Average-case results on heapsort

A new algorithm for rearranging a heap is presented and analysed in the average case. The average case upper bound for deleting the maximum element of a random heap is improved, and is shown to be less than [logn]+0.299+M(n) comparisons, *) whereM(n) is between 0 and 1. It is also shown that a heap can be constructed using 1.650n+O(logn) comparisons with this algorithm, the best result for any algorithm which does not use any extra space. The expected time to sortn elements is argued to be less thann logn+0.670n+O(logn), while simulation result points at an average case ofn log n+0.4n which will make it the fastest in-place sorting algorithm. The same technique is used to show that the average number of comparisons when deleting the maximum element of a heap using Williams' algorithm for rearrangement is 2([logn]–1.299+L(n)) whereL(n) also is between 0 and 1, and the average cost for Floyd-Williams Heapsort is at least 2nlogn–3.27n, counting only comparisons. An analysis of the number of interchanges when deleting the maximum element of a random heap, which is the same for both algorithms, is also presented.     

Covering the vertices of a graph by cycles of prescribed length

The main theorem of that paper is the following: let G be a graph of order n, of size at least (n² - 3n + 6)/2. For any integers k, n₁, n₂,…,n_k such that n = n₁ + n₂ +. + n_k and n_i ? 3, there exists a covering of the vertices of G by disjoint cycles (C_i) =l…k with |C_i| = ni, except when n = 6, n₁ = 3, n₂ = 3, and G is isomorphic to G₁, the complement of G₁ consisting of a C₃ and a stable set of three vertices, or when n = 9, n₁ = n₂ = n₃ = 3, and G is isomorphic to G₂, the complement of G₂ consisting of a complete graph on four vertices and a stable set of five vertices. We prove an analogous theorem for bipartite graphs: let G be a bipartite balanced graph of order 2n, of size at least n² - n + 2. For any integers s, n₁, n₂,…,n_s with n_i ? 2 and n = n₁ + n₂ + ? + n_s, there exists a covering of the vertices of G by s disjoint cycles C_i, with |C_i| = 2n_i.     

Multivalued groups, their representations and Hopf algebras

This paper introduces the concept ofn-valued groups and studies their algebraic and topological properties. We explore a number of examples. An important class consists of those that we calln-coset groups; they arise as orbit spaces of groupsG modulo a group of automorphisms withn elements. However, there are many examples that do not arise from this construction. We see that the theory ofn-valued groups is distinct from that of groups with a given automorphism group. There are natural concepts of the action of ann-valued group on a space and of a representation in an algebra of operators. We introduce the (purely algebraic) notion of ann-Hopf algebra and show that the ring of functions on ann-valued group and, in the topological case, the cohomology has ann-Hopf algebra structure. The cohomology algebra of the classifying space of a compact Lie group admits the structure of ann-Hopf algebra, wheren is the order of the Weyl group; the homology with dual structure is also ann-Hopf algebra. In general the group ring of ann-valued group is not ann-Hopf algebra but it is for ann-coset group constructed from an abelian group. Using the properties ofn-Hopf algebras we show that certain spaces do not admit the structure of ann-valued group and that certain commutativen-valued groups do not arise by applying then-coset construction to any commutative group.     

A Parallel Reduction of Hamiltonian Cycle to Hamiltonian Path in Tournaments

We propose a parallel algorithm which reduces the problem of computing Hamiltonian cycles in tournaments to the problem of computing Hamiltonian paths. The running time of our algorithm is O(log n) using O(n²/log n) processors on a CRCW PRAM, and O(log n log log n) on an EREW PRAM using O(n²/log n log log n) processors. As a corollary, we obtain a new parallel algorithm for computing Hamiltonian cycles in tournaments. This algorithm can be implemented in time O(log n) using O(n²/log n) processors in the CRCW model and in time O(log²n) with O(n²/log n log log n) processors in the EREW model.     

