Pith. sign in

REVIEW 4 major objections 5 minor 35 references

The Capacity of Smartphone Peer-to-Peer Networks

T0 review · 4 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read In smartphone peer-to-peer networks, the maximum degree of a minimum-degree spanning tree fixes broadcast and all-to-all throughput, while pairwise capacity is approximable to within 3/2+ε.

desk verdict The first real capacity characterizations for the mobile telephone model, with solid core proofs and two localized, repairable technical slips; the one-shot gossip claim rests on an external to-appear paper. read the letter →

arxiv 1908.01894 v2 pith:3SO3KP6O submitted 2019-08-05 cs.DC

Michael Dinitz , Magnús M. Halldórsson , Calvin Newport , Alex Weaver This is my paper · ORCID
classification cs.DC
keywords mobiletelephonemodelnetworkcapacitypairwisebroadcastall-to-allmulti-commodityflowminimumdegreespanningtreeone-shotgossip
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper establishes tight capacity bounds for three sustained-communication problems in the mobile telephone model, an abstraction of smartphone peer-to-peer links in which a node can hold at most one connection and transfer one packet per round. The central result is that the maximum degree d(G) of the graph's minimum-degree spanning tree fixes the broadcast throughput at Θ(1/d(G)) and the all-to-all throughput at Θ(1/(n d(G))). For pairwise flows, the paper shows that near-optimal throughput can be computed by reducing the scheduling problem to multi-commodity flow and rounding, achieving a (3/2+ε)-approximation of optimal in every connected graph. The same techniques give a one-shot gossip algorithm that is optimal or within a polylogarithmic factor of optimal in every graph, largely settling a previously open question. These bounds matter because they convert a question about smartphone mesh usefulness into a computable graph quantity that can be evaluated before deployment.

What carries the argument

The load-bearing object is d(G), the minimum possible maximum degree of a spanning tree of the network graph. The paper combines a toughness-based separator lemma with the model's single-connection rule: a separator node that can reach several disconnected components can serve at most one component per round, so d(G) lower-bounds per-packet latency. To match the lower bound, the algorithms construct a low-degree spanning tree, edge-color it with O(d(G)) colors, and use the colors as a round-robin schedule to simulate the CONGEST model, turning standard pipelined flooding or broadcast gossip into a mobile-telephone schedule. For pairwise capacity the operative machinery is different: a reduction to maximum concurrent multi-commodity flow in a node-split digraph, a value-preserving rounding lemma, and edge multicoloring to convert fractional flows into a repeating matching schedule.

What would settle it

Run a sustained broadcast experiment on a star-like network with one central node and many leaves. The model's matching constraint caps throughput at 1/(n-1) packets per round; a schedule or real library that sustains c/(n-1) with c>1 concurrent connections per round would refute the single-connection premise on which the Θ(1/d(G)) lower bound rests.

Watch

Extended reading notes

Core claim

In the mobile telephone model, where connections in a round form a matching and each edge transfers one packet, the paper proves that for any connected graph G the optimal sustainable broadcast throughput is Θ(1/d(G)), where d(G) is the smallest k such that G has a spanning tree of maximum degree k. The upper bound comes from a graph-theoretic separator lemma: if d(G)>k, some node set S splits G into more than (d(G)-2)|S| components, and since each separator node can feed only one component per round, every broadcast packet costs Ω(d(G)) rounds. A matching algorithm constructs a spanning tree of degree O(d(G)), edge-colors it with O(d(G)) colors, and cycles through colors to simulate the CONGEST model, in which pipelined flooding gives constant throughput. All-to-all capacity is then Θ(1/(n d(G))), because n sources each require a broadcast and the same separator argument bounds the total. For pairwise capacity the paper gives a polynomial-time centralized algorithm that routes packets within a (3/2+ε) factor of optimal, via a reduction to maximum concurrent multi-commodity flow followed by rounding and an edge-multicoloring schedule. In random geometric networks with radius above the connectivity threshold, the paper shows broadcast capacity is constant and all-to-all capacity is Θ(1/n), both optimal.

