Abstract
The global structure of the minimal spanning tree (MST) is expected to be universal for a large class of underlying random discrete structures. However, very little is known about the intrinsic geometry of MSTs of most standard models, and so far the scaling limit of the MST viewed as a metric measure space has only been identified in the case of the complete graph (Addario-Berry et al. in Ann Probab 45(5):3075–3144, 2017). In this work, we show that the MST constructed by assigning i.i.d. continuous edge weights to either the random (simple) 3-regular graph or the 3-regular configuration model on n vertices, endowed with the tree distance scaled by \(n^{-1/3}\) and the uniform probability measure on the vertices, converges in distribution with respect to Gromov–Hausdorff–Prokhorov topology to a random compact metric measure space. Further, this limiting space has the same law as the scaling limit of the MST of the complete graph identified in Addario-Berry et al. (2017) up to a scaling factor of \(6^{1/3}\). Our proof relies on a novel argument that proceeds via a comparison between a 3-regular configuration model and the largest component in the critical Erdős–Rényi random graph. The techniques of this paper can be used to establish the scaling limit of the MST in the setting of general random graphs with given degree sequences provided two additional technical conditions are verified.
This is a preview of subscription content, access via your institution.
Notes
Here, the superscript ‘\({{\,\mathrm{er}\,}}\)’ is being used to refer to the Erdős–Rényi random graph. The reason behind using this notation will become clear in Sect. 4.3.
References
Abraham, R., Delmas, J.-F., Hoscheit, P.: A note on the Gromov–Hausdorff–Prokhorov distance between (locally) compact metric measure spaces. Electron. J. Probab. 18(14), 1–21 (2013)
Addario-Berry, L., Broutin, N., Goldschmidt, C.: Critical random graphs: limiting constructions and distributional properties. Electron. J. Probab. 15(25), 741–775 (2010)
Addario-Berry, L., Broutin, N., Goldschmidt, C.: The continuum limit of critical random graphs. Probab. Theory Relat. Fields 152(3–4), 367–406 (2012)
Addario-Berry, L., Bhamidi, S., Sen, S.: A probabilistic approach to the leader problem in random graphs. arXiv preprint arXiv:1703.09908 (2017)
Addario-Berry, L., Broutin, N., Goldschmidt, C., Miermont, G.: The scaling limit of the minimum spanning tree of the complete graph. Ann. Probab. 45(5), 3075–3144 (2017)
Addario-Berry, L., Broutin, N., Reed, B.: Critical random graphs and the structure of a minimum spanning tree. Random Struct. Algorithms 35(3), 323–347 (2009)
Aizenman, M., Burchard, A., Newman, C. M., Wilson, D. B.: Scaling limits for minimal and random spanning trees in two dimensions. Random Struct. Algorithms 15(3–4), 319–367 (1999). Statistical physics methods in discrete probability, combinatorics, and theoretical computer science (Princeton, NJ, 1997)
Aldous, D.: The continuum random tree. I. Ann. Probab. 19, 1–28 (1991)
Aldous, D.: The continuum random tree. III. Ann. Probab. 21, 248–289 (1993)
Aldous, D.: A random tree model associated with random graphs. Random Struct. Algorithms 1(4), 383–402 (1990)
Aldous, D.: Brownian excursions, critical random graphs and the multiplicative coalescent. Ann. Probab. 25(2), 812–854 (1997)
Aldous, D., Steele, J.M.: Asymptotics for Euclidean minimal spanning trees on random points. Probab. Theory Relat. Fields 92(2), 247–258 (1992)
Alexander, K.S.: Rates of convergence of means for distance-minimizing subadditive Euclidean functionals. Ann. Appl. Probab. 4(3), 902–922 (1994)
Alexander, K.S.: Percolation and minimal spanning forests in infinite graphs. Ann. Probab. 23(1), 87–104 (1995)