Small quasimultiple affine and projective planes: Some improved bounds

We improve the known bounds on r(n): = min {λ| an (n², n, λ)-RBIBD exists} in the case where n + 1 is a prime power. In such a case r(n) is proved to be at most n + 1. If, in addition, n − 1 is the product of twin prime powers, then r(n) ${\ \le \ }{n \over 2}$. We also improve the known bounds on p(n): = min{λ| an (n² + n + 1, n + 1, λ)-BIBD exists} in the case where n² + n + 1 is a prime power. In such a case p(n) is bounded at worst by but better bounds could be obtained exploiting the multiplicative structure of GF(n² + n + 1). Finally, we present an unpublished construction by M. Greig giving a quasidouble affine plane of order n for every positive integer n such that n − 1 and n + 1 are prime powers. © 1998 John Wiley & Sons, Inc. J Combin Designs 6: 337–345, 1998     

Fixed Points of Automorphisms of Graphs with 1 – Factorizations

We consider sets of fixed points of finite simple undirected connected graphs with 1 – factorizations. The maximum number of fixed points of complete graphs K_2n (n > 2) is n if n ≡ 0 mod 4, n — 1 if n ≡ 3 mod 4 or n ≡ 5 mod 12, n — 2 if n ≡ 2 mod 4, n — 3 if n ≡ 1 mod 4 and n ≢ 5 (mod 12). The maximum number of fixed points of 1 – factorizations of (non – complete) graphs with 2n vertices is less than or equal to n. If n is a prime number, then there are graphs with 2n vertices whose 1 – factorizations have automorphisms with n fixed points. Moreover, a result on the structure of a group of fixed – point – free automorphisms is presented.     

Steiner centers in graphs

The Steiner distance of a set S of vertices in a connected graph G is the minimum size among all connected subgraphs of G containing S. For n ≥ 2, the n-eccentricity e_n(ν) of a vertex ν of a graph G is the maximum Steiner distance among all sets S of n vertices of G that contains ν. The n-diameter of G is the maximum n-eccentricity among the vertices of G while the n-radius of G is the minimum n-eccentricity. The n-center of G is the subgraph induced by those vertices of G having minimum n-eccentricity. It is shown that every graph is the n-center of some graph. Several results on the n-center of a tree are established. In particular, it is shown that the n-center of a tree is a tree and those trees that are n-centers of trees are characterized.     

The Arcsine Law

Let N _n denote the number of positive sums in the first n trials in a random walk (S _i) and let L _n denote the first time we obtain the maximum in S ₀,..., S _n. Then the classical equivalence principle states that N _n and L _n have the same distribution and the classical arcsine law gives necessary and sufficient condition for (1/n) L _n or (1/n) N _n to converge in law to the arcsine distribution. The objective of this note is to provide a simple and elementary proof of the arcsine law for a general class of integer valued random variables (T _n) and to provide a simple an elementary proof of the equivalence principle for a general class of integer valued random vectors (N _n, L _n).     

Enclosing a Set of Objects by Two Minimum Area Rectangles

In this paper, we face the problem of computing an enclosing pair of axis-parallel rectangles of a set of polygonal objects in the plane, serving as a simple container. We propose anO(nα(n)log n) worst-case time algorithm, where α( ) is the inverse Ackermann's function, for finding, given a setMof points, segments and polygons defined bynvertices, a pair of axis-parallel rectangles (s, t) such thats∪tencloses all objects inMand area(s)+area(t) is minimum. The algorithm works inO(nα(n) log log n) worst-case space. Moreover, we prove an Ω(n log n) lower bound for the one-dimensional version of the problem. We also show that for the special case of enclosing a set of polygons with axis-parallel sides, our algorithm runs in optimal worst-case timeO(n log n), using worst-case spaceO(n log log n).     

