Expansion and contraction functors on matriods
Rahim Rahmati-Asghar

TL;DR
This paper investigates how expansion and contraction functors affect matroid properties, showing that certain properties are preserved under expansions and introducing contraction as a new tool to analyze White's conjecture.
Contribution
It introduces the contraction functor for matroids, establishes the equivalence of properties under expansions, and links White's conjecture to expansions and contractions.
Findings
Expansion preserves graphic, binary, and transversal properties.
White's conjecture holds for a matroid if and only if it holds for all its expansions.
Some classes of matroids are identified that satisfy White's conjecture.
Abstract
Let be a matroid. We study the expansions of mainly to see how the combinatorial properties of and its expansions are related to each other. It is shown that is a graphic, binary or a transversal matroid if and only if an arbitrary expansion of has the same property. Then we introduce a new functor, called contraction, which acts in contrast to expansion functor. As a main result of paper, we prove that a matroid satisfies White's conjecture if and only if an arbitrary expansion of does. It follows that it suffices to focus on the contraction of a given matroid for checking whether the matroid satisfies White's conjecture. Finally, some classes of matroids satisfying White's conjecture are presented.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvanced Graph Theory Research · Commutative Algebra and Its Applications · Advanced Combinatorial Mathematics
Expansion and contraction functors on matriods ††
- Corresponding Author.
2010 Mathematics Subject Classification: Primary 05B35; Secondary 52B40.
Key words and phrases: expansion functor, contraction functor, White’s conjecture.
Rahim Rahmati-Asghar∗
*Department of Mathematics, Faculty of Basic Sciences, University of Maragheh, P. O. Box 55181-83111, Maragheh, Iran.
e-mail* : [email protected]
Abstract. Let be a matroid. We study the expansions of mainly to see how the combinatorial properties of and its expansions are related to each other. It is shown that is a graphic, binary or a transversal matroid if and only if an arbitrary expansion of has the same property. Then we introduce a new functor, called contraction, which acts in contrast to expansion functor. As a main result of paper, we prove that a matroid satisfies White’s conjecture if and only if an arbitrary expansion of does. It follows that it suffices to focus on the contraction of a given matroid for checking whether the matroid satisfies White’s conjecture. Finally, some classes of matroids satisfying White’s conjecture are presented.
Introduction
Matroids are abstract combinatorial structures that capture the notion of independence that is common to a surprisingly large number of mathematical entities. They were introduced by Whitney in 1935 as a common generalization of independence in linear algebra and independence in graph theory [21]. Matroid theory is one of the most fascinating research areas in combinatorics. It was linked to projective geometry by Mac Lane [12], and have found a great many applications in several branches of mathematics [20]. In this regard studying the structural properties of matroids from different point views have been considered by many researchers. Also classifying matroids with a desired property or making modifications to a matroid so that it satisfies a special property is the subject of many research papers, see for example [1, 7, 9, 11, 15, 17, 20].
The notion of expansion is a known notion appeared in different terminologies as parallelization or duplication in combinatorics [5, 4, 13, 16]. In [15], the authors studied behaviors of expansion functor on some algebraic structures associated to discrete polymatroids. It was shown that a nonempty finite set is a discrete polymatroid if and only if its an arbitrary expansion is a discrete polymatroid (c.f. [15, Theorem 1.2.]). The discrete polymatroid is a multiset analogue of the matroid. Moreover, there are several classes of matroids so that the study of each of them is interesting in its own right. This motivates to focus on the behaviors of the expansion functor on some classes of matroids and to investigate some structural properties of them in this paper. Our goal in this paper is to investigate more relations between the exchange property of bases of a matroid and those of its expansions. It turns out that the exchange properties of bases of a matroid are preserved under taking the expansion functor and so this construction is a very good tool to make new matroids with a desired property. Moreover, by taking another functor, contraction functor, which is the opposite to the expansion functor we will able to construct a new matroid, with possibly smaller ground set, from a given matroid and check a desired property on new matroid instead of primary one.
