ArticlePDF Available

Abstract

A $2-(v,k,\lambda)$ directed design (or simply a $2-(v,k,\lambda)DD$) is super-simple if its underlying $2-(v,k,2\lambda)BIBD$ is super-simple, that is, any two blocks of the $BIBD$ intersect in at most two points. A $2-(v,k,\lambda)DD$ is simple if its underlying $2-(v,k,2\lambda)BIBD$ is simple, that is, it has no repeated blocks. A set of blocks which is a subset of a unique $2-(v,k,\lambda)DD$ is said to be a defining set of the directed design. A smallest defining set, is a defining set which has smallest cardinality. In this paper simultaneously we show that the necessary and sufficient condition for the existence of a super-simple $2-(v,4,1)DD$ is $v\equiv1\ ({\rm mod}\ 3)$ and for these values except $v=7$, there exists a super-simple $2-(v,4,1)DD$ whose smallest defining sets have at least a half of the blocks. And also for all $\epsilon > 0$ there exists $v_0(\epsilon)$ such that for all admissible $v>v_0$ there exists a $2-(v,4,1)DD$ whose smallest defining sets have at least $(5/8-\frac{c}{v})\mid \mathcal{B}\mid$ blocks, for suitable positive constant c.
Smallest defining sets of super-simple
2(v, 4,1) directed designs
Nasrin SoltankhahAND Farzane Amirzade
Department of Mathematics
Alzahra University
Vanak Square 19834 Tehran, I.R. Iran
Abstract
A 2 (v, k, λ) directed design (or simply a 2 (v, k, λ)DD) is
super-simple if its underlying 2 (v, k, 2λ)B IB D is super-simple,
that is, any two blocks of the BIBD intersect in at most two points.
A 2 (v, k , λ)DD is simple if its underlying 2 (v, k, 2λ)B IB D is
simple, that is, it has no repeated blocks.
A set of blocks which is a subset of a unique 2(v, k , λ)DD is said
to be a defining set of the directed design. A smallest defining set, is
a defining set which has smallest cardinality. In this paper simulta-
neously we show that the necessary and sufficient condition for the
existence of a super-simple 2 (v, 4,1)DD is v1 (mod 3) and for
these values except v= 7 , there exists a super-simple 2(v, 4,1)DD
whose smallest defining sets have at least a half of the blocks. And
also for all ² > 0 there exists v0(²) such that for all admissible v > v0
there exists a 2 (v, 4,1)D D whose smallest defining sets have at
least ( 5
8c
v)| B | blocks, for suitable positive constant c.
KEYWORDS: super simple directed designs, smallest defining set
1 Introduction
At(v, k, λ) directed design (or simply a t(v , k, λ)DD) is a pair (V, B),
where Vis v-set, and Bis a collection of ordered ktuples of distinct
Corresponding author. E-mail address: soltan@alzahra.ac.ir
1
PLEASE CITE AS: F. Amirzade and N. Soltankhah. Smallest defining sets of
super-simple 2- (v,4,1) directed designs. Utilitas Mathematica, 96(2015), 331-344.
elements of V(called blocks), such that each ordered t-tuple of distinct
elements of Vappears in precisely λblocks. We say that a t-tuple appears
in a k-tuple, if its components appear in that k-tuple as a set, and they
appear with the same order. For example the following 4-tuples form a
2(10,4,1)DD, D.
0132 0467 0589 2680 2791 7843 6942 4510
1498 3861 3970 8752 9653 2354 1576.
Here, for example, the 4-tuple 0132 contains the ordered pairs 01, 03, 02,
13, 12, and 32.
A (v, k, t) directed trade of volume sconsists of two disjoint collections
T1and T2, each of sblocks, such that every t-tuple of distinct elements of
Vis covered by precisely the same number of blocks of T1as of T2. Such
a directed trade is usually denoted by T=T1T2. Blocks in T1(T2) are
called the positive (respectively, negative) blocks of T. If D= (V, B) is a
directed design, and if T1⊆ B, we say that Dcontains the directed trade T.