Fully invariant transformations and associated groups

It is well known that the symmetric group S _ntogether with one idempotent of rank n- 1 on a finite n-element set Nserves as a set of generators for the semigroup T _nof all the total transformations on N. It is also well known that the singular part Sing _n of T _n can be generated by a set of idempotents of rank n- 1. The purpose of this paper is to begin an investigation of the way in which Sing_nand its subsemigroups can be generated by the conjugates of a subset of elements of T _n by a subgroup of S _n . We look for the smallest subset of elements of T _n that will serve and, correspondingly, for a characterization of those subgroups of S _n that will serve. Using some techniques from graph theory we prove our main result:the conjugates of a single transformation of rank n- 1 under Gsuffice to generate Sing_nif and only if Gis what we define to be a 2-block transitive subgroup of S _n .     

The graphical relaxation: A new framework for the symmetric traveling salesman polytope

A present trend in the study of theSymmetric Traveling Salesman Polytope (STSP(n)) is to use, as a relaxation of the polytope, thegraphical relaxation (GTSP(n)) rather than the traditionalmonotone relaxation which seems to have attained its limits. In this paper, we show the very close relationship between STSP(n) and GTSP(n). In particular, we prove that every non-trivial facet of STSP(n) is the intersection ofn + 1 facets of GTSP(n),n of which are defined by the degree inequalities. This fact permits us to define a standard form for the facet-defining inequalities for STSP(n), that we calltight triangular, and to devise a proof technique that can be used to show that many known facet-defining inequalities for GTSP(n) define also facets of STSP(n). In addition, we give conditions that permit to obtain facet-defining inequalities by composition of facet-defining inequalities for STSP(n) and general lifting theorems to derive facet-defining inequalities for STSP(n +k) from inequalities defining facets of STSP(n).Partially financed by P.R.C. Mathématique et Informatique.     

Latin Squares without Orthogonal Mates

In 1779 Euler proved that for every even n there exists a latin square of order n that has no orthogonal mate, and in 1944 Mann proved that for every n of the form 4k + 1, k ≥ 1, there exists a latin square of order n that has no orthogonal mate. Except for the two smallest cases, n = 3 and n = 7, it is not known whether a latin square of order n = 4k + 3 with no orthogonal mate exists or not. We complete the determination of all n for which there exists a mate-less latin square of order n by proving that, with the exception of n = 3, for all n = 4k + 3 there exists a latin square of order n with no orthogonal mate. We will also show how the methods used in this paper can be applied more generally by deriving several earlier non-orthogonality results.     

Cluster tilting for higher Auslander algebras

The concept of cluster tilting gives a higher analogue of classical Auslander correspondence between representation-finite algebras and Auslander algebras. The n-Auslander–Reiten translation functor τn plays an important role in the study of n-cluster tilting subcategories. We study the category Mn of preinjective-like modules obtained by applying τn to injective modules repeatedly. We call a finite-dimensional algebra Λ n-complete if for an n-cluster tilting object M. Our main result asserts that the endomorphism algebra EndΛ(M) is (n+1)-complete. This gives an inductive construction of n-complete algebras. For example, any representation-finite hereditary algebra Λ(1) is 1-complete. Hence the Auslander algebra Λ(2) of Λ(1) is 2-complete. Moreover, for any n?1, we have an n-complete algebra Λ(n) which has an n-cluster tilting object M(n) such that Λ(n+1)=EndΛ(n)(M(n)). We give the presentation of Λ(n) by a quiver with relations. We apply our results to construct n-cluster tilting subcategories of derived categories of n-complete algebras.     

On the construction of odd cycle systems

Three obvious necessary conditions for the existence of a k-cycle system of order n are that if n > 1 then n ? k, n is odd, and 2k divides n(n ? 1). We show that if these necessary conditions are sufficient for all n satisfying k ? n < 3k then they are sufficient for all n. In particular, there exists a 15-cycle system of order n if and only if n ≡ 1, 15, 21, or 25 (mod 30), and there exists a 21-cycle system of order n if and only if n ≡ 1, 7, 15, or 21 (mod 42), n ≠ 7. 15.     