White in 1980 proposed a conjecture about the bases of a matroid [20]. This conjecture has received much attention in recent years and has some algebraic and combinatorial variants, all of which are open problems. Up to now, several mathematicians confirmed only some variants of this conjecture for special classes of matroids (see for example [1, 2, 7, 9, 11, 15, 17, 18]).
White [20] defined three classes , and of matroids and conjectured that the class of all matroids.
We investigate the effect of the expansion functor on the exchange property for bases of matroids and conclude that White’s conjecture is preserved under taking the expansion or contraction functor.
The paper is organized as follows. In Section 1, we review some preliminaries which are needed in the sequel. In Section 2, we investigate the expansion of some classes of matroids. We show that a matroid is graphic, binary or transversal if and only if its an arbitrary expansion has such a property (see Theorems 2.1, 2.4 and 2.7). Also, we prove that the expansion of an uniform matroid is a partition matroid and, conversely, every partition matroid is an expansion of an uniform matroid (see Theorem 2.9). In Section 3, we introduce the contraction functor which acts in contrast to expansion functor. The last section is devoted to the study of unique exchange property. After recalling some notions and notations from [20], we formulate White’s conjecture [20, Conjecture 12]. As one of the main results, we show that a matroid satisfies White’s conjecture if and only if an arbitrary expansion of does (see Theorem 3.6). This concludes that satisfies White’s conjecture if and only if its contraction does (see Corollary 3.11). On the other hand, since the class of contracted matroids is very smaller than the class of all matroids, it follows from Corollary 3.11 that to test White’s conjecture for a given class of matroids it suffices to turn our attention to their contractions. Finally, we give some classes of matroids which satisfy White’s conjecture.
1 Preliminaries
A matroid is a pair consisting of a finite set and a non-empty family of subsets of such that no set in properly contains another set in and, moreover, satisfies the following exchange property:
for every and there exists , such that .
and are, respectively, called the ground set and the basis set of . The background from matroid theory which we use may be obtained from [14] or [19].
Recall from [15] the concept of expansion functor on a family of subsets of . Let . For , the expansion of is defined
[TABLE]
Let be a family of subsets of and . Set . The expansion of the singleton family with respect to is denoted by and it is a family of subsets of defined as follows:
[TABLE]
Also, the expansion of with respect to is denoted by and it is defined
[TABLE]
Let denote the set of all subsets of and let . We define the map by setting .
The following theorem is a direct consequence of [15, Theorem 1.2]:
Theorem 1.1**.**
Let be a nonempty family of subsets of and let . Then is a matroid if and only if is.
The restriction of a matroid to is denoted by and it is a matroid with the ground set and the basis set
[TABLE]
2 The expansion of some classes of matroids
Let be an undirected graph (with possibly loops or parallel edges). A spanning subgraph of a graph is a subgraph whose vertex set is the entire vertex set of . If this spanning subgraph is a tree, it is called a spanning tree of the graph.
Let and . Then is a matroid called a graphic matroid. The graphic matroid associated with the graph is denoted by .
Theorem 2.1**.**
Let be a matroid on and . Then is graphic if and only if is graphic.
We need some notations and an auxiliary lemma:
For any , let be defined as a_{j}=\left\{\begin{array}[]{ll}0&\hbox{if}\ j\neq i\\ 1&\hbox{if}\ j=i.\end{array}\right. Set .
Lemma 2.2**.**
Let be a family of subsets of , and . Then .
Proof.
Note that
[TABLE]
and
[TABLE]
Define given by
[TABLE]
Then induces the bijection
[TABLE]
∎
Now we prove Theorem 2.1:
Proof.
Let be a graph with and . We use induction on .
First, suppose that and for all . Let be a graph with the vertex set and the edge set where for all and and are parallel. It is easy to check that .
Now, suppose that and with if and . Assume that is graphic. Then it follows from induction hypothesis and Lemma 2.2 that where is obtained from by adding a parallel edge.