Alexander, K.S.: The RSW theorem for continuum percolation and the CLT for Euclidean minimal spanning trees. Ann. Appl. Probab. 6(2), 466–494 (1996)
Alexander, K.S., Molchanov, S.A.: Percolation of level sets for two-dimensional random fields with lattice symmetry. J. Statist. Phys. 77(3–4), 627–643 (1994)
Avram, F., Bertsimas, D.: The minimum spanning tree constant in geometrical probability and under the independent model: a unified approach. Ann. Appl. Probab. 2(1), 113–130 (1992)
Bai, Z.D., Lee, S., Penrose, M.D.: Rooted edges of a minimal directed spanning tree on random points. Adv. in Appl. Probab. 38(1), 1–30 (2006)
Beardwood, J., Halton, J.H., Hammersley, J.M.: The shortest path through many points. Proc. Cambr. Philos. Soc. 55, 299–327 (1959)
Bender, E.A., Canfield, E.R.: The asymptotic number of labeled graphs with given degree sequences. J. Comb. Theory Ser. A 24(3), 296–307 (1978)
Beveridge, A., Frieze, A., McDiarmid, C.: Random minimum length spanning trees in regular graphs. Combinatorica 18(3), 311–333 (1998)
Bhamidi, S., Broutin, N., Sen, S., Wang, X.: Scaling limits of random graph models at criticality: universality and the basin of attraction of the Erdös–Rényi random graph. arXiv preprint arXiv:1411.3417 (2014)
Bhamidi, S., Dhara, S., Hofstad, R. v. d., Sen, S.: Universality for critical heavy-tailed network models: metric structure of maximal components, arXiv preprint arXiv:1703.07145 (2017)
Bhamidi, S., van der Hofstad, R., Sen, S.: The multiplicative coalescent, inhomogeneous continuum random trees, and new universality classes for critical random graphs. Probab. Theory Relat. Fields 170(1–2), 387–474 (2018)
Bhamidi, S., Sen, S.: Geometry of the minimal spanning tree in the heavy-tailed regime: new universality classes. arXiv preprint arXiv:2009.10696 (2020)
Bhamidi, S., Sen, S.: Geometry of the vacant set left by random walk on random graphs, Wright’s constants, and critical random graphs with prescribed degrees. Random Struct. Algorithms 56(3), 676–721 (2020)
Bhamidi, S., Sen, S., Wang, X.: Continuum limit of critical inhomogeneous random graphs. Probab. Theory Relat. Fields 169(1–2), 565–641 (2017)
Bhatt, A.G., Roy, R.: On a random directed spanning tree. Adv. Appl. Probab. 36(1), 19–42 (2004)
Bollobás, B.: Random Graphs, Cambridge Studies in Advanced Mathematics, vol. 73, 2nd edn. Cambridge University Press, Cambridge (2001)
Bollobás, B., Janson, S., Riordan, O.: The phase transition in inhomogeneous random graphs. Random Struct. Algorithms 31(1), 3–122 (2007)
Braunstein, L.A., Buldyrev, S.V., Cohen, R., Havlin, S., Stanley, H.E.: Optimal paths in disordered complex networks. Phys. Rev. Lett. 91(16), 168701 (2003)
Braunstein, L.A., Wu, Z., Chen, Y., Buldyrev, S.V., Kalisky, T., Sreenivasan, S., Cohen, R., Lopez, E., Havlin, S., Stanley, H.E.: Optimal path and minimal spanning trees in random weighted networks. Int. J. Bifurc. Chaos 17(07), 2215–2255 (2007)
Broutin, N., Duquesne, T., Wang, M.: Limits of multiplicative inhomogeneous random graphs and Lévy trees. arXiv preprint arXiv:1804.05871 (2018)
Broutin, N., Marckert, J.-F.: Asymptotics of trees with a prescribed degree sequence and applications. Random Struct. Algorithms 44(3), 290–316 (2014)
Burago, D., Burago, Y., Ivanov, S.: A Course in Metric Geometry, Graduate Studies in Mathematics, vol. 33. American Mathematical Society, Providence (2001)
Chatterjee, S., Sen, S.: Minimal spanning trees and Stein’s method. Ann. Appl. Probab. 27(3), 1588–1645 (2017)
Conchon-Kerjan, G., Goldschmidt, C.: The stable graph: the metric space scaling limit of a critical randomgraph with IID power-law degrees. arXiv preprint arXiv:2002.04954 (2020)
Dembo, A., Levit, A., Vadlamani, S.: Component sizes for large quantum Erdös–Rényi graph near criticality. Ann. Probab. 47(2), 1185–1219 (2019)
