pith. sign in

arxiv: 2603.07633 · v2 · pith:CD3M5MJDnew · submitted 2026-03-08 · 💻 cs.GT

Coordination Games on Multiplex Networks: Consensus, Convergence, and Stability of Opinion Dynamics

Ruey-An Shiu , Parinaz Naghizadeh This is my paper

Pith reviewed 2026-05-21 12:41 UTC · model grok-4.3

classification 💻 cs.GT
keywords opinion dynamicsmultiplex networkscoordination gamesconsensusconvergence ratesmultilayer networksstability analysisnetwork perturbations
0
0 comments X

The pith

In multiplex networks, connecting layers can create global consensus in opinions where no single layer reaches it, but only if the layers have aligned node degrees.

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

The paper models opinion updates as coordination games where agents on a multilayer network choose opinions to minimize distance from their neighbors. It compares a merged model that combines layer influences with a switching model that alternates between layers. Analysis via random walks and spectral properties shows that interlayer connections can trigger or speed up agreement that isolated layers never achieve, yet can also destroy agreement when layers are dissimilar. A reader cares because this identifies when real social or communication networks promote unified views versus persistent disagreement.

Core claim

This paper shows that multilayer interactions can induce or accelerate global consensus even when no single layer achieves it alone, and conversely that individually coordinated layers may lose consensus once interconnected. Similarity between the layers as captured by alignments in the weighted degrees of the nodes is a main determinant of whether merging or switching can speed up convergence to consensus compared to when the layers operate in isolation.

What carries the argument

The merged model that aggregates layers through weighted influences and the switching model that periodically alternates across layers; both are analyzed with random-walk and spectral methods on the effective graph to obtain consensus conditions and convergence rates.

If this is right

  • Merging layers promotes faster consensus when their weighted degree sequences align closely.
  • Switching between layers can accelerate agreement under similar degree alignments but may destabilize it under misalignment.
  • Spectral gaps of the effective graph give explicit bounds on convergence speed in both coupling schemes.
  • Small perturbations to interlayer weights can shift the system from consensus to persistent disagreement when layers differ in structure.

Where Pith is reading between the lines

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

  • The same degree-alignment condition could predict when adding cross-platform links in social media unifies or fragments public opinion.
  • Testing the model with asynchronous updates would check whether the consensus results hold in more realistic timing regimes.
  • The framework might extend to directed or weighted interlayer edges to model asymmetric influence between network layers.

Load-bearing premise

The model assumes agents update opinions synchronously by minimizing a local cost based on neighbors' current opinions and that random-walk plus spectral analysis on the resulting graph fully describes convergence and stability.

What would settle it

Run simulations of the opinion dynamics on a two-layer network where node degrees are deliberately misaligned across layers and measure whether consensus fails or slows compared with the isolated layers.

Figures

Figures reproduced from arXiv: 2603.07633 by Parinaz Naghizadeh, Ruey-An Shiu.

Figure 1
Figure 1. Figure 1: Merged-layer consensus experiments on random networks. [PITH_FULL_IMAGE:figures/full_fig_p011_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Switching layer consensus experiments on random networks. [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Merged layers opinion dynamics on the high-school contact network [39] for different weighting factors [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Switching layers opinion dynamics on the high-school contact network [39] with different switching period [PITH_FULL_IMAGE:figures/full_fig_p012_4.png] view at source ↗
read the original abstract

This paper studies opinion dynamics in multilayer (social) networks. Extending a single-layer model, we formulate opinion updates as a synchronous coordination game in which agents minimize a local cost to stay close to their neighbors' opinions. We propose two coupling mechanisms between layers: (i) a merged model that aggregates layers through weighted influences, and (ii) a switching model that periodically alternates across layers. Using random-walk and spectral analysis, we derive sufficient conditions for consensus, characterize convergence rates, and analyze stability under network perturbations. We show that multilayer interactions can induce or accelerate global consensus even when no single layer achieves it alone, and conversely, that individually coordinated layers may lose consensus once interconnected. Notably, we show that similarity between the layers (as captured by alignments in the weighted degrees of the nodes) is a main determinant of if merging or switching can speed up convergence to consensus compared to when the layers operate in isolation, providing guidelines for network design interventions.

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 extends single-layer coordination-game opinion dynamics to multiplex networks. Agents synchronously minimize local costs based on neighbors' opinions under two coupling mechanisms: a merged model aggregating layers via weighted influences and a switching model that alternates periodically across layers. Random-walk and spectral analysis on the resulting effective graphs yield sufficient conditions for consensus, explicit characterizations of convergence rates, and stability results under perturbations. Central claims are that multilayer interactions can induce or accelerate global consensus even when no isolated layer achieves it, and conversely that coordinated layers may lose consensus upon interconnection; layer similarity, captured by alignments in weighted node degrees, is presented as the main determinant of whether merging or switching accelerates convergence relative to isolated operation.