Conversely, suppose is graphic and . Set . Then is graphic by [8, page 842]. ∎
Example 2.3*.*
Consider the matroid on with basis set
[TABLE]
[TABLE]
is a graphic matroid associated with the graph shown in Figure 1. Let . Then is a graphic matroid and is associated with (see Figure 1), obtained from by adding parallel edges to and .
A subset of the ground set of a matroid that is contained in no bases of is called dependent. A circuit in is a minimal dependent subset (with respect to inclusion) of and the set of circuits of is denoted by . A matroid is binary if and only if for every pair of circuits of , their symmetric difference contains another circuit. See [14, Theorem 9.1.2] for other equivalent definitions of binary matroids.
Theorem 2.4**.**
Let be a matroid on and . Then is binary if and only if is.
Proof.
Let be binary and let . If then there exist and with . Set . Then and , and hence the assertion is completed. So suppose that .
If and then since there exists such that . Since we have and so it follows that for some with .
Consider or . Let, for example, . Then for some . Since , thus it should be . Consider the following cases:
- •
: Then for some and . Since we have . Set .
- •
: Then and or where . At the first case, set and at the second one, set .
Thus and . Therefore is binary.
Conversely, suppose that is binary and . Let and . Set and . Then and so there exists with . It follows that and . Therefore is binary. ∎
We recall from [14, page 46.] the definition of a transversal matroid. Let . A set system is a set along with a multiset of (not necessarily distinct) subsets of . If then we may denote by . A transversal of is a subset of for which there is a bijection with for all .
Theorem 2.5**.**
[6]** A finite set system has a transversal if and only if, for all ,
[TABLE]
If , then is a partial transversal of if, for some subset of , is a transversal of . The partial transversals of a are the independent sets of a matroid. We call such a matroid a transversal matroid and denote it by . is called a presentation of .
Theorem 2.6**.**
[3]** Let be a transversal matroid on . Then so is for each . If is a presentation of , then is a presentation of .
If is a family of subsets of then the bipartite graph associated with , denoted by , has the vertex set and the edge set .
A matching in a graph is the set of edges in no two of which have a common endpoint. A subset of is a partial transversal of if and only if there is a matching in which every edge has one endpoint in .
Theorem 2.7**.**
Let be a matroid on and let . Then is transversal if and only if is.
Proof.
“Only if part”: Let be a transversal matroid with where and . Set . We claim that is a presentation of . We associate to and , respectively, the vertices and .
Let . We may assume that . So there exists the maximal matching in the bipartite graph with the partition . It is clear that is a matching in . Suppose, on the contrary, that is not maximal. So for some matching in we have . Let . Since is maximal, we have . Moreover, it is clear that . Let and let . By Theorem 2.6, is transversal with presentation . But . This contradicts Theorem 2.5. Therefore is a maximal matching in and so .
In a similar argument we show that . Therefore , as desired.
“If part”: Let be transversal with the presentation . One may suppose that . By Theorem 2.6, is a transversal matroid with the presentation . Since , the assertion is completed.
∎
Example 2.8*.*
Let and . Let . Then the matchings in are
[TABLE]
and so
[TABLE]
and are shown in Figure 2.
A matroid on of rank is an uniform matroid if all -element subsets of are bases and it is denoted by .
A partition matroid of rank [10] is a matroid associated with a partition of and the basis set
[TABLE]
Theorem 2.9**.**
The expansion of every uniform matroid is a partition matroid. Conversely, every partition matroid is the expansion of an uniform matroid.
Proof.
Let be an uniform matroid on of rank and let . Set for all . Then it is easy to see that where .
Conversely, let be a partition matroid of rank with . Set for all . Then where and for all . ∎
3 The contraction functor
Definition 3.1**.**
Let be a family of subsets of and let denote the maximal elements of (with respect to inclusion). We define the relation ” on in the following form:
[TABLE]
In other words,
[TABLE]
It is easily shown that is an equivalence relation. Let be the set of equivalence classes under .
Let . Set . For , define and a family of subsets of with the set of maximal elements. We call the contraction of by . Clearly, every family of subsets of has an unique contraction.