For example the 2 (10,4,1)DD, Dabove contains the following directed
trade: T1T2
0132 1032
4510 4501
Given a t(v, k, λ) directed design D, a subset of the blocks of Dthat
occurs in no other t(v, k , λ) directed design is called a defining set of
D. A defining set for which no other defining set has a smaller cardinal-
ity, is called a smallest defining set. For example the set of blocks S=
{0132,2354,9653,8752,2791,2680,1498,1576}is a defining set of the 2
(10,4,1)DD above. But the set of blocks R={0132,2354,9653,8752,2791,
2680,1498,0467}can be completed to a 2(10,4,1)DD in two ways: by ad-
joining either {4510,1576,0589,3861,3970,7843,6942}or {4150,5176,0589,
3861,3970,7843,6942}.
Defining sets for directed designs (as suggested by A. P. Street in [12])
are strongly related to trades. This relation is illustrated by the following
result.
PROPOSITION 1.1 [14] Let D= (V, B)be a t(v, k, λ)directed design
and let S⊆ B, then Sis a defining set of Dif only if Scontains a block of
every (v, k, t)directed trade T=T1T2such that Tis contained in D.
Each defining set of a t(v, k, λ)DD , Dcontains at least one block in every
trade in D. In particular, if Dcontains mmutually disjoint directed trades
then the smallest defining set of Dmust contain at least mblocks.
2
The concept of directed trades and defining sets for directed designs
were investigated in articles [14,16].
A 2 (v, k, λ) directed design (or simply a 2 (v, k, λ)DD) is super-
simple if its underlying 2 (v , k, 2λ)BIBD is super-simple, that is, any
two blocks of the BIBD intersect in at most two points. A 2 (v, k, λ)DD
is simple if its underlying 2 (v, k , 2λ)BIBD is simple, that is, it has no
repeated blocks.
The concept of super-simple designs was introduced by Gronau and
Mullin [10]. The existence of super-simple designs is an interesting problem
by itself, but there are also some useful applications.
There are known results for the existence of super simple designs, es-
pecially for existence of super simple 2 (v, k, λ)BIBDs. When k= 4 the
necessary conditions for super simple 2 (v, k , λ)BIBDs are known to be
sufficient for 2 λ6 with few possible exceptions. These known results
can be found in articles [1,3,4,5,6,9,11].
In [14], Mahmoodian, Soltankhah and Street have proved that if Dbe
a 2 (v, 3,1)DD then a defining set of Dhas at least v
2blocks. In [8]
Grannel, Griggs and Quinn have shown that for each admissible value of v,
there exists a 2 (v, 3,1)DD and a simple 2 (v, 3,1)DD whose smallest
defining sets have at least |B|
2blocks. And they have also proved that, for
all ² > 0 and all sufficiently large admissible v, there exists a 2(v , 3,1)DD
whose smallest defining sets have at least (2
3²)| B | blocks.
In this paper we show that for all admissible values, there is a super-
simple 2 (v, 4,1)DD (except v= 7) whose smallest defining sets have at
least a half of the blocks.
In other words we are interested in the quantity
f=number of 4tuples in a smallest defining set in D
number of 4tuples in D.
And we show that, there exists a super-simple 2 (v, 4,1)DD,Dwith
f1
2.
The proofs in this paper use various types of combinatorial objects. The
definitions of these objects are either given in the paper or can be found in
the references.
Several proofs depend on the following result, which involves pairwise
balanced designs (PBDs) and is a special case of a result (the Replacement
Lemma [14]) that is used in several earlier papers on directed designs.
LEMMA 1.1 If there exist a 2(v, K, 1) design and a super-simple (k, 4,1)DD
3
for each kK, then there exists a super-simple 2(v, 4,1)DD.