Load-bearing premise

The model assumes a node can take part in at most one connection per round, so a single bottleneck node can serve at most one disconnected component in a round; if real smartphone libraries allow several simultaneous connections, the broadcast and all-to-all bounds would loosen.

Editorial extensions

If this is right

  • A network's broadcast usefulness can be read off a single tree statistic: graphs admitting a low-degree spanning tree sustain constant broadcast throughput, while graphs that force high-degree trees are proportionally slower.
  • Centralized planners can compute near-optimal pairwise schedules in polynomial time, so before deployment one can estimate whether an intended demand set is feasible.
  • All-to-all throughput in any connected topology is Θ(1/(n d(G))), so per-device sustained throughput cannot beat a 1/n scaling even on the most favorable graphs.
  • One-shot gossip can now be solved optimally or within a polylog factor in every graph, closing the gap left by expansion-based bounds for a large family of topologies.
  • In random geometric deployments just above the connectivity threshold, broadcast capacity reaches constant throughput, so the mobile telephone model is not inherently a bottleneck for one-to-many distribution.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If real smartphone libraries allow a small constant c>1 concurrent connections per node, the broadcast and all-to-all bounds should scale by roughly c where separators dominate; the paper's model uses c=1 even though its motivating platform allows 8.
  • The degree parameter d(G) may be the right complexity measure for other one-shot primitives in this model, such as rumor spreading or token dissemination, whenever a single node must relay to many disconnected components.
  • The pairwise reduction could be used as an offline network planner: collect the topology, solve the multi-commodity flow, and distribute the repeating schedule, since the setup cost is amortized over an infinite run.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper studies three throughput/capacity problems in the mobile telephone model (MTM). It gives a centralized (3/2+ε)-approximation for pairwise capacity via a reduction to concurrent multi-commodity flow and Shannon edge multicoloring, and proves tight Θ(r) bounds for Gupta-Kumar random graphs above the connectivity threshold. For broadcast capacity it proves an O(1/d(G)) upper bound using graph toughness and a matching Ω(1/d(G)) algorithm based on MDST construction and edge-coloring. For all-to-all capacity it proves Θ(1/(n d(G))) bounds. It also derives lower and upper bounds for one-shot gossip, claiming a polylog-optimal algorithm in every graph, with the upper bound relying on a distributed MDST algorithm from a to-appear paper.

Significance. If the main theorems hold, this is the first capacity treatment of the mobile telephone model and it coherently ties pairwise, broadcast, and all-to-all throughput to the MDST parameter d(G). The MCF-rounding-to-scheduling technique and the toughness-based lower bounds are novel and likely useful beyond this paper. The one-shot gossip corollary would largely resolve an open question from prior work. The main fragility is that the headline one-shot gossip upper bound depends on an external unpublished result, and several proof details need repair; none of the core capacity theorems appear fatally flawed.

