Odd Hole Research Articles

Odd Hole Recognition in Graphs of Bounded Clique Size

In a graph G, an odd hole is an induced odd cycle of length at least 5. A clique of G is a set of pairwise adjacent vertices. In this paper we consider the class ${\cal C}_k$ of graphs whose cliques have a size bounded by a constant k. Given a graph G in ${\cal C}_k$, we show how to recognize in polynomial time whether G contains an odd hole.

The Wheels of the Orthogonal Latin Squares Polytope: Classification and Valid Inequalities

A wheel in a graph G(V,E) is an induced subgraph consisting of an odd hole and an additional node connected to all nodes of the hole. In this paper, we study the wheels of the intersection graph of the Orthogonal Latin Squares polytope (P I ). Our work builds on structural properties of wheels which are used to categorise them into a number of collectively exhaustive classes. These classes give rise to families of inequalities that are valid for P I and facet-defining for its set-packing relaxation. The classification introduced allows us to establish the cardinality of the whole wheel class and determine the range of the coefficients of any variable included in a lifted wheel inequality. Finally, based on this classification, a constant-time recognition algorithm for wheel-inducing circulant matrices, is introduced.

On strongly circular-perfectness

Abstract Perfect graphs constitute a well-studied graph class with a rich structure. Circular-perfect graphs as introduced by Zhu are a natural superclass of perfect graphs defined by means of a more general coloring concept and form an important class of χ-bound graphs with the smallest non-trivial χ-binding function χ ( G ) ⩽ ω ( G ) + 1 . Apart from perfect graphs, circular-perfect graphs include all convex-round graphs and outerplanar graphs. A linear description of facets of stable set polytopes of circular-perfect graphs is still unknown, though convex-round graphs and outerplanar graphs are rank-perfect. In this paper, we exhibit an infinite class of non rank-perfect circular-perfect graphs. We introduce strongly circular-perfectness: a circular-perfect graph is said to be strongly circular-perfect if its complement is also circular-perfect. This subclass of circular-perfect graphs includes perfect graphs, odd holes and antiholes. We show that there are infinitely many non rank-perfect strongly circular-perfect graphs and fully characterize triangle-free minimal strongly circular-imperfect graphs: it turns out that these graphs are very close to odd holes.

A class of perfectly contractile graphs

We consider the class A of graphs that contain no odd hole, no antihole, and no “prism” (a graph consisting of two disjoint triangles with three disjoint paths between them). We prove that every graph G ∈ A different from a clique has an “even pair” (two vertices that are not joined by a chordless path of odd length), as conjectured by Everett and Reed [“Even pairs”, in: J.L. Ramírez-Alfonsín, B.A. Reed (eds.), Perfect Graphs, Wiley Interscience, New York, 2001]. Our proof is a polynomial-time algorithm that produces an even pair with the additional property that the contraction of this pair yields a graph in A . This entails a polynomial-time algorithm, based on successively contracting even pairs, to color optimally every graph in A . This generalizes several results concerning some classical families of perfect graphs.

Exploring the Relationship Between Max-Cut and Stable Set Relaxations

The max-cut and stable set problems are two fundamental **-hard problems in combinatorial optimization. It has been known for a long time that any instance of the stable set problem can be easily transformed into a max-cut instance. Moreover, Laurent, Poljak, Rendl and others have shown that any convex set containing the cut polytope yields, via a suitable projection, a convex set containing the stable set polytope. We review this work, and then extend it in the following ways. We show that the rounded version of certain `positive semidefinite' inequalities for the cut polytope imply, via the same projection, a surprisingly large variety of strong valid inequalities for the stable set polytope. These include the clique, odd hole, odd antihole, web and antiweb inequalities, and various inequalities obtained from these via sequential lifting. We also examine a less general class of inequalities for the cut polytope, which we call odd clique inequalities, and show that they are, in general, much less useful for generating valid inequalities for the stable set polytope. As well as being of theoretical interest, these results have algorithmic implications. In particular, we obtain as a by-product a polynomial-time separation algorithm for a class of inequalities which includes all web inequalities.

On the asymmetric representatives formulation for the vertex coloring problem

We consider the vertex coloring problem, which can be stated as the problem of minimizing the number of labels that can be assigned to the vertices of a graph G such that each vertex receives at least one label and the endpoints of every edge are assigned different labels. In this work, the 0-1 integer programming formulation based on representative vertices is revisited to remove symmetry. The previous polyhedral study related to the original formulation is adapted and generalized. New versions of facets derived from substructures of G are presented, including cliques, odd holes and anti-holes and wheels. In addition, a new class of facets is derived from independent sets of G. Finally, a comparison with the independent sets formulation is provided.

