Ponències/Comunicacions de congressos
http://hdl.handle.net/2117/3549
2017-06-25T12:28:38ZDynamic programming for graphs on surfaces
http://hdl.handle.net/2117/104306
Dynamic programming for graphs on surfaces
Rué Perna, Juan José; Sau, Ignasi; Thilikos, Dimitrios
We provide a framework for the design and analysis of dynamic
programming algorithms for surface-embedded graphs on n vertices
and branchwidth at most k. Our technique applies to general families
of problems where standard dynamic programming runs in 2O(k·log k).
Our approach combines tools from topological graph theory and
analytic combinatorics.
2017-05-11T09:30:07ZRué Perna, Juan JoséSau, IgnasiThilikos, DimitriosWe provide a framework for the design and analysis of dynamic
programming algorithms for surface-embedded graphs on n vertices
and branchwidth at most k. Our technique applies to general families
of problems where standard dynamic programming runs in 2O(k·log k).
Our approach combines tools from topological graph theory and
analytic combinatorics.Random cubic planar graphs revisited
http://hdl.handle.net/2117/103660
Random cubic planar graphs revisited
Rué Perna, Juan José; Noy Serrano, Marcos; Requile, Clement
The goal of our work is to analyze random cubic planar graphs according to the
uniform distribution. More precisely, let G be the class of labelled cubic planar
graphs and let gn be the number of graphs with n vertices. Then each graph in G
with n vertices has the same probability 1/gn.
2017-04-24T10:12:03ZRué Perna, Juan JoséNoy Serrano, MarcosRequile, ClementThe goal of our work is to analyze random cubic planar graphs according to the
uniform distribution. More precisely, let G be the class of labelled cubic planar
graphs and let gn be the number of graphs with n vertices. Then each graph in G
with n vertices has the same probability 1/gn.Exploiting symmetry on the Universal Polytope
http://hdl.handle.net/2117/17663
Exploiting symmetry on the Universal Polytope
Pfeifle, Julián
The most successful method to date for finding lower bounds on the
number of simplices needed to triangulate a given polytope P involves optimizing
a linear functional over the associated Universal Polytope U(P). However, as the
dimension of P grows, these linear programs become increasingly difficult to formulate
and solve.
Here we present a method to algorithmically construct the quotient of U(P) by
the symmetry group Aut(P) of P, which leads to dramatic reductions in the size of
the linear program. We compare the power of our approach with older computations
by Orden and Santos, indicate the influence of the combinatorial complexity barrier
on these computations, and sketch some future applications.
2013-02-12T13:49:09ZPfeifle, JuliánThe most successful method to date for finding lower bounds on the
number of simplices needed to triangulate a given polytope P involves optimizing
a linear functional over the associated Universal Polytope U(P). However, as the
dimension of P grows, these linear programs become increasingly difficult to formulate
and solve.
Here we present a method to algorithmically construct the quotient of U(P) by
the symmetry group Aut(P) of P, which leads to dramatic reductions in the size of
the linear program. We compare the power of our approach with older computations
by Orden and Santos, indicate the influence of the combinatorial complexity barrier
on these computations, and sketch some future applications.On two distributions of subgroups of free groups
http://hdl.handle.net/2117/15220
On two distributions of subgroups of free groups
Bassino, Frédérique; Martino, Armando; Nicaud, Cyril; Ventura Capell, Enric; Weil, Pascal
We study and compare two natural distributions of
finitely generated subgroups of free groups. One is
based on the random generation of tuples of reduced
words; that is the one classically used by group theorists.
The other relies on Stallings’ graphical representation
of subgroups and in spite of its naturality, it was
only recently considered. The combinatorial structures
underlying both distributions are studied in this paper
with methods of analytic combinatorics. We use these
methods to point out the differences between these
distributions. It is particularly interesting that certain
important properties of subgroups that are generic in
one distribution, turn out to be negligible in the other.
