The Existence of (ν,6, λ)-Perfect Mendelsohn Designs with λ > 1

The basic necessary conditions for the existence of a (v, k, λ)-perfect Mendelsohn design (briefly (v, k, λ)-PMD) are v ≥ k and λ v(v − 1) ≡ 0 (mod k). These conditions are known to be sufficient in most cases, but certainly not in all. For k = 3, 4, 5, 7, very extensive investigations of (v, k, λ)-PMDs have resulted in some fairly conclusive results. However, for k = 6 the results have been far from conclusive, especially for the case of λ = 1, which was given some attention in papers by Miao and Zhu [34], and subsequently by Abel et al. [1]. Here we investigate the situation for k = 6 and λ > 1. We find that the necessary conditions, namely v ≥ 6 and λ v(v − 1)≡0 (mod 6) are sufficient except for the known impossible cases v = 6 and either λ = 2 or λ odd. Researcher F.E. Bennett supported by NSERC Grant OGP 0005320.     

The spectrum of resolvable holey Mendelsohn triple systems and holey Mendelsohn frames with block size three

A holey Mendelsohn triple system (HMTS) is a decomposition of a complete multipartite directed graph into directed cycles of length 3. If the directed cycles of length 3 can be partitioned into parallel classes, then the HMTS is called an RHMTS. Bennett, Wei and Zhu [J. Combin. Des., 1997] showed that an RHMTS of type $g^n$ exists when $gn\equiv 0(\mathrm{mod}3)$ and $(g,n)\ne (1,6)$ with some possible exceptions. In this paper, motivated by the application in constructing RHMTSs, we investigate the constructions of holey Mendelsohn frames. We prove that a 3-MHF of type $(n,h^t)$ exists if and only if $n\ge 3$, $t\ge 4$ and $nh(t-1)\equiv 0(\mathrm{mod}3)$, and then determine that the necessary condition for the existence of an RHMTS of type $g^n$, namely, $gn\equiv 0(\mathrm{mod}3)$ is also sufficient except for $(g,n)=(1,6)$. New recursive constructions on incomplete RHMTSs via MHFs are introduced to settle this problem completely.     

Existence of (2, 8) GWhD( v) and (4, 8) GWhD( v) with $${ v \equi v 0,1 (mod 8)}$$

(2, 8) Generalized Whist tournament Designs (GWhD) on v players exist only if . We establish that these necessary conditions are sufficient for all but a relatively small number of (possibly) exceptional cases. For there are at most 12 possible exceptions: {177, 249, 305, 377, 385, 465, 473, 489, 497, 537, 553, 897}. For there are at most 98 possible exceptions the largest of which is v = 3696. The materials in this paper also enable us to obtain four previously unknown (4, 8)GWhD(8n+1), namely for n = 16,60,191,192 and to reduce the list of unknown (4, 8) GWhD(8n) to 124 values of v the largest of which is v = 3696.      

Uniformly resolvable designs with index one, block sizes three and five and up to five parallel classes with blocks of size five

Each parallel class of a uniformly resolvable design (URD) contains blocks of only one block size k (denoted k-pc). The number of k-pcs is denoted r_k. The necessary conditions for URDs with v points, index one, blocks of size 3 and 5, and r₃,r₅>0, are . If r_k>1, then v≥k², and r₃=(v−1−4⋅r₅)/2. For r₅=1 these URDs are known as group divisible designs. We prove that these necessary conditions are sufficient for r₅=3 except possibly v=105, and for r₅=2,4,5 with possible exceptions (v=105,165,285,345) New labeled frames and labeled URDs, which give new URDs as ingredient designs for recursive constructions, are the key in the proofs.     

Existence of incomplete canonical Kirkman packing designs

For $u,v$ positive integers with $u\equiv v\equiv 4\left(mod6\right)$, let ICKPD$(u,v)$ denote a canonical Kirkman packing of order $u$ missing one of order $v$. In this paper, it is shown that the necessary condition for existence of an ICKPD$(u,v)$, namely $u\ge 3v+4$, is sufficient with a definite exception $(u,v)=(16,4)$, and except possibly when $v>76$, $v\equiv 4\left(mod12\right)$ and $u\in \{3v+4,3v+10\}$.     

Existence of (q,k, 1) difference families with q a prime power and k = 4, 5

The existence of a (q, k, 1) difference family in GF(q) has been completely solved for k = 3. For k = 4, 5 partial results have been given by Bose, Wilson, and Buratti. In this article, we continue the investigation and show that the necessary condition for the existence of a (q, k, 1) difference family in GF(q), i.e., q ≡ 1 (mod k(k − 1)) is also sufficient for k = 4, 5. For general k, Wilson's bound shows that a (q, k, 1) difference family in GF(q) exists whenever q ≡ 1 (mod k(k − 1)) and q > [k(k − 1)/2]^k(k−1). An improved bound on q is also presented. © 1999 John Wiley & Sons, Inc. J Combin Designs 7: 21–30, 1999     

Finite and -resolvability

A topological space is -resolvable if has disjoint dense subsets. In this paper, we prove that if is -resolvable for each positive integer , then is -resolvable.

     

The existence of HGDDs with block size four and its application to double frames

In this note, we complete the existence spectra of HGDDs with block size four as well as double frames with block size three.     

