Two classical theorems about tournaments state that a tournament with no less than eight vertices admits an antidirected Hamiltonian path and an even cardinality tournament with no less than sixteen vertices admits an antidirected Hamiltonian cycle. Sequential algorithms for finding such a path as well as a cycle follow directly from the proofs of the theorems. Unfortunately, these proofs are inherently sequential and cannot be exploited in a parallel context. In this paper, we propose new proofs leading to efficient parallel algorithms.

In this article, we discuss the number of pairwise orthogonal Latin squares and obtain the estimate \(n_r < 8(r + 1)2^{4r}\) for \(r \geq 2\).

In this paper, we present some results on the existence of balanced arrays (B-arrays) with two symbols and of strength four by using some inequalities involving the statistical concepts of skewness and kurtosis. We demonstrate also, through an illustrative example, that in certain situations, the results given here lead to sharper upper bounds on the number of constraints for B-arrays.

If \(\alpha\) is a primitive root of the finite field \({GF}(2^n)\), we define a function \(\pi_n\) on the set \({E}_n = \{1, 2, \ldots, 2^n – 2\}\) by
\[
\pi_\alpha(i) = j \quad \text{iff} \quad \alpha^i = 1 + \alpha^{j}.
\]
Then \(\pi_\alpha\) is a permutation of \({E}_n\) of order \(2\). The path-length of \(\pi\), denoted \({PL}(\pi)\), is the sum of all the quantities \(|\pi(i) – i|\),
and the rank of \(\pi\) is the number of pairs \((i, j)\) with \(i \pi(j)\). We show that \({PL}(\pi) = {2(2^n – 1)(2^{n-1} – 1)}/{3}\), and the rank of \(\pi\) is \((2^{n-1} – 1)^2\).

If \(\gcd(k, 2^n – 1) = 1\), then \(M_k(x) = kx(\mod{2^n – 1})\) is a permutation of \({E}_n\). We show that a necessary condition for the function \(f_i(x) = 1 + x + \cdots + x^{i}\) to be a permutation of \({GF}(2^n)\), is that the function \(g_k(r) = \pi(M_{k+1}(r)) – \pi(r)\) be a permutation of \({E}_n\) such that exactly half the members \(r\) of \({E}_n\) satisfy \(g_k(r) r\).

Let \((G, \cdot)\) be a group with identity element \(e\) and with a unique element \(h\) of order \(2\). In connection with an investigation into the admissibility of linear groups, one of the present authors was recently asked if, for every cyclic group \(G\) of even order greater than \(6\), there exists a bijection \(\gamma\)
from \(G \setminus \{e, h\}\) to itself such that the mapping \(\delta: g \to g \cdot \gamma(g)\) is again a bijection from \(G \setminus \{e, h\}\) to itself. In the present paper, we answer the above question in the affirmative and we prove the
more general result that every abelian group which has a cyclic Sylow \(2\)-subgroup of order greater than \(6\) has such a partial bijection.

A directed triple system of order \(v\) and index \(\lambda\),denoted \({DTS}_\lambda(v)\), is said to be reverse if it admits an automorphism consisting of \(v/2\) transpositions when \(v\) is even, or a fixed point and \((v-1)/2\) transpositions when \(v\) is odd. We give necessary and sufficient conditions for the existence of a reverse \({DTS}_\lambda(v)\) for all \(\lambda \geq 1\).

A \(1\)-\emph{factor} of a graph \(G\) is a \(1\)-regular spanning subgraph of \(G\).A graph \(G\) has exactly \(t\) \(1\)-factors if the maximum set of edge-disjoint \(1\)-factors is \(t\). For given non-negative integers \(d\), \(t\), and even \(e\), let \(\mathcal{G}(2n; d, e, t)\) be the class of simple connected graphs on \(2n\) vertices, \((2n-1)\) of which have degree \(d\) and one has degree \(d+e\),having exactly \(t\) \(1\)-factors. The problem that arises is that of determining when \(\mathcal{G}(2n; d, e, t) \neq \emptyset ?\) Recently, we resolved the case \(t = 0\). In this paper, we will consider the case \(t = 1\).

In this paper we show that the complete graph \(K_{12}\) is not decomposable into three factors of diameter two, thus resolving a longstanding open problem. This result completes the solution of decomposition of a complete graph into three
factors, one of which has diameter two and the other factors have finite diameters.

The edge-integrity of a graph measures the difficulty of breaking it into pieces through the removal of a set of edges, taking into account both the number of edges removed and the size of the largest surviving component. We develop some techniques for bounding, estimating and computing the edge-integrity of products of graphs, paying particular attention to grid graphs.

We describe an algorithm for computing the number \(h_{m,n}\) of Hamiltonian circuits in a rectangular grid graph consisting of \(m \times n\) squares. For any fixed \(m\), the set of Hamiltonian circuits on such graphs (for varying \(n\)) can be identified via an appropriate coding with the words of a certain regular language \(L_m \subset (\{0,1\}^m)^*\). We show how to systematically construct a finite automaton \(A_m\) recognizing \(L_m\). This allows, in principle, the computation of the (rational) generating function \(h_m(z)\) of \(L_m\). We exhibit a bijection between the states of \(A_m\) and the words of length \(m+1\) of the familiar Motzkin language. This yields an upper bound on the degree of the denominator polynomial of \(h_m(z)\), hence of the order of the linear recurrence satisfied by the coefficients \(h_{m,n}\) for fixed \(m\).

The method described here has been implemented. Numerical data resulting from this implementation are presented at the end of this article.