2012-02-17T14:39:00ZBassino, FrédériqueMartino, ArmandoNicaud, CyrilVentura Capell, EnricWeil, PascalWe study and compare two natural distributions of
finitely generated subgroups of free groups. One is
based on the random generation of tuples of reduced
words; that is the one classically used by group theorists.
The other relies on Stallings’ graphical representation
of subgroups and in spite of its naturality, it was
only recently considered. The combinatorial structures
underlying both distributions are studied in this paper
with methods of analytic combinatorics. We use these
methods to point out the differences between these
distributions. It is particularly interesting that certain
important properties of subgroups that are generic in
one distribution, turn out to be negligible in the other.On the evaluation of the Tutte polynomial at the points (1,-1) and (2,-1)
http://hdl.handle.net/2117/11523
On the evaluation of the Tutte polynomial at the points (1,-1) and (2,-1)
Goodall, Andrew; Merino, Criel; Mier Vinué, Anna de; Noy Serrano, Marcos
C. Merino proved recently the following identity between evaluations of the Tutte polynomial of complete graphs: t($K_{n+2}$; 1,−1) = t($K_n$;2,−1). In this work we extend this result by giving a large class of graphs with this property, that is, graphs G such that there exist two vertices u and v with t(G;1,−1) = t(G−{u,v};2,−1). The class is described in terms of forbidden induced subgraphs and it contains in particular threshold graphs.
2011-02-24T12:39:38ZGoodall, AndrewMerino, CrielMier Vinué, Anna deNoy Serrano, MarcosC. Merino proved recently the following identity between evaluations of the Tutte polynomial of complete graphs: t($K_{n+2}$; 1,−1) = t($K_n$;2,−1). In this work we extend this result by giving a large class of graphs with this property, that is, graphs G such that there exist two vertices u and v with t(G;1,−1) = t(G−{u,v};2,−1). The class is described in terms of forbidden induced subgraphs and it contains in particular threshold graphs.Rotational and dihedral symmetries in Steinhaus and Pascal binary triangles
http://hdl.handle.net/2117/8718
Rotational and dihedral symmetries in Steinhaus and Pascal binary triangles
Brunat Blay, Josep Maria; Maureso Sánchez, Montserrat
We give explicit formulae for obtaining the binary sequences which produce Steinhaus triangles and generalized Pascal triangles with rotational and dihedral symmetries.
2010-09-02T10:25:11ZBrunat Blay, Josep MariaMaureso Sánchez, MontserratWe give explicit formulae for obtaining the binary sequences which produce Steinhaus triangles and generalized Pascal triangles with rotational and dihedral symmetries.An algorithm to design prescribed length codes for single-tracked shaft encoders
http://hdl.handle.net/2117/8428
An algorithm to design prescribed length codes for single-tracked shaft encoders
Balle Pigem, Borja de; Ventura Capell, Enric; Fuertes Armengol, José Mª
Abstract-Maximal-length binary shift register sequences have been known for a long time. They have many interesting properties, one of them is that when taken in blocks of n consecutive positions they form 2n - 1 different codes in a closed circular sequence. This property can be used for measuring absolute angular positions as the circle can be divided in as many parts
as different codes can be retrieved. This paper describes how a closed binary sequence with arbitrary length can be effectively
designed with the minimal possible block-length, using linear feedback shift registers (LFSR). Such sequences can be used
for measuring a specified exact number of angular positions, using the minimal possible number of detectors allowed by linear methods.
2010-07-27T12:04:47ZBalle Pigem, Borja deVentura Capell, EnricFuertes Armengol, José MªAbstract-Maximal-length binary shift register sequences have been known for a long time. They have many interesting properties, one of them is that when taken in blocks of n consecutive positions they form 2n - 1 different codes in a closed circular sequence. This property can be used for measuring absolute angular positions as the circle can be divided in as many parts
as different codes can be retrieved. This paper describes how a closed binary sequence with arbitrary length can be effectively
designed with the minimal possible block-length, using linear feedback shift registers (LFSR). Such sequences can be used
for measuring a specified exact number of angular positions, using the minimal possible number of detectors allowed by linear methods.On polytopality of Cartesian products of graphs
http://hdl.handle.net/2117/8200
On polytopality of Cartesian products of graphs
Pfeifle, Julián; Pilaud, Vincent; Santos Pérez, Francisco Javier
We study the polytopality of Cartesian products of non-polytopal graphs.