major comments (4)
  1. [§5.4, Theorem 5.5] The upper bound of the one-shot gossip result is not self-contained. It depends entirely on the distributed MDST algorithm of [9], cited as 'to appear', and on the assertion in §4.2 that this broadcast-CONGEST algorithm can be run directly in the MTM using O(log n)-bit advertisements. Neither the degree bound O(d(G)+log n), the round bound Õ(D+√n), nor the MTM simulation is proved in this manuscript. Since Theorem 5.5 is advertised as largely resolving an open question, this dependency is load-bearing; the authors should either include a proof or appendix of the required MDST result and its MTM implementation, or explicitly state that the result is conditional on [9].
  2. [§3.2.3, Lemma 3.15] The Chernoff bound contains an arithmetic error. From the stated lower bound r̂ ≥ r/6 and r ≥ β r_c, one obtains exp(-n r̂²/16) ≤ exp(-n(β r_c)²/576), not exp(-n(β r_c)²/96). The stronger bound would require r̂ ≥ r/√6, which is not guaranteed (e.g., r=0.5 gives r̂=0.2, which is less than 0.204). The lemma's high-probability conclusion is still recoverable by using the 576 bound and choosing β larger, but the proof as written is incorrect.
  3. [§3.1, Lemma 3.2] The assertion that 'the value of each flow is trivially bounded from below by v(f_i) ≥ 1/K' is false for an arbitrary feasible MCMF flow, since scaling a feasible flow down by any positive factor preserves feasibility and makes the value arbitrarily small. In the application to Theorem 3.1 the flow is near-optimal and OPT ≥ 1/K because sending 1/K of each commodity along a fixed path is feasible, so the argument is repairable by adding this hypothesis and adjusting N; however, the lemma as stated and proved is not correct.
  4. [§3.1, proof of Theorem 3.1] The bound dr(v) ≤ 2N 'by node capacity constraints' is not justified. For a node that is an endpoint of a commodity and also carries pass-through flow, the sum of f(e) over incident original edges equals 2 f(v_in,v_out) - A(v) - B(v), which can be as large as 2 + ετ/4, so dr(v) can be up to (2+ε/2)N rather than 2N. The discrepancy can be absorbed into the existing ε-approximation, but the displayed derivation as written is incorrect.
minor comments (5)
  1. [§1, Abstract] The abstract says 'wireless network capacity has been active area of research'; it should be 'an active area of research'.
  2. [§3.2.1, Theorem 3.6] The proof text refers to 'a network generated by G(n,k)', which appears to be a typo; it should be 'G ∼ GK(n,r)'.
  3. [§5.2, Figure 3 caption] The caption says broadcast gossip 'terminates in at most αn simulated rounds, for a fixed constant α', but α is not defined in the figure or surrounding text; Theorem 5.2 gives D+n rounds, so the caption should say n+D or define α.
  4. [§4.2, Theorem 4.4] The proof of Theorem 4.4 is only a high-level paragraph; it should specify the failure probability of initialization and give a formal amortization argument that the pipeline fill and setup costs do not affect the asymptotic throughput.
  5. [§3.2.3, Lemma 3.16] The footnote in the proof assumes a model variation with one incoming and one outgoing link per round, whereas the main model in §2.1 allows at most one connection per node per round; since the theorem is stated for the official model, the factor-2 simulation should be incorporated explicitly into the proof or the model definition should be adjusted.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the capacity results are derived from independent graph-theoretic theorems and self-contained proofs; the one self-cited MDST black box supports only the one-shot gossip bound and does not encode the target claims.

full rationale

The paper's central capacity theorems are not circular. Pairwise capacity is obtained by a reduction to concurrent multi-commodity flow (Lemmas 3.2-3.3, Theorem 3.1), using external MCF approximation and Shannon edge-multicoloring; broadcast capacity uses Win's toughness/k-tree theorem (Theorem 2.5) for the upper bound and an MDST-plus-edge-coloring construction for the matching lower bound; all-to-all capacity follows the same in-paper arguments plus the self-contained broadcast-gossip analysis (Theorem 5.2). The quantities d(G), n, and r are graph/input parameters, not fitted values, and the random-network bounds are proved from the GK generation process via concentration and negative association rather than from the claimed throughput. The only author-overlapping citation is the to-appear distributed MDST algorithm [9] used for the one-shot gossip upper bound in Theorem 5.5; it is a parameter-free algorithmic result in broadcast-CONGEST whose stated assumptions do not include the capacity or gossip conclusions, so it is independent support rather than circularity. The unsupported MTM-simulation assertion for [9] and the minor Lemma 3.2 proof gap are correctness risks, not circular reductions. Therefore no step of the derivation is equivalent by construction to its inputs.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The capacity characterizations are derived from the MTM axioms and cited graph-theoretic theorems. There are no fitted parameters or invented entities. The main modeling burden is the one-connection-per-round abstraction; the main external dependencies are Win's theorem, Shannon's theorem, and the cited MDST and MCMF algorithms. Constants such as alpha=1/32 and beta are universal proof constants, not data-fitted parameters.