Dhara, S., van der Hofstad, R., van Leeuwaarden, J. S. H., Sen, S.: Critical window for the configuration model: finite third moment degrees. Electron. J. Probab. 22, Paper No. 16, 33 (2017)
Dhara, S., van derHofstad, R., van Leeuwaarden, J.S.H., Sen, S.: Heavy-tailed configuration models at criticality. Ann. Inst. Henri Poincaré Probab. Stat. 56(3), 1515–1558 (2020)
Duquesne, T.: The coding of compact real trees by real valued functions. arXiv preprint math (2006)
Evans, S.N.: Probability and Real Trees. Lecture Notes in Mathematics, vol. 1920. Springer, Berlin (2008)
Federico, L., van derHofstad, R.: Critical window for connectivity in the configuration model. Combin. Probab. Comput. 1–21 (2017)
Fountoulakis, N.: Percolation on sparse random graphs with given degree sequence. Internet Math. 4(4), 329–356 (2007)
Frieze, A., McDiarmid, C.: Algorithmic theory of random graphs. Random Struct. Algorithms 10(1–2), 5–42 (1997)
Frieze, A., Ruszinkó, M., Thoma, L.: A note on random minimum length spanning trees. Electron. J. Combin. 7(1), R41 (2000)
Frieze, A.M.: On the value of a random minimum spanning tree problem. Discrete Appl. Math. 10(1), 47–56 (1985)
Frieze, A.M., McDiarmid, C.J.: On random minimum length spanning trees. Combinatorica 9(4), 363–374 (1989)
Garban, C., Pete, G., Schramm, O.: The scaling limits of the minimal spanning tree and invasion percolation in the plane. Ann. Probab. 46(6), 3501–3557 (2018)
Goldschmidt, C., Haas, B., Sénizergues, D.: Stable graphs: distributions and line-breaking construction. arXiv preprint arXiv:1811.06940 (2018)
Heydenreich, M., van der Hofstad, R.: Random graph asymptotics on high-dimensional tori. Commun. Math. Phys. 270(2), 335–358 (2007)
Heydenreich, M., van der Hofstad, R.: Random graph asymptotics on high-dimensional tori II: volume, diameter and mixing time. Probab. Theory Relat. Fields 149(3–4), 397–415 (2011)
Janson, S.: The minimal spanning tree in a complete graph and a functional limit theorem for trees in a random graph. Random Struct. Algorithms 7(4), 337–355 (1995)
Janson, S.: On percolation in random graphs with given vertex degrees. Electron. J. Probab. 14, 86–118 (2009)
Janson, S.: The probability that a random multigraph is simple. Comb. Probab. Comput. 18(1–2), 205–225 (2009)
Janson, S., Knuth, D.E., Łuczak, T., Pittel, B.: The birth of the giant component. Random Struct. Algorithms 4(3), 233–358 (1993)
Janson, S., Luczak, M.J.: Susceptibility in subcritical random graphs. J. Math. Phys. 49(12), 125207 (2008)
Janson, S., Luczak, T., Rucinski, A.: Randomgraphs, vol. 45. Wiley, New York (2000)
Joseph, A.: The component sizes of a critical random graph with given degree sequence. Ann. Appl. Probab. 24(6), 2560–2594 (2014)
Kesten, H., Lee, S.: The central limit theorem for weighted minimal spanning trees on random points. Ann. Appl. Probab. 6(2), 495–527 (1996)
Le Gall, J.-F.: Random trees and applications. Probab. Surv. 2, 245–311 (2005)
Lee, S.: The central limit theorem for Euclidean minimal spanning trees. I. Ann. Appl. Probab. 7(4), 996–1020 (1997)
Lee, S.: The central limit theorem for Euclidean minimal spanning trees II. Adv. Appl. Probab. 31(4), 969–984 (1999)
Łuczak, T.: Sparse random graphs with a given degree sequence. In: Proceedings of the Symposium on Random Graphs, Poznan, pp. 165–182 (1989)
Łuczak, T.: Component behavior near the critical point of the random graph process. Random Struct. Algorithms 1(3), 287–310 (1990)
Łuczak, T.: The number of trees with large diameter. J. Austr. Math. Soc. 58(3), 298–311 (1995)
Lyons, R., Peres, Y., Schramm, O.: Minimal spanning forests. Ann. Probab. 34(5), 1665–1692 (2006)
Malik, H.J.: Exact distribution of the product of independent generalized gamma variables with the same shape parameter. Ann. Math. Stat. 39(5), 1751–1752 (1968)
