pith. sign in

arxiv: 2604.13836 · v1 · submitted 2026-04-15 · 🧮 math.OC · math.PR

Potential Games on Unimodular Random Graphs

Eyal Neuman , Sturmius Tuschmann This is my paper

Pith reviewed 2026-05-10 12:34 UTC · model grok-4.3

classification 🧮 math.OC math.PR
keywords potential gamesunimodular random graphsNash equilibriamass transport principlequenched equilibriathermodynamic limitrandom graphsnetwork games
0
0 comments X

The pith

Minimizers of the global potential on unimodular random graphs are exactly the quenched Nash equilibria under convexity.

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

Potential games are studied on unimodular random graphs of bounded degree, where each player interacts locally with neighbors. A global potential is constructed using the unimodular measure to handle both finite and infinite player settings uniformly. The mass transport principle connects changes in this potential to the best-response condition for a typical root player. With convexity assumptions on the payoffs, the minimizers of the potential are shown to be exactly the quenched Nash equilibria, with the converse also true. The potential is further shown to converge in the thermodynamic limit for weakly convergent sequences of such graphs.

Core claim

Using the unimodular measure, a well-posed global potential is defined for potential games on unimodular random graphs that works for both finite- and infinite-player games. The mass-transport principle identifies the first variation of this potential with the first-order condition of a representative root player. Under suitable convexity assumptions, minimizers of the potential coincide with quenched Nash equilibria, and conversely. The thermodynamic limit of the potential is established along weakly convergent sequences of unimodular measures. Examples include semi-explicit equilibria in linear-quadratic games via the Green kernel and in convex settings via nonlinear Poisson equations.

What carries the argument

The global potential defined on configurations via the unimodular measure, with its first variation identified to the root player's Nash condition by the mass-transport principle.

Load-bearing premise

The assumption that the games admit a well-posed global potential via the unimodular measure and that suitable convexity conditions hold so that potential minimizers are exactly the quenched Nash equilibria.

What would settle it

An explicit example on the infinite d-regular tree for a convex linear-quadratic game where a candidate equilibrium minimizes the potential but fails the Nash condition for some players, or vice versa.

Figures

Figures reproduced from arXiv: 2604.13836 by Eyal Neuman, Sturmius Tuschmann.

Figure 1
Figure 1. Figure 1: Left: the n-cycle Cn on [n] (for n = 30). Right: its local weak limit, the two-way infinite path P ∼= Z with a root (red). Example 2.13 (Boxes in Z d ). Fix d ∈ N. For each n ∈ N, let Gn be the nearest-neighbor box graph on V (Gn) = [n] d := {1, . . . , n} d , where two vertices u = (u1, . . . , ud) and v = (v1, . . . , vd) are adjacent if and only if there exists a unique index i ∈ {1, . . . , d} such tha… view at source ↗
Figure 2
Figure 2. Figure 2: Illustration of the case d = 2. Left: the nearest-neighbor box graph Gn on [n] 2 (for n = 5). Right: its local weak limit, the two-dimensional lattice Z 2 with a root (red). 9 [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Illustration of the case D = 4. Left: a transitive realization of the random 4- regular graph Rn on [n] (for n = 8). Middle: a non-transitive realization of the random 4-regular graph Rn on [n] (for n = 8). Right: the local weak limit of {Rn}n∈N given by the infinite 4-regular tree T4 with a root (red). The above examples demonstrate the strength of the representative-player formulation. By identifying pla… view at source ↗
read the original abstract

We study potential games on unimodular random graphs of bounded degree, where players interact through the underlying network. Using the unimodular measure, we define a well-posed global potential that captures both finite- and infinite-player games. A key observation is that the mass-transport principle identifies the first variation of this potential with the first-order condition of a representative (root) player. Under suitable convexity assumptions, we prove that minimizers of the potential coincide with quenched Nash equilibria, and conversely. We also establish the thermodynamic limit of the potential along weakly convergent sequences of unimodular measures. Finally, we present examples with semi-explicit equilibrium descriptions. In linear-quadratic games on unimodular graphs, equilibria are expressed in terms of the Green kernel of the simple random walk operator, while in convex settings, equilibria are characterized by solutions to nonlinear Poisson equations.

