On linear spaces and matroids of arbitrary cardinality

In this paper, we study linear spaces of arbitrary finite dimension on some (possibly infinite) set. We interpret linear spaces as simple matroids and study the problem of erecting some linear space of dimension n to some linear space of dimension n + 1 if possible. Several examples of some such erections are studied; in particular, one of these erections is computed within some infinite iteration process.Dedicated to the memory of Gian-Carlo Rota     

Recipe theorem for the Tutte polynomial for matroids,renormalization group-like approach

Using a quantum field theory renormalization group-like differential equation, we give a new proof of the recipe theorem for the Tutte polynomial for matroids. The solution of such an equation is in fact given by some appropriate characters of the Hopf algebra of isomorphic classes of matroids, characters which are then related to the Tutte polynomial for matroids. This Hopf algebraic approach also allows to prove, in a new way, a matroid Tutte polynomial convolution formula appearing in [W. Kook, V. Reiner, D. Stanton, A convolution formula for the Tutte polynomial, J. Combin. Theory Ser. B 76 (1999) 297–300] and [G. Etienne, M. Las Vergnas, External and internal elements of a matroid basis, Discrete Math. 179 (1998) 111–119].     

On the contraction of vectorial lattice representations

For modular lattices of finite length, vector space representations are shown to give rise to contracted representations of homomorphic imagesDedicated to the memory of Alan Day     

Inequivalent representations of matroids over prime fields

It is proved that for each prime field GF(p)$\mathit{GF}(p)$, there is an integer _np$n_p$ such that a 4-connected matroid has at most _np$n_p$ inequivalent representations over GF(p)$\mathit{GF}(p)$. We also prove a stronger theorem that obtains the same conclusion for matroids satisfying a connectivity condition, intermediate between 3-connectivity and 4-connectivity that we term “k-coherence”.     

A geometric characterization of Coxeter matroids

Coxeter matroids, introduced by Gelfand and Serganova, are combinatorial structures associated with any finite Coxeter group and its parabolic subgroup they include ordinary matroids as a specia case. A basic result in the subject is a geometric characterization of Coxeter matroids first stated by Gelfand and Serganova. This paper presents a self-contained, simple proof of a more general version of this geometric characterization.     

Comparability graphs of lattices

A theorem of N. Terai and T. Hibi for finite distributive lattices and a theorem of Hibi for finite modular lattices (suggested by R.P. Stanley) are equivalent to the following: if a finite distributive or modular lattice of rank d contains a complemented rank 3 interval, then the lattice is (d+1)-connected.In this paper, the following generalization is proved: Let L be a (finite or infinite) semimodular lattice of rank d that is not a chain (d∈N₀). Then the comparability graph of L is (d+1)-connected if and only if L has no simplicial elements, where z∈L is simplicial if the elements comparable to z form a chain.     

Young's natural idempotents as polynomials

Coding permutations as monomials, one obtains a compact expression of representatives of Young's natural idempotents for the symmetric group, or ofq-idempotents in the Hecke algebra.     

Weights of multipartitions and representations of Ariki-Koike algebras

We consider representations of the Ariki-Koike algebra, a q-deformation of the group algebra of the complex reflection group C_r?S_n. The representations of this algebra are naturally indexed by multipartitions of n, and for each multipartition λ we define a non-negative integer called the weight of λ. We prove some basic properties of this weight function, and examine blocks of small weight.     

Skew partial fields,multilinear representations of matroids,and a matrix tree theorem

We extend the notion of representation of a matroid to algebraic structures that we call skew partial fields. Our definition of such representations extends Tutte?s definition, using chain groups. We show how such representations behave under duality and minors, we extend Tutte?s representability criterion to this new class, and we study the generator matrices of the chain groups. An example shows that the class of matroids representable over a skew partial field properly contains the class of matroids representable over a skew field. Next, we show that every multilinear representation of a matroid can be seen as a representation over a skew partial field. Finally we study a class of matroids called quaternionic unimodular. We prove a generalization of the matrix tree theorem for this class.     

h-Vectors of matroids and logarithmic concavity

Let M be a matroid on E, representable over a field of characteristic zero. We show that h-vectors of the following simplicial complexes are log-concave:

1.

The matroid complex of independent subsets of E.     

Symmetrization of Closure Operators and Visibility

For an arbitrary set E and a given closure operator , we want to construct a symmetric closure operator via some – possibly infinite – iteration process. If E is finite, the corresponding symmetric closure operator . defines a matroid. If and is the convex closure operator, turns out to be the affine closure operator. Moreover, we apply the symmetrization process to closure operators induced by visibility. Received March 9, 2005     

Recognizing graphic matroids

There is no polynomially bounded algorithm to test if a matroid (presented by an “independence oracle”) is binary. However, there is one to test graphicness. Finding this extends work of previous authors, who have given algorithms to test binary matroids for graphicness. Our main tool is a new result that ifM′ is the polygon matroid of a graphG, andM is a different matroid onE(G) with the same rank, then there is a vertex ofG whose star is not a cocircuit ofM.     