PROOF. Replacing each block of the 2 (v, K, 1) design with a copy of
a super-simple 2 (k, 4,1)DD with point set the points of that block gives
a super-simple 2 (v, 4,1)DD.
A lower bound for ffor the super-simple 2 (v , 4,1)DD constructed in
Lemma 1.1 can be calculated from lower bounds for ffor the various super-
simple 2 (k, 4,1)DDs. In particular, if there is a constant c such that
each of the super-simple 2 (k, 4,1)DDs has fc, then the resulting
super-simple 2 (v, 4,1)DD also has fc.
2 Main result
A necessary and sufficient condition for the existence of a 2 (v, 4,1)DD
is v1 (mod 3) [17].
In this section simultaneously we show that the necessary and sufficient
condition for the existence of a super-simple 2(v , 4,1)DD is v1 (mod 3)
and for these values except v= 7, there exists a super-simple 2(v , 4,1)DD
with f1
2.
In [13], Soltankhah and Mahmoodian have shown that, up to isomor-
phism, there exist two 2(7,4,1)DDs that they are super-simple and each
of them has smallest defining set of cardinality 2. Consequently, if v= 7
then f=2
7.
In proofs we handle with a special type of directed trade, named a
cyclical trade, as follows.
Let T=T1T2be a (v, 4,2) directed trade of volume s, where T1con-
tains blocks b1, . . . , bssuch that each pair of consecutive 4-tuples (blocks)
of T1,bibi+1 i= 1, . . . , s (mod s) is a trade of volume 2. Therefore if a
directed design Dcontains T1, then any defining set for Dmust contain at
least [s+1
2] blocks of T1.
The following 4-tuples form a super-simple 2 (10,4,1)DD.
0467 0589 2680 2791 7843 6942 0132
1576 1498 3861 3970 8752 9653 4510
2354
Each of the first six columns above contains a trade, hence any defining set
for this directed design must contain at least one 4-tuple of each column.
4
The last column above is a cyclical trade of volume 3. Hence any defining
set for this directed design must contain at least two 4-tuples of the last col-
umn. We can use integer programming problem [2] to find a smallest defin-
ing set for this directed design and S={0132,2354,9653,8752,2791,2680,
1498,1576}is a defining set of size 8. Therefore this super-simple 2
(10,4,1)DD has f=8
15 >1
2.
The following base blocks by (+1 mod 13) form a super-simple 2
(13,4,1)DD with f1
2.
0139,1 0 11 5
The following 4-tuples form a super-simple 2 (16,4,1)DD.
1 3 2 4 1 5 10 14 1 6 12 13 3 8 10 15 1 8 11 16
16 14 3 1 13 9 5 1 15 10 6 1 14 10 8 9 16 11 10 2
5 6 7 8 2 6 11 15 2 5 9 16 4 7 11 14 2 7 10 13
16 13 7 6 14 11 6 5 15 5 2 3 10 7 4 5 8 7 2 1
9 10 12 11 3 7 12 16 2 8 12 14 1 7 9 15 3 6 9 14
13 12 10 3 16 12 8 5 15 14 12 7 11 9 7 3 12 9 6 2
13 14 15 16 4 9 8 13 3 5 11 13 4 6 10 16 5 4 12 15
14 13 4 2 16 15 9 4 15 13 11 8 8 6 4 3 11 12 4 1
This super-simple 2 (16,4,1)DD contains 20 disjoint directed trades of
volume 2. Hence any defining set of this super-simple directed design must
contain at least one 4-tuple of each trade. So for this super-simple 2
(16,4,1)DD we have f20
40 =1
2.
In our future results we need some super-simple directed group divisible
designs with block size four (super-simple 4 DGDDs). We may use a
super-simple 4 DGDD of type (34), a super-simple 4 DGDD of type
(35), a super-simple 4 DGDD of type (36) and a super-simple 4DGDD
of type (24). Now we can construct a super-simple 4 DGDD of type (34)
with groups {0,4,8},{1,5,9},{2,6,10},{3,7,11}, and with the following
blocks.