On digraphs without antidirected cycles

Let t(n) denote the greatest number of arcs in a diagraph of orders n which does not contain any antidrected cycles. We show that [16/5(n ? 1)] ≤ t(n) ≤ 1/2 (n ? 1) for n ≥ 5. Let t_r (n) denote the corresponding quantity for r-colorable digraphs. We show that [16/5(n ? 1)] ≤ t₅(n) ≤ t₆(n) ≤ 10/3(n ? 1) for n ≥ 5 and that t₄(n) = 3(n ? 1) for n ≥ 3.     

Resolvable Coverings of 2-Paths by Cycles

 Let K _n be the complete graph on n vertices. A C(n,k,λ) design is a multiset of k-cycles in K _n in which each 2-path (path of length 2) of K _n occurs exactly λ times. A C(lk,k,1) design is resolvable if its k-cycles can be partitioned into classes so that every vertex appears exactly once in each class. A C(n,n,1) design gives a solution of Dudeney's round table problem. It is known that there exists a C(n,n,1) design when n is even and there exists a C(n,n,2) design when n is odd. In general the problem of constructing a C(n,n,1) design is still open when n is odd. Necessary and sufficient conditions for the existence of C(n,k,λ) designs and resolvable C(lk,k,1) designs are known when k=3,4. In this paper, we construct a resolvable C(n,k,1) design when n=p ^e +1 ( p is a prime number and e≥1) and k is any divisor of n with k≠1,2. Received: October, 2001 Final version received: September 4, 2002 RID="*" ID="*" This research was supported in part by Grant-in-Aid for Scientific Research (C) Japan     

Largest 4-connected components of 3-connected planar triangulations

Let T_n be a 3-connected n-vertex planar triangulation chosen uniformly at random. Then the number of vertices in the largest 4-connected component of T_n is asymptotic to n/2 with probability tending to 1 as n → ∞. It follows that almost all 3-connected triangulations with n vertices have a cycle of length at least n/2 + o(n).     

The structure and representation of <Emphasis Type="Italic">n</Emphasis>-ary algebras of DNA recombination

In this paper we investigate the structure and representation of n-ary algebras arising from DNA recombination, where n is a number of DNA segments participating in recombination. Our methods involve a generalization of the Jordan formalization of observables in quantum mechanics in n-ary splicing algebras. It is proved that every identity satisfied by n-ary DNA recombination, with no restriction on the degree, is a consequence of n-ary commutativity and a single n-ary identity of the degree 3n-2. It solves the well-known open problem in the theory of n-ary intermolecular recombination.     

An intermediate value theorem for graphs with given automorphism group

For a positive integer n and a finite group G, let the symbols e(G, n) and E(G, n) denote, respectively, the smallest and the greatest number of lines among all n-point graphs with automorphism group G. We say that the Intermediate Value Theorem (IVT) holds for G and n, if for each e satisfying e(G, n)≤e≤E(G, n), there exists an n-point graph with group G and e lines. The main result of this paper states that for every group G the IVT holds for all sufficiently large n. We also prove that the IVT holds for the identity group and all n, and exhibit examples of groups for which the IVT fails to hold for small values of n.     

On graphs preserving rectilinear shortest paths in the presence of obstacles

In physical VLSI design, network design (wiring) is the most time-consuming phase. For solving global wiring problems, we propose to first compute from the layout geometry a graph that preserves all shortest paths between pairs of relevant points, and then to operate on that graph for computing shortest paths, Steiner minimal tree approximations, or the like. For a set of points and a set of simple orthogonal polygons as obstacles in the plane, withn input points (polygon corner or other) altogether, we show how a shortest paths preserving graph of sizeO(n logn) can be computed in timeO(n logn) in the worst case, with spaceO(n). We illustrate the merits of this approach with a simple example: If the length of a longest edge in the graph is bounded by a polynomial inn, an assumption that is clearly fulfilled for graphs derived from VLSI layout geometries, then a shortest path can be computed in timeO(n logn log logn) in the worst case; this result improves on the known best one ofO(n(logn)^3/2).