Algorithms for Perfectly Contractile Graphs

We consider the class ${\cal A}$ of graphs that contain no odd hole, no antihole of length at least 5, and no prism (a graph consisting of two disjoint triangles with three disjoint paths between them) and the class ${\cal A}'$ of graphs that contain no odd hole, no antihole of length at least 5, and no odd prism (prism whose three paths are odd). These two classes were introduced by Everett and Reed and are relevant to the study of perfect graphs. We give polynomial-time recognition algorithms for these two classes. In contrast we prove that determining if a general graph contains a prism (or an even prism, or an odd prism) is NP-complete.

A class of web-based facets for the generalized vertex packing problem

The generalized vertex packing problem seeks to identify a largest subset of nodes from an undirected graph, such that the subgraph induced by this subset of nodes contains no more than some threshold number of edges. This paper derives a class of valid inequalities based on certain special subgraphs called webs, which are general structures that subsume cliques, matchings, odd holes, and odd anti-holes. We also provide a set of conditions for this class of valid inequalities to be facet-inducing for the web subgraph polytope. Finally, we prescribe a web subgraph identification procedure, and test the computational benefits obtained by solving generalized vertex packing instances with formulations augmented by these web-based valid inequalities.

Antiwebs are rank-perfect

Webs and antiwebs are natural generalizations of odd holes and odd antiholes with circular symmetry of their maximum cliques and stable sets. Webs and antiwebs turned out to play a crucial role for describing the stable set polytopes for larger graph classes. In this short note we obtain, with the help of a result of Shepherd (1995), a complete description of the stable set polytopes for antiwebs showing that antiwebs are rank-perfect.

Coloring the Maximal Cliques of Graphs

In this paper we are concerned with the so-called clique-colorations of a graph, that is, colorations of the vertices so that no maximal clique is monochromatic. On one hand, it is known to be NP-complete to decide whether a perfect graph is 2-clique-colorable, or whether a triangle-free graph is 3-clique-colorable; on the other hand, there is no example of a perfect graph where more than three colors would be necessary. We first exhibit some simple recursive methods to clique-color graphs and then relate the chromatic number, the domination number, and the maximum cardinality of a stable set to the clique-chromatic number. We show exact bounds and polynomial algorithms that find the clique-chromatic number for some classes of graphs and prove NP-completeness results for some others, trying to find the boundary between the two. For instance, while it is NP-complete to decide whether a graph of maximum degree 3 is 2-clique-colorable, K1,3-free graphs without an odd hole turn out to be always 2-clique-colorable by a polynomial algorithm. Finally, we show that almost all perfect graphs are 3-clique-colorable.

Cliques, holes and the vertex coloring polytope

Certain subgraphs of a given graph G restrict the minimum number χ( G) of colors that can be assigned to the vertices of G such that the endpoints of all edges receive distinct colors. Some of such subgraphs are related to the celebrated Strong Perfect Graph Theorem, as it implies that every graph G contains a clique of size χ( G), or an odd hole or an odd anti-hole as an induced subgraph. In this paper, we investigate the impact of induced maximal cliques, odd holes and odd anti-holes on the polytope associated with a new 0–1 integer programming formulation of the graph coloring problem. We show that they induce classes of facet defining inequalities.

On the maximum feasible subsystem problem, IISs and IIS-hypergraphs

We consider the MAX FS problem: For a given infeasible linear system Ax b, determine a feasible subsystem containing as many inequalities as possible. This problem, which is NP-hard and also difficult to approximate, has a number of interesting applications in a wide range of fields. In this paper we examine structural and algorithmic properties of MAX FS and of Irreducible Infeasible Subsystems (IISs), which are intrinsically related since one must delete at least one constraint from each IIS to attain feasibility. First we provide a new simplex decomposition characterization of IISs and prove that finding a smallest cardinality IIS is very difficult to approximate. Then we discuss structural properties of IIS-hypergraphs, i.e., hypergraphs in which each edge corresponds to an IIS, and show that recognizing IIS-hypergraphs subsumes the Steinitz problem for polytopes and hence is NP-hard. Finally we investigate rank facets of the Feasible Subsystem polytope whose vertices are incidence vectors of feasible subsystems of a given infeasible system. In particular, using the IIS-hypergraph structural result, we show that only two very specific types of rank inequalities induced by generalized antiwebs (which generalize cliques, odd holes and antiholes to general independence systems) can arise as facets.

Graphs without odd holes, parachutes or proper wheels: a generalization of Meyniel graphs and of line graphs of bipartite graphs

We prove that the strong perfect graph conjecture holds for graphs that do not contain parachutes or proper wheels. This is done by showing the following theorem: If a graph G contains no odd hole, no parachute and no proper wheel, then G is bipartite or the line graph of a bipartite graph or G contains a star cutset or an extended strong 2- join or G ̄ is disconnected. To prove this theorem, we prove two decomposition theorems which are interesting in their own rights. The first is a generalization of the Burlet–Fonlupt decomposition of Meyniel graphs by clique cutsets and amalgams. The second is a precursor of the recent decomposition theorem of Chudnovsky, Robertson, Seymour and Thomas for Berge graphs that contain a line graph of a bipartite subdivision of a 3-connected graph.