4 3 10 9 9 2 3 8 2 1 4 11 11 10 5 4 7 6 0 1
91007 11029 3416 4572 6794
10103 3205
8 1 2 7 5 0 6 11 7 8 5 10
10 11 1 8 8 9 11 6 6 5 8 3
5
Each of the first two columns above is a cyclical trade of volume 3, hence
any defining set of this super-simple directed group divisible design must
contain at least two 4-tuples from each of the first two columns and 4-tuples
in the last three columns above form six disjoint directed trades of volume
2 and one 4-tuple from each of them must be in any defining set. Therefore
any defining set of this super-simple 4 DGDD must contain at least 10
blocks, so it has f10
18 >1
2.
A super-simple 4 DGDD of type (35) can be constructed with groups
{0,5,10},{1,6,11},{2,7,12},{3,8,13},{4,9,14}
and with the following base blocks by (+1 mod 15).
1037,0 1 13 9
This super-simple directed group divisible design has 15 disjoint directed
trades of volume 2 and one 4-tuple from each of them must be in any
defining set. Therefore any defining set of this super-simple 4 DGDD
must contain at least 15 blocks, so it has f15
30 =1
2.
A super-simple 4 DGDD of type (36) can be constructed with groups
{0,6,12},{1,7,13},{2,8,14},{3,9,15},{4,10,16},{5,11,17}
and with the following base blocks by (+2 mod 18).
51020 ,14 1 10 5
15 0 4 11 ,11 4 3 1
0132
This super-simple directed group divisible design has 18 disjoint di-
rected trades of volume 2 in the first two rows above and the last row
is a cyclical trade of volume 9. So any defining set for this super-simple
4DGDD must contain at least five 4-tuples of the last row. Hence any
defining set must contain at least 18+5 = 23 blocks, so for this super-simple
4DGDD, we have f23
45 >1
2.
A super-simple 4 DGDD of type (24) with f1
2can be constructed
with groups {1,2},{3,4},{5,6},{7,8}and with the following blocks
4 1 6 7 ,8613
6 8 2 4 ,7542
5 7 3 1 ,2376
1 4 5 8 ,3285.
6
Our principal tool is to apply Wilson’s Fundamental construction(weighting),
that is described in the following lemma.
LEMMA 2.2 If there is a {K} − GDD of type g1u1g2u2. . . gN
uN, there
are super-simple 2(αgi+1,4,1)DDs for each i, i = 1,2, . . . , N and there
are super-simple 4DGDDs of type αkfor each kKthen, there exists
a super-simple 2(αPN
i=1 giui+ 1,4,1)DD.
PROOF. Let (G, B) be a group divisible design with element set U,
blocks of size kKand groups of size g1, g2, . . . , and gN. We form a
super-simple 2(αPN
i=1 giui+1,4,1)DD on the element set U×ZαS{∞}.
Give weight αto all of points. That is replace each point xUwith
αnew points {x1, x2, . . . , xα}. Now replace each block b∈ B of size
kKwith a super-simple 4 DGDD of type αksuch that its groups
are {x1, x2, . . . , xα:xU}to get a super-simple 4 DGDD of type
αg1u1αg2u2. . . αgN
uN. Finally fill in the holes with a new point , using
super-simple 2 (αgi+ 1,4,1)DDs for i= 1,2, . . . , N .
In particular, if there is a constant c such that each of the super-simple
4DGDD of type αkand the super-simple 2 (αgi+ 1,4,1)DDs has
fc, then the resulting super-simple 2 (v , 4,1)DD in Lemma 2.2 also
has fc.
THEOREM 2.1 For all v1 (mod 3) except v= 7 there exists a super-
simple 2(v, 4,1)DD with f1
2.