Significance. If the derivations hold, the work supplies analytical guidelines for when interlayer coupling promotes or disrupts consensus, with direct relevance to social-network design and intervention. The use of standard spectral tools on explicitly constructed effective graphs is a strength, as is the provision of both positive and negative results (induction of consensus versus loss of it). The paper does not appear to include machine-checked proofs or fully reproducible code, but the random-walk/spectral framing makes the convergence claims falsifiable via the second-largest eigenvalue of the effective transition matrix.

major comments (2)
  1. [§4] §4 (Convergence-rate analysis for merged and switching models): the assertion that alignments in weighted degrees constitute a main determinant of acceleration is load-bearing for the central claim yet rests on the spectrum of the effective graph. Degree sequences alone do not control the second-largest eigenvalue; edge placement, cuts, and higher-order structure also matter. If two layers share identical weighted-degree vectors but differ in connectivity, the merged Laplacian spectrum (and thus the reported speed-up condition) can change without degree misalignment. A counter-example or additional structural assumptions (e.g., regular graphs or identical adjacency matrices up to permutation) should be supplied to substantiate the determinant claim.
  2. [§3.2] §3.2 (Switching model): the periodic alternation is modeled by an effective transition matrix whose spectral gap is used to bound convergence. The derivation assumes that the time-averaged or lifted operator fully captures the long-run rate; however, the paper does not appear to provide explicit error bounds separating the transient from the asymptotic regime or to verify that the synchronous-update assumption remains valid under the chosen coupling weights. This gap affects the quantitative comparison with isolated-layer rates.
minor comments (2)
  1. [Notation] Notation for the effective Laplacian in the merged model (Eq. (8) or equivalent) should explicitly state how the layer-influence weights enter the row-stochastic normalization; current presentation leaves ambiguity about whether the weights are node-specific or layer-specific.
  2. [Figures] Figure captions for the numerical illustrations of convergence trajectories should report the exact parameter values (layer weights, initial opinion vectors) used to generate each panel so that the experiments are reproducible from the text alone.

Simulated Author's Rebuttal

2 responses · 0 unresolved

Thank you for the referee's detailed review and valuable suggestions. We address the major comments point by point below, indicating where revisions will be made to the manuscript.

read point-by-point responses

  1. Referee: [§4] §4 (Convergence-rate analysis for merged and switching models): the assertion that alignments in weighted degrees constitute a main determinant of acceleration is load-bearing for the central claim yet rests on the spectrum of the effective graph. Degree sequences alone do not control the second-largest eigenvalue; edge placement, cuts, and higher-order structure also matter. If two layers share identical weighted-degree vectors but differ in connectivity, the merged Laplacian spectrum (and thus the reported speed-up condition) can change without degree misalignment. A counter-example or additional structural assumptions (e.g., regular graphs or identical adjacency matrices up to permutation) should be supplied to substantiate the determinant claim.

    Authors: We agree that the second-largest eigenvalue depends on the full graph structure beyond degree sequences alone. In our derivation, weighted-degree alignment serves as a sufficient condition for comparing convergence rates in the merged and switching models, leveraging the random-walk stationary distribution. To address the concern, we will revise Section 4 to explicitly qualify the claim as holding under degree alignment (with a note that it is not exhaustive), and we will either supply a brief counter-example showing spectrum variation despite aligned degrees or add a structural assumption such as regularity of the layers. This will be incorporated as a partial revision to strengthen the central claim without overstatement. revision: partial

  2. Referee: [§3.2] §3.2 (Switching model): the periodic alternation is modeled by an effective transition matrix whose spectral gap is used to bound convergence. The derivation assumes that the time-averaged or lifted operator fully captures the long-run rate; however, the paper does not appear to provide explicit error bounds separating the transient from the asymptotic regime or to verify that the synchronous-update assumption remains valid under the chosen coupling weights. This gap affects the quantitative comparison with isolated-layer rates.