On the divisibility of graphs

A graph G is k- divisible if for each induced subgraph H of G with at least one edge, there is a partition of the vertex set of H into sets V 1,…, V k such that no V i contains a maximum clique of H. We show that a claw-free graph is 2-divisible if and only if it does not contain an odd hole: we conjecture that this result is true for any graph, and present further conjectures relating 2-divisibility to the strong perfect graph conjecture. We also present related results involving the chromatic number and the stability number, with connections to Ramsey theory.

A Transformation Which Preserves the Clique Number

We introduce a graph transformation which preserves the clique number. When applied to graphs containing no odd hole and no cricket (a particular graph on 5 vertices) the transformation also preserves the chromatic number. Using this transformation we derive a polynomial algorithm for the computation of the clique number of all graphs in a class which strictly contains diamond-free graphs. Furthermore, the transformation leads to a proof that the Strong Perfect Graph Conjecture is true for two new classes of graphs and yields a polynomial time algorithm for the computation of the clique number and the chromatic number for both classes. One of these two classes strictly contains claw-free graphs.

Recognition of quasi-Meyniel graphs

We present a polynomial-time algorithm for recognizing quasi-Meyniel graphs. A hole is a chordless cycle with at least four vertices. A cap is a cycle with at least five vertices, with a single chord that forms a triangle with two edges of the cycle. A graph G is quasi-Meyniel if it contains no odd hole and for some x∈ V( G), the chord of every cap in G has x as an endvertex. Our recognition algorithm is based on star cutset decompositions.

Bull-Free Weakly Chordal Perfectly Orderable Graphs

Using doubly lexical orders and the notion of box partition due to de Figueiredo, Maffray, and Porto, we show that a certain subclass of bull-free weakly chordal graphs is perfectly orderable. This together with results of de Figueiredo, Maffray, and Porto confirms Chvatal's conjecture that bull-free graphs with no anti-hole and no odd hole are perfectly orderable; here hole means induced cycle with five or more vertices.

Graph Imperfection

We are interested in colouring a graph G=(V, E) together with an integral weight or demand vector x=(xv:v∈V) in such a way that xv colours are assigned to each node v, adjacent nodes are coloured with disjoint sets of colours, and we use as few colours as possible. Such problems arise in the design of cellular communication systems, when radio channels must be assigned to transmitters to satisfy demand and avoid interference. We are particularly interested in the ratio of chromatic number to clique number when some weights are large. We introduce a relevant new graph invariant, the “imperfection ratio” imp(G) of a graph G, present alternative equivalent descriptions, and show some basic properties. For example, imp(G)=1 if and only if G is perfect, imp(G)=imp(G) where G denotes the complement of G, and imp(G)=g/(g−1) for any line graph G where g is the minimum length of an odd hole (assuming there is an odd hole).

On the Structure of Bull-Free Perfect Graphs, 2: the Weakly Chordal Case

A bull is a graph obtained by adding a pendant vertex at two vertices of a triangle. A graph is perfectly orderable if it admits an ordering such that the greedy sequential method applied on this ordering produces an optimal coloring for every induced subgraph. Chvatal conjectured that every bull-free graph with no odd hole or antihole is perfectly orderable. In a previous paper we studied the structure of general bull-free perfect graphs, and reduced Chvatal's conjecture to the case of weakly chordal graphs. Here we focus on weakly chordal graphs, and we reduce Chvatal's conjecture to a restricted case. Our method lays out the structure of all bull-free weakly chordal graphs. These results have been used recently by Hayward to establish Chvatal's conjecture for this restricted case and therefore in full.

Steiner trees and polyhedra

In this paper we study the dominant of the Steiner tree polytope. We introduce a new class of valid inequalities that generalizes the so-called odd hole, wheel, bipartite, anti-hole and Steiner partition inequalities introduced by Chopra and Rao (Math. Programming 64 (1994) 209–229, 231–246), and we give sufficient conditions for these inequalities to define facets. We describe some procedures that permit to construct facets from known ones for the dominant of the Steiner tree polytope and the closely related Steiner connected subgraph polytope. Using these methods we give a counterexample to a conjecture of Chopra and Rao on the dominant of the Steiner tree polytope on 2-trees. We also describe the dominant of the Steiner tree polytope and the Steiner connected subgraph polytope on special classes of graphs. In particular, we show that if the underlying graph is series–parallel and the terminals satisfy certain conditions, then both polyhedra are given by the trivial inequalities and the Steiner partition inequalities.