Identities of cofinal sublattices

If V is a variety of lattices and L a free lattice in V on uncountably many generators, then any cofinal sublattice of L generates all of V. On the other hand, any modular lattice without chains of order-type +1 has a cofinal distributive sublattice. More generally, if a modular lattice L has a distributive sublattice which is cofinal modulo intervals with ACC, this may be enlarged to a cofinal distributive sublattice. Examples are given showing that these existence results are sharp in several ways. Some similar results and questions on existence of cofinal sublattices with DCC are noted.This work was done while the first author was partly supported by NSF contract MCS 82-02632, and the second author by an NSF Graduate Fellowship.     

The expressive power of voting polynomials

We consider the problem of approximating a Boolean functionf∶{0,1} ⁿ →{0,1} by the sign of an integer polynomialp of degreek. For us, a polynomialp(x) predicts the value off(x) if, wheneverp(x)≥0,f(x)=1, and wheneverp(x)<0,f(x)=0. A low-degree polynomialp is a good approximator forf if it predictsf at almost all points. Given a positive integerk, and a Boolean functionf, we ask, “how good is the best degreek approximation tof?” We introduce a new lower bound technique which applies to any Boolean function. We show that the lower bound technique yields tight bounds in the casef is parity. Minsky and Papert [10] proved that a perceptron cannot compute parity; our bounds indicate exactly how well a perceptron canapproximate it. As a consequence, we are able to give the first correct proof that, for a random oracleA, PP ^A is properly contained in PSPACE ^A . We are also able to prove the old AC⁰ exponential-size lower bounds in a new way. This allows us to prove the new result that an AC⁰ circuit with one majority gate cannot approximate parity. Our proof depends only on basic properties of integer polynomials.     

Polar Graphs and Corresponding Involution Sets, Loops and Steiner Triple Systems

A 1-factorization (or parallelism) of the complete graph with loops is called polar if each 1-factor (parallel class) contains exactly one loop and for any three distinct vertices x₁, x₂, x₃, if {x₁} and {x₂, x₃} belong to a 1-factor then the same holds for any permutation of the set {1, 2, 3}. To a polar graph there corresponds a polar involution set , an idempotent totally symmetric quasigroup (P, *), a commutative, weak inverse property loop (P, + ) of exponent 3 and a Steiner triple system . We have: satisfies the trapezium axiom is self-distributive ⇔ (P, + ) is a Moufang loop is an affine triple system; and: satisfies the quadrangle axiom is a group is an affine space.     

Euclideaness and final polynomials in oriented matroid theory

This paper deals with a geometric construction of algebraic non-realizability proofs for certain oriented matroids. As main result we obtain an algorithm which generates a (bi-quadratic) final polynomial [3], [5] for any non-euclidean oriented matroid. Here we apply the results of Edmonds, Fukuda and Mandel [6], [7] concerning non-degenerate cycling of linear programs in non-euclidean oriented matroids.     

Sublattices of product spaces: Hulls, representations and counting

The Cartesian product of lattices is a lattice, called a product space, with componentwise meet and join operations. A sublattice of a lattice L is a subset closed for the join and meet operations of L. The sublattice hullLQ of a subset Q of a lattice is the smallest sublattice containing Q. We consider two types of representations of sublattices and sublattice hulls in product spaces: representation by projections and representation with proper boundary epigraphs. We give sufficient conditions, on the dimension of the product space and/or on the sublattice hull of a subset Q, for LQ to be entirely defined by the sublattice hulls of the two-dimensional projections of Q. This extends results of Topkis (1978) and of Veinott [Representation of general and polyhedral subsemilattices and sublattices of product spaces, Linear Algebra Appl. 114/115 (1989) 681-704]. We give similar sufficient conditions for the sublattice hull LQ to be representable using the epigraphs of certain isotone (i.e., nondecreasing) functions defined on the one-dimensional projections of Q. This also extends results of Topkis and Veinott. Using this representation we show that LQ is convex when Q is a convex subset in a vector lattice (Riesz space), and is a polyhedron when Q is a polyhedron in Rⁿ.We consider in greater detail the case of a finite product of finite chains (i.e., totally ordered sets). We use the representation with proper boundary epigraphs and provide upper and lower bounds on the number of sublattices, giving a partial answer to a problem posed by Birkhoff in 1937. These bounds are close to each other in a logarithmic sense. We define a corner representation of isotone functions and use it in conjunction with the representation with proper boundary epigraphs to define an encoding of sublattices. We show that this encoding is optimal (up to a constant factor) in terms of memory space. We also consider the sublattice hull membership problem of deciding whether a given point is in the sublattice hull LQ of a given subset Q. We present a good characterization and a polynomial time algorithm for this sublattice hull membership problem. We construct in polynomial time a data structure for the representation with proper boundary epigraphs, such that sublattice hull membership queries may be answered in time logarithmic in the size |Q| of the given subset.     

A generalization of the ingleton—Main lemma and a class of non-algebraic matroids

Dedicated to professor Jan-Erik Roos on his fifthieth birthday     

Decomposition of binary matroids

We prove results relating to the decomposition of a binary matroid, including its uniqueness when the matroid is cosimple. We extend the idea of “freedom” of an element in a matroid to “freedom” of a set, and show that there is a unique maximal integer polymatroid inducing a given binary matroid.