PROOF. For all nexcept nA={7,8,9,10,11,12,14,15,18,19,23}
there exists a P BD (n, {4,5,6}) [15]. We can delete one point from this
P BD to form a {4,5,6} − GDD of order n1 of type 3a4b5c, where a,
b, c are non-negative integers. We then apply lemma 2.2 to this GDD
using a weight of 3 to get a super-simple 4 DGDD of type 9a12b15cwith
f1
2. Finally fill in the holes with a new point , using super-simple
2(m, 4,1)DDs for m= 10,13,16 with f1
2.
Now for the remaining values v∈ {19,22,25,28,31,34,40,43,52,55,67}
we construct a super-simple 2 (v, 4,1)DD with f1
2as follows.
v= 19: the following 4-tuples form a super-simple 2 (19,4,1)DD.
7
3 6 4 15 2 7 4 17 10 3 11 13 11 16 5 12 14 12 17 8
6 3 16 1 5 4 7 16 9 3 10 14 13 15 12 5 6 17 12 13
5 8 17 6
6 11 0 7 17 9 4 5 3 9 12 7 10 12 15 6
11 6 8 2 16 9 17 0 1 8 7 12 12 10 9 2 2 1 14 6
13 4 1 2
4 0 18 12 13 16 7 11 8 1 9 11 2 9 13 8 18 14 1 4
18 0 2 11 16 13 6 14 9 1 15 16 15 13 9 18
0 5 1 13
18 7 9 6 7 14 18 13 8 18 16 15 8 4 13 10 2 5 0 15
0 4 6 9 17 14 7 15 2 12 16 18 16 4 8 3 15 0 8 14
7 10 8 0
12 1 3 0 5 14 11 9 15 11 17 1 14 0 10 16 10 7 1 5
13 0 3 17 12 4 11 14 11 15 10 4
1 18 10 17 14 5 3 2 17 11 3 18 17 16 2 10
6 5 10 18 15 7 2 3 18 3 8 5
This super-simple 2(19,4,1)DD has 22 disjoint directed trades of volume
2 in the first four columns above and the last column has three cyclical
trades of volume 3, 3 and 5, respectively. So any defining set for this DD
must contain at least two 4-tuples from each of the cyclical trades of volume
3 and three 4-tuples of the cyclical trade of volume 5. Hence any defining
set must contain at least 22+ 2×2 + 3 = 29 blocks, so for this super-simple
2(19,4,1)DD, we havef29
57 >1
2.
v= 22: the following base blocks by (,+1 mod 11) form a super-simple
2(22,4,1)DD.
(0,0)(0,3)(0,9)(0,10) ,(1,0)(1,4)(0,9)(0,3)
(0,0)(1,0)(1,7)(1,2) ,(1,0)(0,0)(1,9)(1,10)
(0,0)(0,2)(1,5)(1,8) ,(1,0)(1,5)(0,2)(0,6)
(1,7)(0,3)(0,0)(1,4)
This super-simple 2(22,4,1)DD has 33 disjoint directed trades of volume
2 in the first three rows above and the last row is a cyclical trade of volume
11. So any defining set for this DD must contain at least six 4-tuples of the
last row. Hence any defining set must contain at least 33 + 6 = 39 blocks,
so for this super-simple 2 (22,4,1)DD, we have f39
77 >1
2.
v= 25: given base blocks in [10] form a super-simple 2 (25,4,1)DD
with f1
2.
v= 28: the following 4-tuples form a super-simple 2 (28,4,1)DD.