Marckert, J.-F., Mokkadem, A.: The depth first processes of Galton–Watson trees converge to the same Brownian excursion. Ann. Probab. 31, 1655–1678 (2003)
McDiarmid, C.: Concentration. Probabilistic methods for algorithmic discrete mathematics, pp. 195–248 (1998)
Miermont, G.: Tessellations of random maps of arbitrary genus. Ann. Sci. Éc. Norm. Supér. (4) 42(5), 725–781 (2009)
Miermont, G., Sen, S.: On breadth-first constructions of scaling limits of random graphs and random unicellular maps. arXiv preprint arXiv:1908.04403 (2019)
Molloy, M., Reed, B.: A critical point for random graphs with a given degree sequence. Random Struct. Algorithms 6(2–3), 161–180 (1995)
Nachmias, A., Peres, Y.: Critical random graphs: diameter and mixing time. Ann. Probab. 36(4), 1267–1286 (2008)
Nachmias, A., Peres, Y.: Critical percolation on random regular graphs. Random Struct. Algorithms 36, 111–148 (2010)
Newman, C., Tassion, V., Wu, W.: Critical percolation and the minimal spanning tree in slabs. Commun. Pure Appl. Math. 70(11), 2084–2120 (2017)
Penrose, M.D.: The random minimal spanning tree in high dimensions. Ann. Probab. 24(4), 1903–1925 (1996)
Penrose, M.D.: The longest edge of the random minimal spanning tree. Ann. Appl. Probab. 7(2), 340–361 (1997)
Penrose, M.D.: Random minimal spanning tree and percolation on the n-cube. Random Struct. Algorithms 12(1), 63–82 (1998)
Penrose, M.D., Wade, A.R.: Random minimal directed spanning trees and Dickman-type distributions. Adv. Appl. Probab. 36(3), 691–714 (2004)
Riordan, O.: The phase transition in the configuration model. Comb. Probab. Comput. 21(1–2), 265–299 (2012)
Steele, J.M.: Growth rates of Euclidean minimal spanning trees with power weighted edges. Ann. Probab. 16(4), 1767–1787 (1988)
van der Hofstad, R., Nachmias, A.: Hypercube percolation. Preprint. J. Eur. Math. Soc. (2012)
van der Hofstad, R.: Random Graphs and Complex Networks, vol. I. Cambridge University Press, Cambridge (2017)
van der Hofstad, R., Sapozhnikov, A.: Cycle structure of percolation on high-dimensional tori. Ann. Inst. Henri Poincaré Probab. Stat. 50(3), 999–1027 (2014)
Villani, C.: Optimal transport, Grundlehren der Mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences], vol. 338, Springer, Berlin, Old and new (2009)
Acknowledgements
The authors thank two anonymous referees whose careful reading and detailed comments led to significant improvements in the paper. LAB was supported in part by an NSERC Discovery Grant and Discovery Accelerator Supplement, and by an FRQNT Team Grant, during the preparation of this research. SS was partially supported by the Infosys Foundation, Bangalore and by MATRICS grant MTR/2019/000745 from SERB.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix A
Appendix A
Our aim in this section is to briefly describe the ideas needed to prove Theorems 3.11 and 3.13 .
1.1 Sketch of proof of Theorem 3.11
Suppose \({\mathbf {t}}\) is a rooted tree with vertices labeled by [m] and let \(R(\cdot ,{\mathbf {t}})\) be as in (6.59). For \(s\ge 1\), let
Note that \(s!\times |A_s({\mathbf {t}})|\le |A_1({\mathbf {t}})|^s\). Let \(T_m\) denote a uniform rooted labeled tree on [m], and let \({\overline{T}}_m\) be distributed as
Then we have the following decomposition of \({\mathscr {H}}_{m, s}\):
Theorem A.1
Fix \(s\ge 1\). Sample \({\overline{T}}_m\), and conditional on the realization, sample \(\big ({\overline{v}}_{1,m}, {\overline{u}}_{1,m},\ldots , {\overline{v}}_{s,m},{\overline{u}}_{s,m}\big )\) from \(A_s({\overline{T}}_m)\) uniformly. Place an edge between \({\overline{v}}_{i,m}\) and \({\overline{u}}_{i,m}\) for \(1\le i\le s\), and then forget about the root of \({\overline{T}}_m\). Call the resulting graph \(\overline{{\mathscr {H}}}_{m,s}\). Then \(\overline{{\mathscr {H}}}_{m,s}{\mathop {=}\limits ^{\text {d}}}{\mathscr {H}}_{m, s}\).