assumptions (6)
  • domain assumption Mobile telephone model constraints: each node has at most one connection per round, one packet per connection, and O(log n)-bit advertisements.
    All capacity bounds are stated for this model; the one-connection-per-round restriction drives the d(G) and n*d(G) throughput denominators. The paper motivates it via smartphone libraries but does not prove real devices obey it.
  • standard math Win's theorem: if t(G) >= 1/(k-2) then G has a k-tree, with the corollary that d(G) > k implies a separator S with c(G \ S) > (k-2)|S|.
    Used in Theorem 4.1 and Theorem 5.1 to obtain separator sets giving the broadcast and all-to-all upper bounds.
  • standard math Shannon edge multicoloring theorem from [32].
    Converts rounded multicommodity flows into schedules in the pairwise capacity algorithm.
  • standard math Chernoff bounds and negative association of balls-and-bins indicators (Theorems 2.3 and 3.7).
    Used in all random-network high-probability arguments.
  • standard math Existence and efficiency of prior algorithms: MCMF approximation [27], MDST approximation [13], distributed MDST [9], maximal matching [24].
    Used as black boxes for constructing flows, trees, and colorings; the paper does not re-prove them.
  • domain assumption Gupta-Kumar random network process GK(n,r): uniform node placement and edge when distance is at most r, with random pairwise flow sets.
    Defines the random-network setting; theorems characterize throughput for this specific generative process.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The Capacity of Smartphone Peer-to-Peer Networks." pith.science (2026). https://pith.science/paper/3SO3KP6O

@misc{pith2026190801894,
  author       = {Pith},
  title        = {Pith review of: The Capacity of Smartphone Peer-to-Peer Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3SO3KP6O}},
  note         = {Machine review of arXiv:1908.01894}
}
read the original abstract

We study three capacity problems in the mobile telephone model, a network abstraction that models the peer-to-peer communication capabilities implemented in most commodity smartphone operating systems. The capacity of a network expresses how much sustained throughput can be maintained for a set of communication demands, and is therefore a fundamental bound on the usefulness of a network. Because of this importance, wireless network capacity has been active area of research for the last two decades. The three capacity problems that we study differ in the structure of the communication demands. The first problem is pairwise capacity, where the demands are (source, destination) pairs. Pairwise capacity is one of the most classical definitions, as it was analyzed in the seminal paper of Gupta and Kumar on wireless network capacity. The second problem we study is broadcast capacity, in which a single source must deliver packets to all other nodes in the network. Finally, we turn our attention to all-to-all capacity, in which all nodes must deliver packets to all other nodes. In all three of these problems we characterize the optimal achievable throughput for any given network, and design algorithms which asymptotically match this performance. We also study these problems in networks generated randomly by a process introduced by Gupta and Kumar, and fully characterize their achievable throughput. Interestingly, the techniques that we develop for all-to-all capacity also allow us to design a one-shot gossip algorithm that runs within a polylogarithmic factor of optimal in every graph. This largely resolves an open question from previous work on the one-shot gossip problem in this model.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

35 extracted references · 34 canonical work pages

  1. [9]

    Halld\' o rsson, Taisuke Izumi, and Calvin Newport

    Michael Dinitz, Magn\' u s M. Halld\' o rsson, Taisuke Izumi, and Calvin Newport. Distributed minimum degree spanning trees. In Proceedings of the ACM Symposium on the Principles of Distributed Computing (PODC) , 2019 (to appear.)

    work page 2019

  2. [13]

    Approximating the minimum-degree Steiner tree to within one of optimal

    Martin F \"u rer and Balaji Raghavachari. Approximating the minimum-degree Steiner tree to within one of optimal. Journal of Algorithms , 17(3):409 -- 423, 1994

    work page 1994

  3. [35]

    On a connection between the existence of k-trees and the toughness of a graph

    Sein Win. On a connection between the existence of k-trees and the toughness of a graph. Graphs and Combinatorics , 5(1):201--205, 1989

    work page 1989

  4. [1]

    Spontaneous smartphone networks as a user-centric solution for the future internet

    Gianluca Aloi, Marco Di Felice, Valeria Loscr \` , Pasquale Pace, and Giuseppe Ruggeri. Spontaneous smartphone networks as a user-centric solution for the future internet. IEEE Communications Magazine , 52(12):26--33, 2014

    work page 2014

  5. [2]

    Multipeerconnectivity | apple developer documentation

    Apple. Multipeerconnectivity | apple developer documentation. Internet: https://developer.apple.com/documentation/multipeerconnectivity [Accessed : 07/20/2018] , 2018

    work page 2018

  6. [3]

