, In this subsection we prove that, at the difference from other hyperbolicity parameters
, Note that if we are given a BFS-tree T rooted at w, we can easily check whether ? w,T (G) ? 1, and thus deciding whether ? ? (G) ? 1 is in NP
, Up to preprocessing the formula, we can suppose that ? satisfies the following properties (otherwise, ? can be reduced to a formula satisfying these conditions): ? no clause c j can be reduced to a singleton, vol.1
, ? every literal x i , x i is contained in at least one clause
, ? no clause c j can be strictly contained in another clause c k
, ? every clause c j is disjoint from some other clause c k (otherwise, a trivial satisfiability assignment for ? is to set true every literal in c j )
, ? if two clauses c j , c k are disjoint, then there exists another clause c p that intersects c j in exactly one literal, and similarly, that also intersects c k in exactly one literal (otherwise, we add the two new clauses x ? y and x ? y, with x, y being fresh new variables; then, we replace every clause c j by the two new clauses c j ? x ? y and c j ? x ? y)
For simplicity, in what follows, we often denote x i , x i by 2i?1 , 2i . Let C := {c 1 , . . . , c m } be the clause-set of ?. Finally, let w and V = {v 1 , v 2 , . . . , v 2n } be additional vertices. We construct a graph G ? with V (G ? ) = {w} ? V ? X ? C and where E(G ? ) is defined as follows: ? N (w) = V and V is a clique, ? for every i, i, = i ,
, ? for every i, i , i and i are adjacent if and only if i = i
, ? for every i, j, v i and c j are not adjacent
, ? for every i, j, i and c j are adjacent if and only if i ? c j
, By construction, d(c j , c k ) = 2, and the parent node of c k contained in c j ? c p . We have (c k | i ) c j = 2 ? d(c k , i )/2 ? {1, 3/2}. In particular, (c k | i ) c j = 1. As a result
, For every BFS-tree T rooted at i ? X, we have ? i
, Since there is a perfect matching between X and V = N (w), the parent node of w must be v i . We claim that the parent node of v i , for every i = i, must be also v i . Indeed, otherwise this should be i . However, (w|v i ) i = (2 + 2 ? 1)/2 = 3/2. In particular, (w|v i ) i = 1, and so, ? i ,T (G ? ) ? 1 implies v i and i should be adjacent, that is a contradiction. So, the claim is proved. Then, let c j ? C be nonadjacent to i . We have d(c j , i ) = 2, and the parent node p j of c j, vol.1
, For every BFS-tree T rooted at v i ? V , we have ? v i
, In particular, the parent of i must be v i . Furthermore, there exists c j ? C such that d(v i , c j ) = 2. In particular, i must be the parent of c j . However, ( i |c j ) v i = 2 ? d( i , c j )/2 ? {1, 3/2}. In particular, ( i |c j ) v i = 1. So, ? v i ,T (G ? ) ? d(v i , i ) = 2. From now on, There exists i = i such that i , i are nonadjacent
, If s ? V and t is arbitrary
, Since w is a simplicial vertex and s ? N (w), we have d(w, t) ? 1 ? d(s, t) ? d(w, t), and consequently
, We have (s|t) w = 2 ? d(s, t)/2. In particular, (s|t) w ? 1. As a result
, 2 ? {3/2, 2}. In particular, if d(s, t) = 2 then (s|t) w = 1, and so, we are done because s t , t s ? N [w] and w is simplicial. Otherwise, d(s, t) = 1, and so, (s|t) w = 2. In particular, s t = s and s t
, We have (s|t) w = 3 ? d(s, t)/2 ? {2, 5/2}. In particular
Metric tree-like structures in real-world networks: an empirical study, Networks, vol.67, pp.49-68, 2016. ,
Tree-like structure in large social and information networks, ICDM 2013, pp.1-10, 2013. ,
Notes on word hyperbolic groups, Group Theory from a Geometrical Viewpoint, pp.3-63, 1990. ,
On computing the hyperbolicity of real-world graphs, ESA 2015, vol.9294, pp.215-226, 2015. ,
URL : https://hal.archives-ouvertes.fr/hal-01199860
Into the square: on the complexity of some quadratic-time solvable problems, Electron. Notes Theor. Comput. Sci, vol.322, pp.51-67, 2015. ,