    Authors: The effective transition matrix for the switching model is the product of the per-layer transition matrices over a full period, and its spectral gap governs the asymptotic rate. We acknowledge the absence of explicit transient error bounds. In the revision we will add a paragraph in §3.2 deriving a simple bound on the deviation from the asymptotic regime after a finite number of periods, drawing on standard results for periodic Markov chains. The synchronous-update rule holds by construction of the model (each discrete time step applies one layer's neighbors), and we will add a clarifying sentence on the validity under the chosen weights. These additions will support the quantitative comparisons with isolated layers. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation relies on independent spectral analysis of effective graphs

full rationale

The paper derives consensus conditions, convergence rates, and stability results from random-walk and spectral properties (e.g., second-largest eigenvalue of the transition matrix or Laplacian) of the merged or switched effective graph. These follow directly from the defined coupling mechanisms and standard graph theory without reducing to self-definitional fits, parameter renaming, or load-bearing self-citations. The role of degree alignments is presented as a derived sufficient condition on the spectrum rather than an input assumption, and the analysis remains self-contained against external benchmarks of multiplex network spectra. No steps exhibit the enumerated circular patterns.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

Ledger entries are inferred from the abstract description only; full paper would be needed to list all fitted weights or lemmas. The model rests on the standard coordination-game cost minimization and graph-theoretic assumptions for random walks and spectra.

free parameters (1)
  • layer influence weights
    The merged model aggregates layers through weighted influences, requiring selection of relative weights between layers.
axioms (2)
  • domain assumption Agents perform synchronous updates that minimize a local cost to match neighbors' opinions.
    This is the core extension of the single-layer coordination game to the multilayer setting.
  • standard math Random-walk and spectral analysis on the effective coupled graph characterize convergence and stability.
    Invoked to derive sufficient conditions for consensus.

pith-pipeline@v0.9.0 · 5698 in / 1391 out tokens · 62076 ms · 2026-05-21T12:41:09.945986+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

41 extracted references · 41 canonical work pages · 2 internal anchors

  1. [1]

    Dynamics of information diffusion and its applications on complex networks,

    Z.-K. Zhang, C. Liu, X.-X. Zhan, X. Lu, C.-X. Zhang, and Y .-C. Zhang, “Dynamics of information diffusion and its applications on complex networks,”Physics Reports, vol. 651, pp. 1–34, 2016

    work page 2016

  2. [2]

    The structure and dynamics of multilayer networks,

    S. Boccaletti, G. Bianconi, R. Criado, C. I. Del Genio, J. G ´omez- Gardenes, M. Romance, I. Sendina-Nadal, Z. Wang, and M. Zanin, “The structure and dynamics of multilayer networks,”Physics reports, vol. 544, no. 1, pp. 1–122, 2014

    work page 2014

  3. [3]

    Multilayer networks in a nutshell,

    A. Aleta and Y . Moreno, “Multilayer networks in a nutshell,”Annual Review of Condensed Matter Physics, vol. 10, no. 1, pp. 45–62, 2019

    work page 2019

  4. [4]

    Reaching a consensus,

    M. H. DeGroot, “Reaching a consensus,”Journal of the American Statistical association, vol. 69, no. 345, pp. 118–121, 1974

    work page 1974

  5. [5]

    Social influence and opinions,

    N. E. Friedkin and E. C. Johnsen, “Social influence and opinions,” Journal of mathematical sociology, vol. 15, no. 3-4, pp. 193–206, 1990. 12 (a)k= 0 (b)k= 3 (c)k= 5 Fig. 4: Switching layers opinion dynamics on the high-school contact network [39] with different switching periodk+ 1

    work page 1990

  6. [6]

    Gossip algorithms,

    D. Shahet al., “Gossip algorithms,”Foundations and Trends® in Networking, vol. 3, no. 1, pp. 1–125, 2009

    work page 2009

  7. [7]

    Opinion dynamics in social networks with stubborn agents: Equilibrium and convergence rate,

    J. Ghaderi and R. Srikant, “Opinion dynamics in social networks with stubborn agents: Equilibrium and convergence rate,”Automatica, vol. 50, no. 12, pp. 3209–3215, 2014

    work page 2014

  8. [8]

    A tutorial on modeling and analysis of dynamic social networks. part i,

    A. V . Proskurnikov and R. Tempo, “A tutorial on modeling and analysis of dynamic social networks. part i,”Annual Reviews in Control, vol. 43, pp. 65–79, 2017

    work page 2017

  9. [9]

    Group consensus in multi- layer networks,

    F. Sorrentino, L. Pecora, and L. Trajkovi ´c, “Group consensus in multi- layer networks,”IEEE Transactions on Network Science and Engineer- ing, vol. 7, no. 3, pp. 2016–2026, 2020

    work page 2016

  10. [10]