    Device-to-device communications with wi-fi direct: overview and experimentation

    Daniel Camps-Mur, Andres Garcia-Saavedra, and Pablo Serrano. Device-to-device communications with wi-fi direct: overview and experimentation. IEEE Wireless Communications , 20(3):96--104, 2013

    work page 2013

  7. [4]

    Rumour spreading and graph conductance

    Flavio Chierichetti, Silvio Lattanzi, and Alessandro Panconesi. Rumour spreading and graph conductance. In Proceedings of the ACM-SIAM symposium on Discrete Algorithms (SODA) , 2010

    work page 2010

  8. [5]

    Tough graphs and hamiltonian circuits

    Vasek Chv \'a tal. Tough graphs and hamiltonian circuits. Discrete Mathematics , 5(3):215--228, 1973

    work page 1973

Show all 35 references
  1. [6]

    Leader election in shared spectrum radio networks

    Sebastian Daum, Seth Gilbert, Fabian Kuhn, and Calvin Newport. Leader election in shared spectrum radio networks. In Proceedings of the 2012 ACM Symposium on the Principles of Distributed Computing , pages 215--224. ACM, 2012

    2012

  2. [7]

    Rumor spreading with bounded in-degree

    Sebastian Daum, Fabian Kuhn, and Yannic Maus. Rumor spreading with bounded in-degree. In Proceedings of the International Colloquium on Structural Information and Communication Complexity , 2016

    2016

  3. [8]

    Load balancing with bounded convergence in dynamic networks

    Michael Dinitz, Jeremy Fineman, Seth Gilbert, and Calvin Newport. Load balancing with bounded convergence in dynamic networks. In INFOCOM , pages 1--9, 2017

    2017

  4. [10]

    Balls and bins: A study in negative dependence

    Devdatt Dubhashi and Desh Ranjan. Balls and bins: A study in negative dependence. Random Structures & Algorithms , 13(2):99--124, 1998

    1998

  5. [11]

    Rumor spreading on random regular graphs and expanders

    Nikolaos Fountoulakis and Konstantinos Panagiotou. Rumor spreading on random regular graphs and expanders. In Approximation, Randomization, and Combinatorial Optimization. Algorithms and Techniques , pages 560--573. Springer, 2010

    2010

  6. [12]

    The shortest-path problem for graphs with random arc-lengths

    Alan M Frieze and Geoffrey R Grimmett. The shortest-path problem for graphs with random arc-lengths. Discrete Applied Mathematics , 10(1):57--77, 1985

    1985

  7. [14]

    Firechat

    Open Garden. Firechat. Internet: https://www.opengarden.com/ [Accessed: 07/20/2018] , 2018

    2018

  8. [15]

    The open garden hotspot

    Open Garden. The open garden hotspot. Internet: https://www.opengarden.com/ [Accessed: 07/20/2018] , 2018

    2018

  9. [16]

    Multi-message broadcast with abstract mac layers and unreliable links

    Mohsen Ghaffari, Erez Kantor, Nancy Lynch, and Calvin Newport. Multi-message broadcast with abstract mac layers and unreliable links. In Proceedings of the ACM Symposium on the Principles of Distributed Computing , pages 56--65. ACM, 2014

    2014

  10. [17]

    How to discreetly spread a rumor in a crowd

    Mohsen Ghaffari and Calvin Newport. How to discreetly spread a rumor in a crowd. In Proceedings of the International Symposium on Distributed Computing (DISC) , 2016

    2016

  11. [18]

    Tight bounds for rumor spreading in graphs of a given conductance

    George Giakkoupis. Tight bounds for rumor spreading in graphs of a given conductance. In Proceedings of the Symposium on Theoretical Aspects of Computer Science (STACS) , 2011

    2011

  12. [19]

    Tight bounds for rumor spreading with vertex expansion

    George Giakkoupis. Tight bounds for rumor spreading with vertex expansion. In Proceedings of the ACM-SIAM Symposium on Discrete Algorithms (SODA) , 2014

    2014

  13. [20]

