Let \(S\) and \(T\) be sets with \(|S| = m\) and \(|T| = n\). Let \(S_3, S_2\) and \(T_3, T_2\) be the sets of all \(3\)-subsets (\(2\)-subsets) of \(S\) and \(T\), respectively. Define \(Q((m, 2, 3), (n, 2, 3))\) as the smallest subset of \(S_2 \times T_2\) needed to cover all elements of \(S_3 \times T_3\). A more general version of this problem is initially defined, but the bulk of the investigation is devoted to studying this number. Its property as a lower bound for a planar crossing number is the reason for this focus.

Under some assumptions on the incidence matrices of symmetric designs, we prove a non-existence theorem for symmetric designs. The approach generalizes Wilbrink’s result on difference sets \([7]\).

In this paper, we derive some inequalities which the parameters of a two-symbol balanced array \(T\) (\(B\)-array) of strength four must satisfy for \(T\) to exist.

This paper considers Latin squares of order \(n\) having \(0, 1, \ldots, n-1\) down the main diagonal and in which the back diagonal is a permutation of these symbols (diagonal squares). It is an open question whether or not such a square which is self-orthogonal (i.e., orthogonal to its transpose) exists for order \(10\). We consider two possible constraints on the general concept: self-conjugate squares and strongly symmetric squares. We show that relative to each of these constraints, a corresponding self-orthogonal diagonal Latin square of order \(10\) does not exist. However, it is easy to construct self-orthogonal diagonal Latin squares of orders \(8\) and \(12\) which satisfy each of the constraints respectively.

It has been conjectured by D. R. Stinson that an incomplete Room square \((n, s)\)-IRS exists if and only if \(n\) and \(s\) are both odd and \(n \geq 3s + 2\), except for the nonexistent case \((n, s) = (5, 1)\). In this paper we shall improve the known results and show that the conjecture is true except for \(45\) pairs \((n, s)\) for which the existence of an \((n, s)\)-IRS remains undecided.

Let \(f(n)\) denote the number of essentially different factorizations of \(n\). In this paper, we prove that for every odd number \( > 1\), we have \(f(n) \leq c\frac{n}{\log n},\) where \(c\) is a positive constant.

A partition of the edge set of a hypergraph \(H\) into subsets inducing hypergraphs \(H_1,\ldots,H_r\) is said to be a \({decomposition}\) of \(H\) into \(H_1,\ldots,H_r\). A uniform hypergraph \(F = (\bigcup \mathcal{F}, \mathcal{F})\) is a \(\Delta\)-\({system}\) if there is a set \(K \subseteq V(F)\), called the \({kernel}\) of \(F\), such that \(A \cap B = K\) for every \(A, B \in \mathcal{F}\), \(A \neq B\). A disjoint union of \(\Delta\)-systems whose kernels have the same cardinality is said to be a \(constellation\). In the paper, we find sufficient conditions for the existence of a decomposition of a hypergraph \(H\) into:

a) \(\Delta\)-systems having almost equal sizes and kernels of the same cardinality,

b) isomorphic copies of constellations such that the sizes of their components are relatively prime.

In both cases, the sufficient conditions are satisfied by a wide class of hypergraphs \(H\).

The binding number of a graph \(G\) is defined to be the minimum of \(|N(S)|/|S|\) taken over all nonempty \(S \subseteq V(G)\) such that \(N(S) \neq V(G)\). In this paper, another look is taken at the basic properties of the binding number. Several bounds are established, including ones linking the binding number of a tree to the “distribution” of its end-vertices. Further, it is established that under some simple conditions, \(K_{1,3}\)-free graphs have binding number equal to \((p(G) – 1)/(p(G) – \delta(G))\) and applications of this are considered.

Strongly regular graphs are graphs in which every adjacent pair of vertices share \(\lambda\) common neighbours and every non-adjacent pair share \(\mu\) common neighbours. We are interested in strongly regular graphs with \(\lambda = \mu = k\) such that every such set of \(k\) vertices common to any pair always induces a subgraph with a constant number \(x\) of edges. The Friendship Theorem proves that there are no such graphs when \(\lambda = \mu = 1\). We derive constraints which such graphs must satisfy in general, when \(\lambda = \mu > 1\), and \(x \geq 0\), and we find the set of all parameters satisfying the constraints. The result is an infinite, but sparse, collection of parameter sets. The smallest parameter set for which a graph may exist has \(4896\) vertices, with \(k = 1870\).

The idea of a domino square was first introduced by J. A. Edwards et al. in [1]. In the same paper, they posed some problems on this topic. One problem was to find a general construction for a whim domino square of side \(n \equiv 3 \pmod{4}\). In this paper, we solve this problem by using a direct construction. It follows that a whim domino square exists for each odd side [1].