A family of subsets of is called contracted if and coincide up to a relabeling.
Remark 3.2*.*
Note that the contraction functor behaves exactly the opposite to expansion functor. Actually, if is a family of subsets of and is the contraction of by , then and coincide up to a relabeling of . Also, for every , two families and coincide. Therefore is a matroid if and only if is a matroid. Equivalently, by Theorem 1.1, is a matroid.
Remark 3.3*.*
All of uniform matroids of rank are contracted.
Corollary 3.4**.**
The contraction of a partition matroid is an uniform matroid.
Proof.
Let be a partition matroid on . By Theorem 2.9, is the expansion of an uniform matroid with respect to some . It follows from Remarks 3.2 and 3.3 that . ∎
In view of Theorems 2.1, 2.4, 2.7 and 2.9 we have the following:
Corollary 3.5**.**
The contraction of a graphic (resp. binary, transversal) matroid is graphic (resp. binary, transversal).
4. Unique exchange property for bases with a view towards White’s conjecture
In this section, we investigate the preservation of White’s conjecture under taking the expansion and contraction functors. First, we recall some notions and notations from [20].
Let be a matroid. Two sequences of bases and are compatible if . Let be a sequence of bases of a matroid and let and . Set
[TABLE]
Let . Then we say
[TABLE]
is obtained from by a symmetric exchange. For two sequences of bases of a matroid , means that is obtained from by a symmetric exchange. Also, implies to is obtained from by a symmetric exchange or a permutation of the order of the bases.
If , we let
[TABLE]
Let . Then we say is obtained from by a symmetric subset exchange. We write if is obtained from by a symmetric subset exchange.
For , set the class of matroids with the property that for every matroid and every two compatible sequences and of bases of , there exist the sequences , for , of bases of such that
[TABLE]
It is easy to check that . White conjectured that theses classes are equal to the class of all matroids [20, Conjecture 12]. We will say that a matroid satisfies White’s conjecture if , for all .
Theorem 3.6**.**
For a matroid on , and we have
[TABLE]
Before proving Theorem 3.6, we present some auxiliary results.
Lemma 3.7**.**
Let be a matroid on and .
- (i)
For any two compatible sequences and of bases of , there exist two compatible sequences and of bases of such that and ; 2. (ii)
If and are two compatible sequences of bases of then and are two compatible sequences of bases of .
Lemma 3.8**.**
Let be a matroid on and let . For , if is obtained from by a symmetric subset exchange then
- (i)
* for or* 2. (ii)
* is obtained from by a symmetric subset exchange.*
Lemma 3.9**.**
Let be a matroid on and let . Suppose that is obtained from by a symmetric subset exchange. Let and be two compatible sequences of bases of with , . Then is obtained from by a symmetric subset exchange.
Now we prove Theorem 3.6 in three parts:
Membership in :
Assume that and let and be two compatible sequences of bases of . Let and for . By Lemma 3.7(ii), and are compatible and, by the assumption, is obtained from by a composition of symmetric subset exchanges. It follows from Lemma 3.9 that is obtained from by a composition of symmetric subset exchanges. Therefore .
Similarly, the converse direction follows from Lemmas 3.7(i) and 3.8.
Membership in :
Let . Let and be two compatible sequences of . By Lemma 3.7(ii), and are two compatible sequences of bases of and so, by the assumption, is obtained from by a composition of symmetric exchanges and permutations of the order of the bases. It follows from Lemma 3.9 that is obtained from by a composition of symmetric exchanges and permutations of the order of the bases.
The converse direction obtains in a similar argument by using Lemmas 3.7(i) and 3.8.
Membership in :
Lemma 3.10**.**
([11]) Let be a matroid. Then if and only if and any pair of bases of is obtained from by a composition of symmetric exchanges.
Since if and only if , it suffices to show that any pair of bases of is obtained from by a composition of symmetric exchanges if and only if any pair of bases of has this property.