8
1 2 14 16 1 8 22 23 2 3 19 22 6 8 9 14 4 7 14 15
21 9 2 1 25 4 8 1 25 14 3 2 27 8 6 3 23 21 15 14
12 14 20 24 23 22 9 3 13 14 19 26 3 6 13 17 2 6 21 23
16 14 12 10 10 9 22 24 11 14 13 9 27 25 17 13 11 10 6 2
6 7 10 12 8 4 21 24 7 13 8 11 7 16 25 27 10 11 23 25
24 15 10 7 25 24 21 5 22 19 11 8 17 16 7 4 23 11 7 1
2 8 10 15 1 3 24 25 2 7 18 24 16 17 20 22 1 7 9 17
17 9 10 8 18 10 3 1 12 8 7 2 20 17 14 1 26 9 7 5
6 15 24 27 0 3 10 14 9 11 21 26 14 17 21 25 5 7 19 20
27 24 16 9 24 11 3 0 27 21 11 4 26 21 17 6 20 19 9 4
4 3 9 16 1 6 20 26 5 6 22 25 3 12 21 27 13 16 23 24
26 15 3 4 22 16 6 1 19 14 6 5 19 17 12 3 24 23 17 2
0 2 12 17 1 4 11 12 2 4 25 26 11 17 19 24 3 5 11 15
20 16 2 0 26 20 12 11 26 25 22 10 17 11 18 5 16 13 5 3
4 6 18 19 0 5 24 26 3 8 18 20 18 15 17 23 0 4 20 23
24 12 6 4 26 24 14 8 24 22 20 18 22 17 15 0 25 23 20 6
0 8 19 25 5 8 12 16 2 11 20 27 4 13 22 27 18 14 4 0
25 19 18 7 23 8 5 0 27 20 10 5 27 22 14 7 14 18 11 22
8 17 26 27 3 7 23 26 0 1 13 15 10 13 20 21 9 15 20 25
27 26 1 0 26 23 19 16 24 19 13 1 21 20 7 3 20 15 8 13
15 18 9 6 2 5 9 13 6 0 11 16 12 15 22 26 13 7 0 6
0 9 18 27 22 4 5 2 25 15 16 11 15 12 5 1 0 7 21 22
16 15 19 21 12 13 18 25 1 10 19 27 5 14 23 27 5 4 10 17
21 19 10 0 22 21 13 12 27 19 15 2 27 23 18 12 23 13 10 4
1 5 18 21 10 16 18 26 25 12 9 0
21 18 16 8 26 18 13 2 9 12 19 23
This super-simple 2 (28,4,1)DD contains 63 disjoint directed trades.
Hence, any defining set of this super simple directed design must contain
at least one 4-tuple of each trade. So for this super-simple 2 (28,4,1)DD
we have f63
126 =1
2.
9
v= 31: the following base blocks by (+1 mod 31) form a super-simple
2(31,4,1)DD with f1
2.
2 0 4 1 ,1 4 19 9
4 0 20 13 ,13 20 8 30 ,14 28 8 20
This super-simple 2(31,4,1)DD has 31 disjoint directed trades of volume
2 in the first row above and the second row is a cyclical trade of volume 93.
So any defining set for this DD must contain at least 47, 4-tuples of the
cyclical trade. Hence any defining set must contain at least 31 + 47 = 78
blocks, so for this super-simple 2 (31,4,1)DD we have f78
155 >1
2.
v= 34: the following base blocks by (+1 mod 34) form a super-simple
2(34,4,1)DD with f1
2.
0 22 1 24 ,31 29 24 1
25 1 19 0 ,19 1 15 22 ,23 15 1 20
0 17 8 25
This super-simple 2 (34,4,1)DD has 34 disjoint directed trades of vol-
ume 2 in the first row above and the second row is a cyclical trade of
volume 102. So any defining set for this DD must contain at least 51,
4-tuples of the second row. The last row form a cyclical trade of volume
34. Hence any defining set must contain at least 34 + 51 + 17 = 102
blocks, so for this super-simple 2 (34,4,1)DD we have f102
204 1
2.
v= 40: begin with a 4 GDD(54) [15]. We apply lemma 2.2 using
a weight 2, using a super-simple 4 DGDD of type (24) with f1
2,
to get a super-simple 4 DGDD of type 104with f1
2. Now fill in
the holes of size 10 with a super-simple 2 (10,4,1)DD with f1
2.
v= 43: begin with a 4 GDD(27) [15]. We apply lemma 2.2 using a
weight 3, using a super-simple 4DGDD of type (34) with f5
9, to get a
super-simple 4DGDD of type 67with f1
2. Finally fill in the hole with a
new point , using a super-simple 2(7,4,1)DD. This yield a super-simple
2(43,4,1)DD with f154
301 >1
2.
v= 52: the following base blocks form a super-simple (52,4,1)DD with
f1
2.