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

0 major / 3 minor

Summary. The paper studies potential games on unimodular random graphs of bounded degree. It uses the unimodular measure to define a global potential that applies to both finite- and infinite-player settings. The mass-transport principle is invoked to identify the first variation of this potential with the first-order condition of a representative root player. Under suitable convexity assumptions, the paper proves that minimizers of the potential coincide with quenched Nash equilibria (and conversely). It further establishes the thermodynamic limit of the potential along weakly convergent sequences of unimodular measures and provides two families of examples: linear-quadratic games whose equilibria are expressed via the Green kernel of the simple random walk, and convex games characterized by solutions to nonlinear Poisson equations.

Significance. If the central equivalence holds, the work supplies a variational characterization of equilibria for network games on random graphs that bridges finite and infinite regimes. The thermodynamic-limit result and the semi-explicit descriptions in the linear-quadratic and convex cases are concrete strengths that could facilitate analysis in applications. The approach correctly leverages the mass-transport principle on unimodular graphs and avoids circularity by defining the potential directly from the payoffs.

minor comments (3)
  1. The precise statement of the convexity assumptions (e.g., strict convexity of the potential or convexity of the payoff functions) should be isolated in a dedicated definition or proposition rather than being referenced only in the abstract and the main theorem statement.
  2. In the linear-quadratic example, the relation between the Green kernel and the equilibrium strategy is stated but the verification that this strategy indeed satisfies the first-order condition for every root player could be expanded by one or two lines of direct computation.
  3. Notation for the unimodular measure and the quenched probability space is introduced without an explicit comparison table to the corresponding objects in the finite-player case; adding such a table would improve readability.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive assessment of the manuscript, including the recognition that the central equivalence provides a variational characterization bridging finite and infinite regimes, and for recommending minor revision. No major comments were raised in the report.

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The derivation defines the global potential directly from payoffs via the unimodular measure and invokes the standard external mass-transport principle to equate its first variation to the root player's first-order condition. Under stated convexity assumptions, this yields the equivalence to quenched Nash equilibria without any self-definitional loops, fitted inputs renamed as predictions, or load-bearing self-citations. The thermodynamic limit and example characterizations (Green kernel, nonlinear Poisson) follow from the same variational setup and are not forced by construction from the inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claims rest on the standard properties of unimodular measures and convexity assumptions for the payoff functions; no free parameters or new entities are introduced in the abstract.

axioms (1)
  • standard math Unimodular random graphs of bounded degree satisfy the mass-transport principle
    Invoked to equate the first variation of the global potential with the first-order condition of a representative root player.

pith-pipeline@v0.9.0 · 5440 in / 1261 out tokens · 44550 ms · 2026-05-10T12:34:38.053080+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

46 extracted references · 46 canonical work pages

  1. [1]

    Aldous and R

    D. Aldous and R. Lyons. Processes on unimodular random networks.Electronic Journal of Probability, 12:1454–1508, 2007

    work page 2007

  2. [2]

    Aldous and J

    D. Aldous and J. M. Steele. The objective method: probabilistic combinatorial opti- mization and local weak convergence. InProbability on Discrete Structures, pages 1–72. Springer, 2004

    work page 2004

  3. [3]

    Aurell, R

    A. Aurell, R. Carmona, and M. Laurière. Stochastic graphon games: II. the linear- quadratic case.Applied Mathematics & Optimization, 85(3):39, 2022

    work page 2022

  4. [4]

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

    work page 2017

  5. [5]

    Benjamini and O

    I. Benjamini and O. Schramm. Recurrence of distributional limits of finite planar graphs.Electronic Journal of Probability, 6:1–13, 2001

    work page 2001

  6. [6]

    Benjamini, R

    I. Benjamini, R. Lyons, and O. Schramm. Unimodular random trees.Ergodic Theory and Dynamical Systems, 35(2):359–373, 2015

    work page 2015

  7. [7]