This can be seen as follows: Consider a simple, connected, rooted graph G on [m] with \(\text {sp}(G)=s\). Let \({\mathbf {t}}\) be the tree constructed by following a depth-first exploration of G starting at its root, and let \(v_i, u_i\), \(1\le i\le s\), be the endpoints of the s edges that need to be added to \({\mathbf {t}}\) to recover G. We can arrange \(v_1, u_1,\ldots , v_s, u_s\) in a unique way so that the resulting sequence becomes an element of \(A_s({\mathbf {t}})\). It thus follows that the set of simple, connected, rooted graphs on [m] having s surplus edges is in bijective correspondence with the set
Then one can show that if we root \({\mathscr {H}}_{m, s}\) at a uniform vertex, then its corresponding element in the set (A.1) will be distributed as \(\big ({\overline{T}}_m, {\overline{v}}_{1,m},{\overline{u}}_{1,m},\ldots ,{\overline{v}}_{s,m},{\overline{u}}_{s,m}\big )\). We omit the details as similar ideas have already been used in [3, 26, 27].
For any tree \({\mathbf {t}}\) on [m] rooted at \(\rho \), endow the children of each vertex in \({\mathbf {t}}\) with the linear order induced by their labels. Let \(\rho =w_0, w_1,\ldots , w_{m-1}\) be the vertices of \({\mathbf {t}}\) in order of appearance in a depth-first exploration of \({\mathbf {t}}\) using the above order. Let \({{\,\mathrm{Ht}\,}}_{{\mathbf {t}}}:[0,m]\rightarrow {\mathbb {R}}\) be the height function of \({\mathbf {t}}\) given by \({{\,\mathrm{Ht}\,}}_{{\mathbf {t}}}(m)=0\), and
The following lemma is a collection of some standard results about \(T_m\):
Lemma A.2
- (i) :
-
The following convergences hold:
$$\begin{aligned}&m^{-1/2}{{\,\mathrm{Ht}\,}}_{T_m}\big (m\cdot \big ){\mathop {\longrightarrow }\limits ^{\text {d}}}2{\mathbf {e}}(\cdot ), \ \ \ \ \text { and} \nonumber \\&m^{-1/2}\max _{v\in [m]}\big |2\big |R(v, T_m)\big |-{{\,\mathrm{ht}\,}}(v, T_m)\big |{\mathop {\longrightarrow }\limits ^{\text {P}}}0, \end{aligned}$$(A.2)where the convergence in (A.2) is w.r.t. the Skorohod \(J_1\) topology.
- (ii) :
-
For all \(m\ge 1\), \({{\,\mathrm{{\mathbb {P}}}\,}}\big ({{\,\mathrm{ht}\,}}(T_m)\ge x\sqrt{m}\big )\le c x^3\exp \big (-x^2/2\big )\).
- (iii) :
-
For all \(x\ge 0\) and \(m\ge 1\),
$$\begin{aligned} {{\,\mathrm{{\mathbb {P}}}\,}}\big (\max _{v\in [m]}|R(v, T_m)|\ge x\sqrt{m}\big )\le c\exp (-c'x^2). \end{aligned}$$Using the bounds \(|A_s(T_m)|\times s!\le |A_1(T_m)|^s\) and \(|A_1(T_m)|\le m\cdot \max _{v\in [m]}|R(v, T_m)|\), we further have
$$\begin{aligned} {{\,\mathrm{{\mathbb {P}}}\,}}\big (|A_s(T_m)|\ge x m^{3s/2}\big )\le c\exp \big (-c'x^{2/s}\big ) \end{aligned}$$for any \(s\ge 1\), \(x\ge 0\), and \(m\ge 1\).
- (iv) :
-
For any \(s\ge 1\), \(m^{-3s/2}\big (|A_1(T_m)|^s-|A_s(T_m)|\times s!\big ){\mathop {\longrightarrow }\limits ^{\text {P}}}0\).
Lemma A.2((i)) follows from [69]. ((ii)) follows from [66, Corollary 1]. ((iii)) is the content of [3, Lemma 13]. The proof of ((iv)) is similar to that of [26, Lemma 7.3 (iii)].
