerin Esfahani and D

    P. Mohajerin Esfahani and D. Kuhn. Data-driven distributionally robust optimization using the Wasserstein metric:performanceguaranteesandtractablereformulations. Mathematical Programming, 171(1-2):115–166, 2018

    work page 2018

  32. [32]

    C. Qu, H. Jia, and P. You. Decision-dependent distributionally robust optimization with application to dynamic pricing. InIEEE Conf. on Decision and Control, pages 693–698, 2025

    work page 2025

  33. [33]

    Li and S

    D. Li and S. Martínez. Online optimization and data assimilation with performance guarantees.IEEE Transactions on Automatic Control, 66(5):2115–2129, 2020

    work page 2020

  34. [34]

    Boskos, J

    D. Boskos, J. Cortés, and S. Martínez. High-confidence data- driven ambiguity sets for time-varying linear systems.IEEE Transactions on Automatic Control, 69(2):797–812, 2023

    work page 2023

  35. [35]

    S. Qu, D. Meng, Y. Zhou, and Y. Dai. Distributionally robust games with an application to supply chain.Journal of Intelligent & Fuzzy Systems, 33(5):2749–2762, 2017

    work page 2017

  36. [36]

    Y. Liu, H. Xu, S. S. Yang, and J. Zhang. Distributionally robust equilibrium for continuous games: Nash and Stackelberg models.European Journal of Operational Research, 265(2):631–643, 2018

    work page 2018

  37. [37]

    Gao and H

    J. Gao and H. Tembine. Distributionally robust games: Wasserstein metric. InInt. Joint Conf. on Neural Networks, pages 1–8, 2018

    work page 2018

  38. [38]

    G. Peng, T. Zhang, and Q. Zhu. A data-driven distributionally robust game using Wasserstein distance. In Int. Conf. on Decision and Game Theory for Security, pages 405–421. Springer, 2020

    work page 2020

  39. [39]

    Fabiani and B

    F. Fabiani and B. Franci. On distributionally robust generalized Nash games defined over the Wasserstein ball. Journal of Optimization Theory & Applications, 199(1):298– 309, 2023

    work page 2023

  40. [40]

    Z. Wang, G. Pantazis, S. Grammatico, M. M. Zavlanos, and K. H. Johansson. Wasserstein distributionally robust nash equilibrium seeking with heterogeneous data: A lagrangian approach.arXiv preprint arXiv:2511.14048, 2025

    work page arXiv 2025

  41. [41]

    Lanzetti, S

    N. Lanzetti, S. Fricker, S. Bolognani, F. Dörfler, and D. Paccagnan. Strategically robust game theory via optimal transport.arXiv preprint arXiv:2507.15325, 2025

    work page arXiv 2025

  42. [42]

    L. Liu, S. Liu, H. Xu, and D. E. Quevedo. Incomplete- information dynamic Stackelberg equilibrium seeking by a distributed distributionally robust feedback approach.IEEE Transactions on Cybernetics, 2025

    work page 2025

  43. [43]

    Mandal, A

    N. Mandal, A. Suresh, C. Nieto, and S. Martínez. Behaviorally heterogeneous multi-agent exploration using distributed task allocation.IEEE Robotics and Automation Letters, pages 2983 – 2990, 2026

    work page 2026

  44. [44]

    Santambrogio.Optimal Transport for Applied Mathematicians

    F. Santambrogio.Optimal Transport for Applied Mathematicians. Springer, 2015

    work page 2015

  45. [45]

    Boskos, J

    D. Boskos, J. Cortés, and S. Martínez. Dynamic evolution of distributional ambiguity sets and precision tradeoffs in data assimilation. InEuropean Control Conference, pages 2252– 2257, Naples, Italy, June 2019. 21

    work page 2019

  46. [46]

    E. Sontag. Remarks on input to state stability of perturbed gradient flows motivated by model-free feedback control learning.Systems and Control Letters, 161:105138, 2022

    work page 2022

  47. [47]

    C. M. Kellett and P. M. Dower. A generalization of input- to-state stability. InIEEE Conf. on Decision and Control, pages 2970–2975, 2012

    work page 2012

  48. [48]

    Noroozi, R

    N. Noroozi, R. Geiselhart, L. Grüne, B. S. Rüffer, and F. R. Wirth. Nonconservative discrete-time iss small-gain conditions for closed sets.IEEE Transactions on Automatic Control, 63(5):1231–1242, 2017

    work page 2017

  49. [49]

    D Sontag and Y

    E. D Sontag and Y. Wang. Notions of input to output stability.Systems and Control Letters, 38(4):235–248, 1999

    work page 1999

  50. [50]

    R. Diestel. Graph theory.Graduate Texts in Mathematics, pages 173–207, 2017

    work page 2017

  51. [51]

    H. L. Royden and P. Fitzpatrick.Real analysis, volume 32. Macmillan New York, 1988

    work page 1988

  52. [52]

    Clason.Introduction to functional analysis

    C. Clason.Introduction to functional analysis. Springer Nature, 2020

    work page 2020

  53. [53]

    Boyd and L

    S. Boyd and L. Vandenberghe.Convex Optimization. Cambridge University Press, 2004

    work page 2004

  54. [54]

    Jiang, E

    Z. Jiang, E. Sontag, and Y. Wang. Input-to-state stability for discrete-time nonlinear systems.IF AC Proceedings and Volumes, 32(2):2403–2408, 1999

    work page 1999

  55. [55]

    J. M. Danskin.The theory of max-min and its application to weapons allocation problems, volume 5. Springer Science & Business Media, 2012

    work page 2012

  56. [56]

    Cherukuri and J

    A. Cherukuri and J. Cortés. Distributed algorithms for convex network optimization under non-sparse equality constraints. InAllerton Conf. on Communications, Control and Computing, pages 452–459, Monticello, IL, September 2016. 22

    work page 2016