    Rumor spreading and vertex expansion

    George Giakkoupis and Thomas Sauerwald. Rumor spreading and vertex expansion. In Proceedings of the ACM-SIAM symposium on Discrete Algorithms (SODA) , pages 1623--1641. SIAM, 2012

    2012

  14. [21]

    Overview and evaluation of bluetooth low energy: An emerging low-power wireless technology

    Carles Gomez, Joaquim Oller, and Josep Paradells. Overview and evaluation of bluetooth low energy: An emerging low-power wireless technology. Sensors , 12(9):11734--11753, 2012

    2012

  15. [22]

    The capacity of wireless networks

    Piyush Gupta and Panganmala R Kumar. The capacity of wireless networks. IEEE Transactions on Information Theory , 46(2):388--404, 2000

    2000

  16. [23]

    Padoc: Enabling social networking in proximity

    Adrian Holzer, Sven Reber, Jonny Quarta, Jorge Mazuze, and Denis Gillet. Padoc: Enabling social networking in proximity. Computer Networks , 111:82--92, 2016

    2016

  17. [24]

    Amos Israeli and A. Itai. A fast and simple randomized parallel algorithm for maximal matching. Inf. Process. Lett. , 22(2):77--80, February 1986

    1986

  18. [25]

    Networking smartphones for disaster recovery

    Zongqing Lu, Guohong Cao, and Thomas La Porta. Networking smartphones for disaster recovery. In Proceedings of the IEEE International Conference on Pervasive Computing and Communications (PerCom) , pages 1--9. IEEE, 2016

    2016

  19. [26]

    M. Luby. A simple parallel algorithm for the maximal independent set problem. SIAM Journal on Computing , 15(4):1036--1053, 1986. http://dx.doi.org/10.1137/0215074 doi:10.1137/0215074

    1986 doi

  20. [27]

    Faster approximation schemes for fractional multicommodity flow problems via dynamic graph algorithms

    Aleksander Madry. Faster approximation schemes for fractional multicommodity flow problems via dynamic graph algorithms. In Proceedings of the Forty-Second ACM Symposium on Theory of Computing (STOC) , pages 121--130. ACM, 2010

    2010

  21. [28]

    Peer-to-peer using multipeer connectivity

    David Mark, Jayant Varma, Jeff LaMarche, Alex Horovitz, and Kevin Kim. Peer-to-peer using multipeer connectivity. In More iPhone Development with Swift , pages 239--280. Springer, 2015

    2015

  22. [29]

    Leader election in a smartphone peer-to-peer network

    Calvin Newport. Leader election in a smartphone peer-to-peer network. In Proceedings of the IEEE International Parallel and Distributed Processing Symposium (IPDPS) , 2017. Full version available online at: http://people.cs.georgetown.edu/ cnewport/pubs/le-IPDPS2017.pdf

    2017

  23. [30]

    Random gossip processes in smartphone peer-to-peer networks

    Calvin Newport and Alex Weaver. Random gossip processes in smartphone peer-to-peer networks. In Proceedings of the International Conference on Distributed Computing in Sensor Systems (DCOSS) , 2019

    2019

  24. [31]

    A survey on multihop ad hoc networks for disaster response scenarios

    DG Reina, Mohamed Askalani, SL Toral, Federico Barrero, Eleana Asimakopoulou, and Nik Bessis. A survey on multihop ad hoc networks for disaster response scenarios. International Journal of Distributed Sensor Networks , 11(10):647037, 2015

    2015

  25. [32]

    A theorem on coloring the lines of a network

    Claude E Shannon. A theorem on coloring the lines of a network. Journal of Mathematics and Physics , 28(1-4):148--152, 1949

    1949

  26. [33]

    Soscast: Location estimation of immobilized persons through sos message propagation

    Noriyuki Suzuki, Jane Louie Fresco Zamora, Shigeru Kashihara, and Suguru Yamaguchi. Soscast: Location estimation of immobilized persons through sos message propagation. In Proceedings of the International Conference on Intelligent Networking and Collaborative Systems (INCoS) ,...

    2012

  27. [34]

    V. G. Vizing. On an estimate of the chromatic class of a p -graph. Diskret. Analiz No. , 3:25--30, 1964

    1964

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.