On equidistant binary codes of lengthn=4k+1 with distanced=2k

The aim of this paper is to provide a short proof of the main result (Theorem 2.12) of [3], using standard methods from the theory of combinatorial designs. This paper was submitted to Combinatorica at the request of the editors.     

Maximum uniformly resolvable designs with block sizes 2 and 4

A central question in design theory dating from Kirkman in 1850 has been the existence of resolvable block designs. In this paper we will concentrate on the case when the block size k=4. The necessary condition for a resolvable design to exist when k=4 is that v≡4mod12; this was proven sufficient in 1972 by Hanani, Ray-Chaudhuri and Wilson [H. Hanani, D.K. Ray-Chaudhuri, R.M. Wilson, On resolvable designs, Discrete Math. 3 (1972) 343-357]. A resolvable pairwise balanced design with each parallel class consisting of blocks which are all of the same size is called a uniformly resolvable design, a URD. The necessary condition for the existence of a URD with block sizes 2 and 4 is that v≡0mod4. Obviously in a URD with blocks of size 2 and 4 one wishes to have the maximum number of resolution classes of blocks of size 4; these designs are called maximum uniformly resolvable designs or MURDs. So the question of the existence of a MURD on v points has been solved for by the result of Hanani, Ray-Chaudhuri and Wilson cited above. In the case this problem has essentially been solved with a handful of exceptions (see [G. Ge, A.C.H. Ling, Asymptotic results on the existence of 4-RGDDs and uniform 5-GDDs, J. Combin. Des. 13 (2005) 222-237]). In this paper we consider the case when and prove that a exists for all u≥2 with the possible exception of u∈{2,7,9,10,11,13,14,17,19,22,31,34,38,43,46,47,82}.     

Embeddings of resolvable group divisible designs with block size 3

It is proved in this paper that an RGD(3, g;v) can be embedded in an RGD(3, g;u) if and only if , , , v ≥ 3g, u ≥ 3v, and (g,v) ≠ (2,6),(2,12),(6,18).     

Abelian <Emphasis Type="Italic">ℓ</Emphasis>-Groups with Strong Unit and Perfect MV-Algebras

We investigate the class of abelian ℓ-groups with strong unit corresponding to perfect MV-algebras via the Γ functor, showing that this is a universal subclass of the class of all abelian ℓ-groups with strong unit and describing the formulas that axiomatize it. We further describe results for classes of abelian ℓ-groups with strong unit corresponding to local MV-algebras with finite rank.     

On perfect <Emphasis Type="Italic">p</Emphasis>-ary codes of length <Emphasis Type="Italic">p</Emphasis> + 1

Let p be a prime number and assume p ≥ 5. We will use a result of L. Redéi to prove, that every perfect 1-error correcting code C of length p + 1 over an alphabet of cardinality p, such that C has a rank equal to p and a kernel of dimension p − 2, will be equivalent to some Hamming code H. Further, C can be obtained from H, by the permutation of the symbols, in just one coordinate position.      

Uniformly resolvable designs with index one and block sizes three and four — with three or five parallel classes of block size four

Each parallel class of a uniformly resolvable design (URD) contains blocks of only one block size. A URD with v points and with block sizes three and four means that at least one parallel class has block size three and at least one has block size four. Danziger [P. Danziger, Uniform restricted resolvable designs with r=3, ARS Combin. 46 (1997) 161-176] proved that for all there exist URDs with index one, some parallel classes of block size three, and exactly three parallel classes with block size four, except when v=12 and except possibly when . We extend Danziger’s work by showing that there exists a URD with index one, some parallel classes with block size three, and exactly three parallel classes with block size four if, and only if, , v≠12. We also prove that there exists a URD with index one, some parallel classes of block size three, and exactly five parallel classes with block size four if, and only if, , v≠12. New labeled URDs, which give new URDs as ingredient designs for recursive constructions, are the key in the proofs. Some ingredient URDs are also constructed with difference families.     

Construction of BIB designs with various support sizes—with special emphasis for v = 8 and k = 4

This paper deals with methods of constructing new BIB designs from known designs. First, we provide a method of finding trades on a given BIB design so that new designs with the same parameters ν, b, k can be constructed. Second, we point out that the method of combining designs mentioned in [9] is particularly useful for constructing designs based on ν = 2k, k even, when designs based on ν = 2k ? 1 are available. By utilizing the above methods with the aid of Table I in [4], we have constructed a table of 53 BIB designs based on V = 8 and k = 4 with support sizes $\text{14 ? b 1 (}\text{8}\text{4}\text{)=70}$ except for $\text{b}{}^1\text{= 15, 16, 17, 19}$. BIB designs with ν = 8, k = 4 and $\text{b}{}^1\text{< 14}$ are shown to be nonexistent.     

The metric of anti-Thorpe spaces, n = 4k, k > 1

We show that in dimension 8, a semiflat metric is flat and that in dimension (8k+4) higher than 8, a semiflat metric does not necessarily imply flat. This revised version was published online in June 2006 with corrections to the Cover Date.     

Kite-group Divisible Designs of Type g t u 1

Necessary and sufficient conditions are given to the existence for kite-group divisible designs of type g^tu¹. Research supported by NSFC Grant 10371002.