    Bordenave

    C. Bordenave. Lecture notes on random graphs and probabilistic combinatorial opti- mization. Lecture notes, 2016

    work page 2016

  8. [8]

    Busoniu, R

    L. Busoniu, R. Babuska, and B. De Schutter. A comprehensive survey of multiagent reinforcement learning.IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews, 38(2):156–172, 2008

    work page 2008

  9. [9]

    P. E. Caines and M. Huang. Graphon mean field games and the GMFG equations:ε- Nash equilibria. InIEEE 58th Conference on Decision and Control (CDC 2019), pages 286–292, 2019

    work page 2019

  10. [10]

    P. E. Caines and M. Huang. Graphon mean field games and their equations.SIAM Journal on Control and Optimization, 59(6):4373–4399, 2021

    work page 2021

  11. [11]

    Carmona, D

    R. Carmona, D. Cooney, C. Graves, and M. Lauriere. Stochastic graphon games: I. the static case.Mathematics of Operations Research, 47(1):750–778, 2022. 32

    work page 2022

  12. [12]

    X. Di, A. Hu, Z. Wang, and Y. Zhang.α-potential games for decentralized control of connected and automated vehicles. Preprint, 2025

    work page 2025

  13. [13]

    Dugundji

    J. Dugundji. An extension of Tietze’s theorem.Pacific Journal of Mathematics, 1(3): 353–367, 1951

    work page 1951

  14. [14]

    Feller.An Introduction to Probability Theory and its Applications, Volume I

    W. Feller.An Introduction to Probability Theory and its Applications, Volume I. John Wiley & Sons, 1968

    work page 1968

  15. [15]

    G. B. Folland.Real Analysis: Modern Techniques and Their Applications. John Wiley & Sons, 1999

    work page 1999

  16. [16]

    Ganguly and K

    A. Ganguly and K. Ramanan. Hydrodynamic limits of non-Markovian interacting particle systems on sparse graphs.Electronic Journal of Probability, 29:1–63, 2024

    work page 2024

  17. [17]

    Guo and Y

    X. Guo and Y. Zhang. Towards an analytical framework for dynamic potential games. SIAM Journal on Control and Optimization, 63(2):1213–1242, 2025

    work page 2025

  18. [18]

    X. Guo, X. Li, and Y. Zhang. Anα-potential game framework forN-player dynamic games.SIAM Journal on Control and Optimization, 63(4):2964–3005, 2025

    work page 2025

  19. [19]

    X. Guo, X. Li, and Y. Zhang. Distributed games with jumps: Anα-potential game approach. Preprint, 2025

    work page 2025

  20. [20]

    Guo and Q

    Y. Guo and Q. Yan. Particle systems with local interactions via hitting times and cascades on graphs. Preprint, 2025

    work page 2025

  21. [21]

    P. Hang, C. Lv, Y. Xing, C. Huang, and Z. Hu. Human-like decision making for autonomous driving: A noncooperative game theoretic approach.IEEE Transactions on Intelligent Transportation Systems, 22(4):2076–2087, 2020

    work page 2076

  22. [22]

    Lacker and A

    D. Lacker and A. Soret. A label-state formulation of stochastic graphon games and approximate equilibria on large networks.Mathematics of Operations Research, 48(4): 1987–2018, 2023

    work page 1987

  23. [23]

    Lacker, K

    D. Lacker, K. Ramanan, and R. Wu. Locally interacting diffusions as Markov random fields on path space.Stochastic Processes and their Applications, 140:81–114, 2021

    work page 2021

  24. [24]

    Lacker, K

    D. Lacker, K. Ramanan, and R. Wu. Marginal dynamics of interacting diffusions on unimodular Galton–Watson trees.Probability Theory and Related Fields, 187(3-4):817– 884, 2023

    work page 2023

  25. [25]

    Lacker, K

    D. Lacker, K. Ramanan, and R. Wu. Local weak convergence for sparse networks of interacting processes.The Annals of Applied Probability, 33(2):843–888, 2023

    work page 2023

  26. [26]