0 5 1 3 ,48 5 0 45 (+1 mod 52)
30 6 0 13 ,13 0 23 42 (+1 mod 52)
0 6 21 37 ,4 37 21 45 (+1 mod 52)
8 38 20 0 ,34 9 0 20 (+1 mod 52)
27 1 0 26 ,26 0 25 51 (+2 mod 52)
10
v= 55: a RBIBD(16,4,1) has 5 parallel classes. For two parallel
classes add a new point to each of the blocks in that parallel class. This
yields a {4,5} − GDD of type 4421. We apply lemma 2.2 using a weight 3,
using a super-simple 4DGDD of type (34) with f5
9and a super-simple
4DGDD of type (35) with f1
2, to get a super-simple 4 DGDD of
type 12461with f1
2. Finally fill in the holes with a new point , using
a super-simple (13,4,1)DD with f1
2and a super-simple 2 (7,4,1)DD.
v= 67: begin with a 4 GDD(6491) [15]. Applying lemma 2.2 using a
weight 2, we obtain a super-simple 2 (67,4,1)DD with f1
2.
3 Asymptotic results
In previous section, we showed that for all admissible values of v, except
v= 7, there exists a super-simple 2(v, 4,1)DD with f1
2. In this section
we show that these results can be improved for 2 (v, 4,1)DDs. And we
prove that for all sufficiently large admissible v, there exists a 2(v, 4,1)DD
with f5
8c
vfor suitable positive constant c. For this result we need a
directed group divisible design with block size four (4 DGDD) of type
(24) with f5
8.
A 4DGDD of type (24) can be constructed with groups {1,2},{3,4},
{5,6},{7,0}, and with the following blocks.
blocks : 5 1 0 3 3 7 1 5 4 0 2 5
6 0 1 4 4 1 7 6 5 2 7 4
2 3 0 6 6 7 3 2
Each of first two columns above is a cyclical trade of volume 3, hence any
defining set of this directed group divisible design must contain at least two
4-tuples from each of the first two columns. The 4-tuples in the last column
above form a trade and one of them must be in any defining set. Therefore
any defining set of this 4DGDD must contain at least 5 blocks, so it has
f5
8.
LEMMA 3.3 For all v1 (mod 12),v61, there exists a 2(v, 4,1)DD
with f5
813
8v.
PROOF. For all k > 4 there exists a 4 GDD(6k) [15]. Let (G, B) be
a group divisible design with element set U, blocks of size 4 and groups of
11
size 6. We replace each group gGwith a 2 (13,4,1)DD with f1
2
on g×Z2S{∞} and each block b∈ B with 4 DGDD of type (24) with
f5
8on b×Z2, such that its groups are {x} × Z2,{y} × Z2,{z} × Z2and
{w} × Z2. Since a 4 GDD(6k) has 3k(k1) blocks, so a 2 (v , 4,1)DD
has
f(5
8.24k(k1) + 1
2.26k)/((12k+1)(12k)
6).
Simplifying gives f5
813
8v.
LEMMA 3.4 For all v4 (mod 12),v64, there exists a 2(v, 4,1)DD
with f5
81
2v.
PROOF. For all k > 4 there exists a 4 GDD(23k+1) [15]. Let (G, B)
be a group divisible design with element set U, blocks of size 4 and groups
of size 2. We replace each group gGwith a 2 (4,4,1)DD on g×Z2
and each block b∈ B with 4 DGDD of type (24) with f5
8on b×Z2,
such that its groups are {x} × Z2,{y} × Z2,{z} × Z2and {w} × Z2. Since
a 4 GDD(23k+1 ) has k(3k+ 1) blocks, so a 2 (v, 4,1)DD has
f(5
8.8k(3k+ 1) + 1
2.4(3k+ 1))/((12k+4)(12k+3)
6).