URL : https://hal.archives-ouvertes.fr/hal-01390131
Notes on Gromov's hyperbolicity criterion for path-metric spaces, Group Theory from a Geometrical Viewpoint, pp.64-167, 1990. ,
Metric Spaces of Non-Positive Curvature, Grundlehren Math. Wiss, vol.319, 1999. ,
Fast approximation and exact computation of negative curvature parameters of graphs, :15. Schloss Dagstuhl -Leibniz-Zentrum fuer Informatik, vol.99, pp.1-22, 2018. ,
URL : https://hal.archives-ouvertes.fr/hal-01836063
Cop and robber game and hyperbolicity, SIAM J. Discrete Math, vol.28, 1987. ,
DOI : 10.1137/130941328
URL : https://hal.archives-ouvertes.fr/hal-01198887
Diameters, centers, and approximating trees of delta-hyperbolic geodesic spaces and graphs, SoCG, pp.59-68, 2008. ,
Additive spanners and distance and routing labeling schemes for hyperbolic graphs, Algorithmica, vol.62, pp.713-732, 2012. ,
DOI : 10.1007/s00453-010-9478-x
URL : https://hal.archives-ouvertes.fr/hal-01194836
Core congestion is inherent in hyperbolic networks, SODA 2017, pp.2264-2279, 2017. ,
URL : https://hal.archives-ouvertes.fr/hal-01769624
Packing and covering ?-hyperbolic spaces by balls, APPROX-RANDOM 2007, vol.4627, pp.59-73, 2007. ,
DOI : 10.1007/978-3-540-74208-1_5
On computing the Gromov hyperbolicity, ACM J. Exp. Algorithmics, vol.20, p.18, 2015. ,
URL : https://hal.archives-ouvertes.fr/hal-01182890
Recognition of C4-free and 1/2-hyperbolic graphs, SIAM J. Discrete Math, vol.28, pp.1601-1617, 2014. ,
URL : https://hal.archives-ouvertes.fr/hal-01070768
Fully polynomial FPT algorithms for some classes of bounded cliquewidth graphs, SODA 2018, pp.2765-2784, 2018. ,
URL : https://hal.archives-ouvertes.fr/hal-01676187
Effect of Gromov-hyperbolicity parameter on cuts and expansions in graphs and some algorithmic implications, Algorithmica, vol.80, pp.772-800, 2018. ,
Courbure mésoscopique et théorie de la toute petite simplification, J. Topol, vol.1, pp.804-836, 2008. ,
DOI : 10.1112/jtopol/jtn023
Approximation algorithms for the Gromov hyperbolicity of discrete metric spaces, LATIN 2014, vol.8392, pp.285-293, 2014. ,
Fast approximation algorithms for p-centers in large ?-hyperbolic graphs, Algorithmica, vol.80, pp.3889-3907, 2018. ,
When can graph hyperbolicity be computed in linear time?, WADS 2017, vol.10389, pp.397-408, 2017. ,
Computing the Gromov hyperbolicity of a discrete metric space, Inform Process. Lett, vol.115, pp.576-579, 2015. ,
, Les groupes hyperboliques d'après M. Gromov, vol.83, 1990.
Hyperbolic groups, Group Theory, vol.8, pp.75-263, 1987. ,
Weak hyperbolicity of cube complexes and quasi-arboreal groups, J. Topology, vol.7, issue.2, pp.385-418, 2014. ,
On the hyperbolicity of large-scale networks and its estimation, pp.3344-3351, 2016. ,
Large-scale curvature of networks, Phys. Rev. E, vol.84, p.66108, 2011. ,
Strongly geodesically automatic groups are hyperbolic, Invent. Math, vol.121, issue.2, pp.323-334, 1995. ,
DOI : 10.1007/bf01884301
An algorithm detecting hyperbolicity, Geometric and computational perspectives on infinite groups, vol.25, pp.193-200, 1994. ,
DOI : 10.1090/dimacs/025/10
On infinite bridged graphs and strongly dismantlable graphs, Discrete Math, vol.211, pp.153-166, 2000. ,
DOI : 10.1016/s0012-365x(99)00142-9
URL : https://doi.org/10.1016/s0012-365x(99)00142-9
Hyperbolic embedding of Internet graph for distance estimation and overlay construction, IEEE/ACM Trans. Netw, vol.16, issue.1, pp.25-36, 2008. ,
, Quelques propriétés topologiques des graphes et applicationsà Internet et aux réseaux, 2011.
Metric embedding, hyperbolic space, and social networks, pp.501-510, 2014. ,
DOI : 10.1016/j.comgeo.2016.08.003
URL : https://manuscript.elsevier.com/S0925772116300712/pdf/S0925772116300712.pdf
An improved combinatorial algorithm for boolean matrix multiplication, ICALP 2015, vol.9134, pp.1094-1105, 2015. ,