Sketch of proof of (3.9)
In view of Theorem A.1, \(s\cdot \big (\text {ht}({\overline{T}}_m)+1\big )\) dominates \(L({\mathscr {H}}_{m, s})\) stochastically for any \(s\ge 1\). Thus, (3.9) follows from Lemma A.2 ((ii)) and ((iii)). \(\square \)
To prove the other assertions in Theorem 3.11 it will be convenient to work with two slightly different spaces \({\mathscr {H}}_{m, s}^{\circ }\) and \({\mathscr {H}}_{m, s}^{\dagger }\) which we define next. Recall the notation \(R(\cdot ,\cdot ,\cdot )\) from (6.58). Sample \(T_m^{\circ }\) according to distribution
Conditional on \(T_m^{\circ }\), sample an i.i.d. sequence of triples \((v_{i,m}^{\circ }, u_{i,m}^{\circ }, f_{i,m}^{\circ })\), \(1\le i\le s\), where
Let \({\mathscr {H}}_{m, s}^{\dagger }\) (resp. \({\mathscr {H}}_{m, s}^{\circ }\)) be the space obtained by adding an edge between \(v_{i,m}^{\circ }\) and \(f_{i,m}^{\circ }\) (resp. between \(v_{i,m}^{\circ }\) and \(u_{i,m}^{\circ }\)) for \(1\le i\le s\), and then forgetting about the root of \(T_m^{\circ }\). It follows from Lemma A.2 ((iii)) and ((iv)) that the total variation distance between the laws of \({\overline{{\mathscr {H}}}}_{m, s}\) (as defined in Theorem A.1) and \({\mathscr {H}}_{m, s}^{\dagger }\) tends to zero as \(m\rightarrow \infty \). It thus follows from Theorem A.1 that there exists a coupling of \({\mathscr {H}}_{m, s}\) and \({\mathscr {H}}_{m, s}^\dagger \) such that
We will now recall an alternate construction of \({\mathscr {H}}^{(s)}\) which is essentially given in [3]; see also the discussion below [2, Equation (1)]. We first introduce some notation. For any \(f:[0,1]\rightarrow {\mathbb {R}}, x\in [0,1]\), and \(h>0\), let
where \(\sup \{\ \}=-\infty \) and \(\inf \{\ \}=\infty \) by convention.
Construction A.3
(Alternate construction of \({\mathscr {H}}^{(s)}\)) Fix an integer \(s\ge 2\).
-
(a)
Sample \({\mathbf {e}}^{\circ }\) with law given by
$$\begin{aligned} {\mathbb {E}}\big [f({\mathbf {e}}^{\circ })\big ]=\frac{{\mathbb {E}}\big [f({\mathbf {e}})\big (\int _0^1 {\mathbf {e}}(t)dt\big )^s\big ]}{{\mathbb {E}}\big [\big (\int _0^1 {\mathbf {e}}(t)dt\big )^s\big ]}. \end{aligned}$$ -
(b)
Conditional on \({\mathbf {e}}^{\circ }\), sample i.i.d. points \(y_1^{\circ },\ldots , y_s^{\circ }\) having density \({\mathbf {e}}^{\circ }(y)\big /\int _0^1 {\mathbf {e}}^{\circ }(t)dt \).
-
(c)
Conditional on the above, sample \(h_1^{\circ },\ldots ,h_s^{\circ }\) independently, where \(h_i^{\circ }\sim \text {Unif}[0,{\mathbf {e}}^{\circ }(y_i^\circ )]\). Set \(x_i^\circ ={{\,\mathrm{prev}\,}}(y_i^\circ , h_i^\circ ; {\mathbf {e}}^\circ )\).
-
(d)
Form the quotient space \({\mathscr {T}}_{{\mathbf {e}}^\circ }/\sim \), where \(\sim \) is the equivalence relation under which \(q_{{\mathbf {e}}^\circ }(x_i^\circ )\sim q_{{\mathbf {e}}^\circ }(y_i^\circ )\), \(1\le i\le s\).
Then \({\mathscr {H}}^{(s)}{\mathop {=}\limits ^{\text {d}}}2\cdot \big ({\mathscr {T}}_{{\mathbf {e}}^\circ }/\sim \big )\).