On the one hand, we prove that a product of graphs is the graph of a simple polytope
if and only if its factors are. On the other hand, we provide a general construction of
polytopal products of a polytopal graph by a non-polytopal graph.
2010-07-16T09:44:24ZPfeifle, JuliánPilaud, VincentSantos Pérez, Francisco JavierWe study the polytopality of Cartesian products of non-polytopal graphs.
On the one hand, we prove that a product of graphs is the graph of a simple polytope
if and only if its factors are. On the other hand, we provide a general construction of
polytopal products of a polytopal graph by a non-polytopal graph.Graph operations and Laplacian eigenpolytopes
http://hdl.handle.net/2117/8198
Graph operations and Laplacian eigenpolytopes
Padrol Sureda, Arnau; Pfeifle, Julián
We introduce the Laplacian eigenpolytopes ("L-polytopes") associated to a simple undirected graph G, investigate how they change under basic operations such as taking the union, join, complement, line graph and cartesian product of graphs, and show how several "famous" polytopes arise as L-polytopes of "famous" graphs.
Eigenpolytopes have been previously introduced by Godsil, who studied them in
detail in the context of distance-regular graphs. Our focus on the Laplacian matrix,
as opposed to the adjacency matrix of G, permits simpler proofs and descriptions of
the result of operations on not necessarily distance-regular graphs. Additionally, it
motivates the study of new operations on polytopes, such as the Kronecker product.
Thus, we open the door to a detailed study of how combinatorial properties of G are reflected in its L-polytopes. Subsequent papers will use these tools to construct interesting polytopes from interesting graphs, and vice versa.
2010-07-16T09:29:16ZPadrol Sureda, ArnauPfeifle, JuliánWe introduce the Laplacian eigenpolytopes ("L-polytopes") associated to a simple undirected graph G, investigate how they change under basic operations such as taking the union, join, complement, line graph and cartesian product of graphs, and show how several "famous" polytopes arise as L-polytopes of "famous" graphs.
Eigenpolytopes have been previously introduced by Godsil, who studied them in
detail in the context of distance-regular graphs. Our focus on the Laplacian matrix,
as opposed to the adjacency matrix of G, permits simpler proofs and descriptions of
the result of operations on not necessarily distance-regular graphs. Additionally, it
motivates the study of new operations on polytopes, such as the Kronecker product.
Thus, we open the door to a detailed study of how combinatorial properties of G are reflected in its L-polytopes. Subsequent papers will use these tools to construct interesting polytopes from interesting graphs, and vice versa.Overlapping community search for social networks
http://hdl.handle.net/2117/7797
Overlapping community search for social networks
Padrol Sureda, Arnau; Perarnau Llobet, Guillem; Pfeifle, Julián; Muntés Mulero, Víctor
Finding decompositions of a graph into a family of clusters is crucial to understanding its underlying structure.
While most existing approaches focus on partitioning the nodes, real-world datasets suggest the presence of overlapping communities. We present OCA, a novel algorithm to detect overlapped communities in large data graphs. It outperforms previous proposals in terms of execution time, and efficiently handles large graphs containing more than 108 nodes and edges.
2010-06-22T14:08:12ZPadrol Sureda, ArnauPerarnau Llobet, GuillemPfeifle, JuliánMuntés Mulero, VíctorFinding decompositions of a graph into a family of clusters is crucial to understanding its underlying structure.
While most existing approaches focus on partitioning the nodes, real-world datasets suggest the presence of overlapping communities. We present OCA, a novel algorithm to detect overlapped communities in large data graphs. It outperforms previous proposals in terms of execution time, and efficiently handles large graphs containing more than 108 nodes and edges.