Suppose that any pair of bases of is obtained from by a composition of symmetric exchanges. Consider a pair of bases of . Let . By the assumption, is obtained from by a composition of symmetric exchanges. Therefore
[TABLE]
By Lemma 3.7(i), one can choose the bases with such that are pairwise compatible. It follows from Lemma 3.9 that is obtained from by a composition of symmetric exchanges.
The converse follows from Lemmas 3.7(ii) and 3.8 in a similar argument.
Corollary 3.11**.**
Let be a matroid on and let . Then satisfies White’s conjecture if and only if the contraction of does.
Remark 3.12*.*
Note that the class of contracted matroids is very smaller than the class of all matroids. It follows from Corollary 3.11 that to test White’s conjecture for a class of matroids it suffices to turn our attention to their contractions.
Corollary 3.13**.**
Every partition matroid satisfies White’s conjecture.
Proof.
Let be a partition matroid. By Corollary 3.4, the contraction of is an uniform matroid. In view of Corollary 3.11, if we show that then the assertion is completed. It was shown in [18] that the toric ideal of any uniform matroid is generated by quadratic binomials corresponding to symmetric exchanges and this is an algebraic meaning of the property . Hence . Now suppose that is a pair of bases of . Let and . Let and for . Then
It follows from Lemma 3.10 that , as desired. ∎
Let and be matroids on disjoint ground sets. The direct sum of and is denoted by and it is a matroid, by [14, Proposition 4.2.12], on with the basis set
[TABLE]
Lemma 3.14**.**
Let and be matroids on disjoint ground sets. If and satisfy White’s conjecture then does, too.
The proofs of above lemma is easy and we leave them to the reader.
Remark 3.15*.*
(i) Let be a matroid on and let be disjoint from ’s. Then satisfies White’s conjecture if and only if the matroid with the basis set does.
(ii) Let be a matroid on . Consider a set disjoint from . Let . Then the set
[TABLE]
is the basis set of a matroid . In fact, where and for all . Especially, if satisfies White’s conjecture then it follows from Lemma 3.14 and Corollary 3.13 that satisfies White’s conjecture, too.
Combining Theorem 2.4 with Theorem 3.6 we obtain
Corollary 3.16**.**
A binary matroid satisfies White’s conjecture if and only if its contraction does.
Example 3.17*.*
Consider the matroid on with the basis set
[TABLE]
[TABLE]
[TABLE]
It is easy to check that is binary. To see that satisfies White’s conjecture, we first contract . The contraction of is a binary matroid on with the basis set
[TABLE]
On the other hand, can be obtained by adding to all bases of the uniform matroid with the ground set . It follows from Remark 3.15(i) that satisfies White’s conjecture and so satisfies White’s conjecture, too.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] J. Blasiak, The toric ideal of a graphic matroid is generated by quadrics , Combinatorica 28 (3) (2008) 283-297.
- 2[2] J. Bonin, Basis-exchange properties of sparse paving matroids , Adv. in Appl. Math. 50 (2013) no. 1, 6-15.
- 3[3] R. A. Brualdi, Transversal matroids, in: Combinatorial Geometries , N. White, ed. (Cambridge Univ. Press, Cambridge (1987)) 72-97.
- 4[4] C. A. Francisco; H. T. Hà, A. Van Tuyl, Coloring of hypergraphs, perfect graphs and associated primes of powers of monomial ideals. J. Algebra 331 (2011), 224-242.
- 5[5] M. C. Golumbic, Algorithmic graph theory and perfect graphs , second edition, Annals of Discrete Mathematics 57 , Elsevier Science B.V., Amsterdam, 2004.
- 6[6] P. Hall, On representatives of subsets. J. London Math. Soc. 10 (1935) 26-30.
- 7[7] W. Hochstättler, The toric ideal of a cographic matroid is generated by quadrics , Technical Report feu-dmo 023.10, Fern Universitat in Hagen Fakultät für Mathematik und Informatik.
- 8[8] K.D. Joshi, Applied Discrete Structures , New Age International(P) limited, publisher (1997).