Now observe that \({\mathscr {H}}_{m, s}^{\circ }\) has a similar alternate construction: First sample \(T_m^\circ \) as in (A.3). Let \(w_0,\ldots ,w_{m-1}\) be the vertices of \(T_m^\circ \) in order of appearance in a depth-first exploration of \(T_m^\circ \). Let \({{\,\mathrm{Ht}\,}}^\circ \) be the height function of \(T_m^\circ \). Conditional on \(T_m^\circ \), sample i.i.d. random variables \(y_{1,m}^\circ ,\ldots ,y_{s,m}^\circ \), where
Conditional on the above, sample \(h_{1,m}^\circ ,\ldots , h_{s,m}^\circ \) independently via
where \(v=w_{y_{i,m}^\circ }\). Let \(x_{i,m}^\circ ={{\,\mathrm{prev}\,}}(y_{i,m}^\circ , h_{i,m}^{\circ }; {{\,\mathrm{Ht}\,}}^\circ )-1\). Then \({\mathscr {H}}_{m, s}^\circ \) has the same distribution as the space obtained by placing an edge in \(T_m^\circ \) between \(w_{y_{i,m}^\circ }\) and \(w_{x_{i,m}^\circ }\) for \(1\le i\le s\).
Sketch of proof of (3.8)
Using Lemma A.2 ((i)) and ((iii)), it can be shown that the following convergences hold jointly:
as \(m\rightarrow \infty \). Using Construction A.3 and the above alternate construction of \({\mathscr {H}}_{m, s}^\circ \), it is now routine to prove the assertion in (3.8) for \({\mathscr {H}}_{m, s}^\circ \), from which it follows that the same is true for \({\mathscr {H}}_{m, s}^\dagger \). The desired result now follows from (A.4). \(\square \)
Let \(y_{(i),m}^\circ \) (resp. \(y_{(i)}^\circ \)), \(1\le i\le s\), be \(y_{i,m}^\circ \) (resp. \(y_i^\circ \)), \(1\le i\le s\), arranged in an increasing order. For \(1\le i\le s-1\) define \(z_{i,m}^\circ \) and \(z_i^\circ \) via
Further, define
Sketch of proof of (3.11)
From (A.5) it follows that the following convergence holds jointly with the convergence in (A.5): As \(m\rightarrow \infty \),
Arrange \(x_i^\circ , x_i^{\circ , +}, y_i^\circ \), \(1\le i\le s\), and \(z_i^\circ , z_i^{\circ , -}, z_i^{\circ , +}\), \(1\le i\le s-1\), (resp. \(x_{i,m}^\circ , x_{i,m}^{\circ , +}, y_{i,m}^\circ \), \(1\le i\le s\), and \(z_{i,m}^\circ , z_{i,m}^{\circ , -}, z_{i,m}^{\circ , +}\), \(1\le i\le s-1\)) in increasing order as \(a_1,\ldots ,a_{6s-3}\) (resp. as \(a_{1,m},\ldots ,a_{6s-3, m}\)). Let
Then it follows from (A.6) and the second convergence in (A.5) that
Recall the notation used in (3.11), and note that there exists a partition \({\mathscr {P}}=\{{\mathscr {P}}_1,\ldots ,{\mathscr {P}}_r\}\) of \([6s-4]\) that depends only on the realizations of \({\mathbf {e}}^\circ \) and \(x_i^\circ ,y_i^\circ \), \(1\le i\le s\), such that
Further, it follows from (A.5) that for large m, the vector consisting of the numbers of vertices in \({\mathscr {H}}_{m, s}^\circ \) that are connected to the different elements of \(e({\mathscr {H}}_{m, s}^\circ )\) is given by \(\big (\sum _{j\in {\mathscr {P}}_i}\Delta _{j,m},\ 1\le i\le r\big )\), where the common endpoints of multiple \(e\in e({\mathscr {H}}_{m, s}^\circ )\) and the vertices in their pendant subtrees have been accounted for in \(\sum _{j\in {\mathscr {P}}_i}\Delta _{j,m}\) for exactly one value of i in a specific way. Using (A.7) and (A.8), we get the analogue of (3.11) for \({\mathscr {H}}_{m, s}^\circ \) for the above specific way of assigning the common endpoints of multiple \(e\in e({\mathscr {H}}_{m, s}^\circ )\) and the vertices in their pendant subtrees to the different terms \(\sum _{j\in {\mathscr {P}}_i}\Delta _{j,m}\).
This together with (A.4) would complete the proof if we can show that the sizes of the pendant subtrees of the common endpoints of multiple \(e\in e({\mathscr {H}}_{m, s}^\circ )\) are asymptotically negligible. This negligibility claim follows from the following facts:
-
(A)
\(Y_{i,m}=o_P(m)\), \(1\le i\le s\), where \(Y_{i,m}\) denotes the number of descendants of \(v_{i,m}^\circ \) in \(T_m^\circ \).