    Lacker, K

    D. Lacker, K. Ramanan, and R. Wu. Marginal dynamics of probabilistic cellular au- tomata on trees. Preprint, 2025

    work page 2025

  27. [27]

    Leng and M

    M. Leng and M. Parlar. Game theoretic applications in supply chain management: a review.INFOR: Information Systems and Operational Research, 43(3):187–220, 2005. 33

    work page 2005

  28. [28]

    M. L. Littman. Markov games as a framework for multi-agent reinforcement learning. InMachine Learning Proceedings 1994, pages 157–163. Elsevier, 1994

    work page 1994

  29. [29]

    Lovász.Large Networks and Graph Limits

    L. Lovász.Large Networks and Graph Limits. American Mathematical Society, 2012

    work page 2012

  30. [30]

    Lovász and B

    L. Lovász and B. Szegedy. Limits of dense graph sequences.Journal of Combinatorial Theory, Series B, 96(6):933–957, 2006

    work page 2006

  31. [31]

    J. R. Marden and J. S. Shamma. Revisiting log-linear learning: Asynchrony, complete- ness and payoff-based implementation.Games and Economic Behavior, 75(2):788–808, 2012

    work page 2012

  32. [32]

    Monderer and L

    D. Monderer and L. S. Shapley. Potential games.Games and Economic Behavior, 14 (1):124–143, 1996

    work page 1996

  33. [33]

    Neuman and S

    E. Neuman and S. Tuschmann. Stochastic graphon games with memory. Preprint, 2024

    work page 2024

  34. [34]

    Neuman and S

    E. Neuman and S. Tuschmann. Stochastic graphon games with interventions. Preprint, 2025

    work page 2025

  35. [35]

    Neuman and S

    E. Neuman and S. Tuschmann. Stochastic games on large sparse graphs. Preprint, 2026

    work page 2026

  36. [36]

    R. I. Oliveira and G. H. Reis. Interacting diffusions on random graphs with diverging average degrees: Hydrodynamics and large deviations.Journal of Statistical Physics, 176(5):1057–1087, 2019

    work page 2019

  37. [37]

    R. I. Oliveira, G. H. Reis, and L. M. Stolerman. Interacting diffusions on sparse graphs: Hydrodynamics from local weak limits.Electronic Journal of Probability, 25:1–35, 2020

    work page 2020

  38. [38]

    Parise and A

    F. Parise and A. Ozdaglar. Analysis and interventions in large network games.Annual Review of Control, Robotics, and Autonomous Systems, 4(1):455–486, 2021

    work page 2021

  39. [39]

    Parise and A

    F. Parise and A. Ozdaglar. Graphon games: A statistical framework for network games and interventions.Econometrica, 91(1):191–225, 2023

    work page 2023

  40. [40]

    Plank and Y

    P. Plank and Y. Zhang. Learning distributed equilibria in linear-quadratic stochastic differential games: Anα-potential approach. Preprint, 2026

    work page 2026

  41. [41]

    K. Ramanan. Interacting stochastic processes on sparse random graphs.Proceedings of the 2022 ICM, 6:4394–4425, 2023

    work page 2022

  42. [42]

    A. Sayedi. Real-time bidding in online display advertising.Marketing Science, 37(4): 553–568, 2018

    work page 2018

  43. [43]

    Tangpi and X

    L. Tangpi and X. Zhou. Optimal investment in a large population of competitive and heterogeneous agents.Finance and Stochastics, 28(2):497–551, 2024

    work page 2024

  44. [44]

    van der Hofstad.Random Graphs and Complex Networks: Volume 2

    R. van der Hofstad.Random Graphs and Complex Networks: Volume 2. Cambridge University Press, 2024. 34

    work page 2024

  45. [45]

    Woess.Random Walks on Infinite Graphs and Groups

    W. Woess.Random Walks on Infinite Graphs and Groups. Cambridge University Press, 2000

    work page 2000

  46. [46]

    Zhang, Z

    K. Zhang, Z. Yang, and T. Başar. Multi-agent reinforcement learning: A selective overview of theories and algorithms.Handbook of Reinforcement Learning and Control, pages 321–384, 2021. 35

    work page 2021