    Public goods game across multilayer populations with different densities,

    C. Wang and C. Sun, “Public goods game across multilayer populations with different densities,”Chaos, Solitons & Fractals, vol. 168, p. 113154, 2023

    work page 2023

  11. [11]

    Stability of multi-layer ecosystems,

    Y . Wang, Y . Yang, A. Li, and L. Wang, “Stability of multi-layer ecosystems,”Journal of the Royal Society Interface, vol. 20, no. 199, p. 20220752, 2023

    work page 2023

  12. [12]

    Weak ergodicity in non-homogeneous markov chains,

    J. Hajnal and M. S. Bartlett, “Weak ergodicity in non-homogeneous markov chains,” inMathematical Proceedings of the Cambridge Philo- sophical Society, vol. 54, no. 2. Cambridge University Press, 1958, pp. 233–246

    work page 1958

  13. [13]

    Exponential convergence of products of stochastic matrices,

    J. M. Anthonisse and H. Tijms, “Exponential convergence of products of stochastic matrices,”Journal of Mathematical Analysis and Applications, vol. 59, no. 2, pp. 360–364, 1977

    work page 1977

  14. [14]

    Products of generalized stochastic sarymsakov matrices,

    W. Xia, J. Liu, M. Cao, K. H. Johansson, and T. Bas ¸ar, “Products of generalized stochastic sarymsakov matrices,” in54th IEEE Conference on Decision and Control (CDC). IEEE, 2015, pp. 3621–3626

    work page 2015

  15. [15]

    Sets of Stochastic Matrices with Converging Products: Bounds and Complexity

    P.-Y . Chevalier, V . V . Gusev, R. M. Jungers, and J. M. Hendrickx, “Sets of stochastic matrices with converging products: bounds and complexity,”arXiv preprint arXiv:1712.02614, 2017

    work page internal anchor Pith review Pith/arXiv arXiv 2017

  16. [16]

    Geometric bounds for eigenvalues of markov chains,

    P. Diaconis and D. Stroock, “Geometric bounds for eigenvalues of markov chains,”The annals of applied probability, pp. 36–61, 1991

    work page 1991

  17. [17]

    Mathematical aspects of mixing times in markov chains,

    R. Montenegro, P. Tetaliet al., “Mathematical aspects of mixing times in markov chains,”Foundations and Trends® in Theoretical Computer Science, vol. 1, no. 3, pp. 237–354, 2006

    work page 2006

  18. [18]

    Br ´emaud,Markov chains: Gibbs fields, Monte Carlo simulation, and queues

    P. Br ´emaud,Markov chains: Gibbs fields, Monte Carlo simulation, and queues. Springer Science & Business Media, 2013, vol. 31

    work page 2013

  19. [19]

    Reversible markov chains and random walks on graphs,

    D. Aldous and J. A. Fill, “Reversible markov chains and random walks on graphs,” 2002, unfinished monograph, recompiled 2014, available at http://www.stat.berkeley.edu/∼aldous/RWG/book.html

    work page 2002

  20. [20]

    D. A. Levin and Y . Peres,Markov chains and mixing times. American Mathematical Soc., 2017, vol. 107

    work page 2017

  21. [21]

    Perturbation theory and finite markov chains,

    P. J. Schweitzer, “Perturbation theory and finite markov chains,”Journal of Applied Probability, vol. 5, no. 2, pp. 401–413, 1968

    work page 1968

  22. [22]

    Entrywise perturbation theory and error analysis for markov chains,

    C. A. O’cinneide, “Entrywise perturbation theory and error analysis for markov chains,”Numerische Mathematik, vol. 65, no. 1, pp. 109–120, 1993

    work page 1993

  23. [23]

    R. A. Horn and C. R. Johnson,Matrix analysis. Cambridge university press, 2012

    work page 2012

  24. [24]

    Opinion dynamics and learning in social networks,

    D. Acemoglu and A. Ozdaglar, “Opinion dynamics and learning in social networks,”Dynamic Games and Applications, vol. 1, pp. 3–49, 2011

    work page 2011

  25. [25]

    Linear Time Average Consensus on Fixed Graphs and Implications for Decentralized Optimization and Multi-Agent Control

    A. Olshevsky, “Linear time average consensus on fixed graphs and implications for decentralized optimization and multi-agent control,” arXiv preprint arXiv:1411.4186, 2014

    work page internal anchor Pith review Pith/arXiv arXiv 2014

  26. [26]

    A theory of non- bayesian social learning,

    P. Molavi, A. Tahbaz-Salehi, and A. Jadbabaie, “A theory of non- bayesian social learning,”Econometrica, vol. 86, no. 2, pp. 445–490, 2018

    work page 2018

  27. [27]