-
(B)
\(X_{i,m}=o_P(m)\), \(1\le i\le s\), where \(X_{i,m}\) denotes the number of descendants of \(u_{i,m}^\circ \) in \(T_m^\circ \) that are not in the subtree that contains \(v_{i,m}^\circ \).
-
(C)
For every \(\varepsilon >0\),
$$\begin{aligned} {{\,\mathrm{{\mathbb {P}}}\,}}\big (\exists v\in T_m^\circ : v\text { has at least three subtrees in }T_m^\circ \text { each of size }\ge \varepsilon m\big ) \rightarrow 0\ \ \text { as }\ \ m\rightarrow \infty . \end{aligned}$$
(A) and (C) follow from (A.5) and the facts that \(q_{{\mathbf {e}}^\circ }(y_i^\circ )\) is almost surely a leaf in \({\mathscr {T}}_{{\mathbf {e}}^\circ }\) and that \({\mathscr {T}}_{{\mathbf {e}}^\circ }\) is almost surely binary. The proof of (B) is also routine. \(\square \)
1.2 Sketch of proof of Theorem 3.13
Assume that for each \(m\ge 1\), \({\varvec{k}}^{\scriptscriptstyle (m)}=(k_i^{\scriptscriptstyle (m)}, i\ge 0)\), where \(k_i^{\scriptscriptstyle (m)}\) are nonnegative integers satisfying \(\sum _{i\ge 0}k_i^{\scriptscriptstyle (m)}=m\) and \(\sum _{i\ge 0}ik_i^{\scriptscriptstyle (m)}=m-1\). Then there exist trees on m vertices in which for each \(i\ge 0\), there are exactly \(k_i^{\scriptscriptstyle (m)}\) many vertices with i many children. We call \({\varvec{k}}^{\scriptscriptstyle (m)}\) the child sequence of such a tree. Assumption 3.4 gives the criterion for graphs with given degree sequences to be critical. The following assumption gives the analogous criterion for plane trees with given child sequences.
Assumption A.4
There exists a pmf \((p_0, p_1,\ldots )\) with
such that
We will write \(\sigma ^2 = \sum _i i^2 p_i - 1\) for the variance associated with the pmf \((p_0, p_1,\ldots )\).
Let \({\mathbb {T}}_{{\varvec{k}}^{\scriptscriptstyle (m)}}\) be the set of plane trees with child sequence \({\varvec{k}}^{\scriptscriptstyle (m)}\). Let \({\mathscr {T}}_{{\varvec{k}}^{\scriptscriptstyle (m)}}\) be a uniform element of \({\mathbb {T}}_{{\varvec{k}}^{\scriptscriptstyle (m)}}\) endowed with the tree distance and the uniform probability measure on m vertices and viewed as a metric measure space. Broutin and Marckert [34] showed that under Assumption A.4, \(\sigma m^{-1/2}{\mathscr {T}}_{{\varvec{k}}^{\scriptscriptstyle (m)}}{\mathop {\longrightarrow }\limits ^{\text {d}}}{\mathscr {T}}_{2{\mathbf {e}}}\) w.r.t. GHP topology. The following variant of this result follows from [26, Lemma 7.4 and Lemma 7.6]:
Lemma A.5
Suppose \({\varvec{k}}^{\scriptscriptstyle (m)}\) satisfies Assumption A.4. Further, suppose \(f_m:\{0, 1,\ldots \}\rightarrow [0,1]\) is such that
Let \({\mathscr {T}}_{{\varvec{k}}^{\scriptscriptstyle (m)}}^{f_m}\) be a uniform element of \({\mathbb {T}}_{{\varvec{k}}^{\scriptscriptstyle (m)}}\) endowed with the tree distance and the measure that assigns probability \(f_m(i)\) to any node that has i children, \(i\ge 0\). Then
Now we can prove Theorem 3.13 using the above lemma and the techniques used in the proof of [26, Theorem 2.2].
Rights and permissions
About this article
Cite this article
Addario-Berry, L., Sen, S. Geometry of the minimal spanning tree of a random 3-regular graph. Probab. Theory Relat. Fields 180, 553–620 (2021). https://doi.org/10.1007/s00440-021-01071-3
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00440-021-01071-3
Keywords
- Minimal spanning tree
- Gromov–Hausdorff distance
- Critical percolation
- Real tree
- Random regular graphs
- Graphs with prescribed degree sequence
- Configuration model
Mathematics Subject Classification
- Primary 60C05
- 05C80