    Opinion dynamics in social networks: A local interaction game with stubborn agents,

    J. Ghaderi and R. Srikant, “Opinion dynamics in social networks: A local interaction game with stubborn agents,” in2013 American control conference. IEEE, 2013, pp. 1982–1987

    work page 2013

  28. [28]

    Binary opinion dynamics with stubborn agents,

    E. Yildiz, A. Ozdaglar, D. Acemoglu, A. Saberi, and A. Scaglione, “Binary opinion dynamics with stubborn agents,”ACM Transactions on Economics and Computation (TEAC), vol. 1, no. 4, pp. 1–30, 2013

    work page 2013

  29. [29]

    Degree-weighted degroot learning,

    C. Cheng, X. Han, X. Tong, Y . Wu, and Y . Xing, “Degree-weighted degroot learning,”Available at SSRN 5195626, 2025

    work page 2025

  30. [30]

    Mathematical formulation of multilayer networks,

    M. De Domenico, A. Sol ´e-Ribalta, E. Cozzo, M. Kivel ¨a, Y . Moreno, M. A. Porter, S. G ´omez, and A. Arenas, “Mathematical formulation of multilayer networks,”Physical Review X, vol. 3, no. 4, p. 041022, 2013

    work page 2013

  31. [31]

    Multilayer networks,

    M. Kivel ¨a, A. Arenas, M. Barthelemy, J. P. Gleeson, Y . Moreno, and M. A. Porter, “Multilayer networks,”Journal of complex networks, vol. 2, no. 3, pp. 203–271, 2014

    work page 2014

  32. [32]

    Multilayer network games: A cooperative approach,

    P. Boruah, S. Borkotokey, and R. Kumar, “Multilayer network games: A cooperative approach,”Network Science, vol. 12, no. 4, pp. 392–403, 2024

    work page 2024

  33. [33]

    United we fall: On the nash equilibria of multiplex and multilayer network games,

    R. Ebrahimi and P. Naghizadeh, “United we fall: On the nash equilibria of multiplex and multilayer network games,”IEEE Transactions on Control of Network Systems, 2025

    work page 2025

  34. [34]

    Evolution of cooperation in multiplex networks,

    J. G ´omez-Gardenes, I. Reinares, A. Arenas, and L. M. Flor´ıa, “Evolution of cooperation in multiplex networks,”Scientific reports, vol. 2, no. 1, p. 620, 2012

    work page 2012

  35. [35]

    Evolutionary games on multilayer networks: a colloquium,

    Z. Wang, L. Wang, A. Szolnoki, and M. Perc, “Evolutionary games on multilayer networks: a colloquium,”The European physical journal B, vol. 88, pp. 1–15, 2015

    work page 2015

  36. [36]

    Stability and robustness of time- varying opinion dynamics: A graph-theoretic approach,

    M. H. Abedinzadeh and E. Akyol, “Stability and robustness of time- varying opinion dynamics: A graph-theoretic approach,”arXiv preprint arXiv:2510.01580, 2025

    work page arXiv 2025

  37. [37]

    Coordination games on multiplex networks: Consensus and convergence of opinion dynamics,

    R.-A. Shiu and P. Naghizadeh, “Coordination games on multiplex networks: Consensus and convergence of opinion dynamics,” in2025 61st Annual Allerton Conference on Communication, Control, and Computing (Allerton). IEEE, 2025, pp. 1–8

    work page 2025

  38. [38]

    On the products of stochastic and diagonal matrices,

    A. Hallak and G. Dalal, “On the products of stochastic and diagonal matrices,”arXiv preprint arXiv:2304.11634, 2023

    work page arXiv 2023

  39. [39]

    Contact patterns in a high school: a comparison between data collected using wearable sensors, contact diaries and friendship surveys,

    R. Mastrandrea, J. Fournet, and A. Barrat, “Contact patterns in a high school: a comparison between data collected using wearable sensors, contact diaries and friendship surveys,”PloS one, vol. 10, no. 9, p. e0136497, 2015

    work page 2015

  40. [40]

    Topology of evolving networks: local events and universality,

    R. Albert and A.-L. Barab ´asi, “Topology of evolving networks: local events and universality,”Physical review letters, vol. 85, no. 24, p. 5234, 2000

    work page 2000

  41. [41]

    Scale-free networks,

    A.-L. Barab ´asi and E. Bonabeau, “Scale-free networks,”Scientific amer- ican, vol. 288, no. 5, pp. 50–9, 2003

    work page 2003