Simplifying gives f5
81
8(4k+1) >5
81
2v.
LEMMA 3.5 For all v7 (mod 12),v67, there exists a 2(v, 4,1)DD
with f5
813
8v21
2v(v+11) .
PROOF. For all k > 4 there exists a 4 GDD(6k91) [7]. Let (G, B) be
a group divisible design with element set U, blocks of size 4 and groups of
size 6 and a single group of size 9. We replace each group gGof size
6 and 9 with a 2 (13,4,1)DD and a 2 (19,4,1)DD with f1
2on
g×Z2S{∞}, respectively, and each block b∈ B with 4 DGDD of type
(24) with f5
8on b×Z2, such that its groups are {x} × Z2,{y} × Z2,
{z} × Z2and {w} × Z2. Since a 4 GDD(6k91) has 3k(k+ 2) blocks, so
a 2 (v, 4,1)DD has
f(5
8.24k(k+ 2) + 1
2.26k+ 29)/((12k+19)(12k+18)
6).
Simplifying gives f5
813
8v21
2v(v1) .
12
LEMMA 3.6 For all v10 (mod 12),v70 there exists a 2(v, 4,1)DD
with f5
81
4v1
2v.
PROOF. For all k > 4 there exists a 4 GDD(23k51) [7]. Let (G, B) be
a group divisible design with element set U, blocks of size 4 and groups of
size 2 and a single group of size 5. We replace each group gGof size 2
and 5 with a 2 (4,4,1)DD and a 2 (10,4,1)DD with f1
2on g×Z2,
respectively, and each block b∈ B with 4 DGDD of type (24) with f5
8
on b×Z2, such that its groups are {xZ2,{yZ2,{zZ2and {wZ2.
Since a 4 GDD(23k51) has k(3k+ 4) blocks, so a 2 (v , 4,1)DD has
f(5
8.8k(3k+ 4) + 1
2.4(3k) + 8)/((12k+10)(12k+9)
6).
Simplifying gives f5
81
4vk+4
2v(4k+3) or f > 5
81
2v.
THEOREM 3.2 For all ² > 0there exists v0(²)such that for all admis-
sible v > v0there exists a 2(v, 4,1)DD with f5
8c
vwhere c=13
8+².
References
[1] P. Adams, D. Bryant and A. Khodkar, On the existence of super-simple
designs with block size 4, Aequationes Math., 51 (1996), 230–246.
[2] P. Adams, A. Khodkar, C.Ramsay, Smallest defining sets of some
STS(19), J. Combin. Math. Combin. Comput., 37 (2001), 225–237.
[3] H. Cao, K. Chen and R. Wei, Super-simple balanced incomplete block
designs with block size 4 and index 5, Discrete Math., 309 (2009), 2808–
2814.
[4] K. Chen, On the existence of super-simple (v, 4,3)BIBDs, J. Combin.
Math. Combin. Comput., 17 (1995), 149–159.
[5] K. Chen, On the existence of super-simple (v, 4,4) BIBDs, J. Statis.
Plann. Inference, 51 (1996), 339–350.
[6] K. Chen, Z. Cao, and R. Wei, Super-simple balanced incomplete block
designs with block size 4 and index 6, J. Statis. Plann. Inference, 133
(2005), 537–554.
[7] G. Ge, A. C. H. Ling, Group divisible designs with block size four and
type gum1for small g, Discrete Math., 285 (2004), 97–120.
13
[8] M. J. Grannell, T. S. Griggs and K. A. S. Quinn, Smallest defining sets
of directed triple systems , Discrete Math., 309 (2009), 4810–4818.
[9] H. D. O. F. Gronau, Super-simple designs, in The CRC Handbook of
Combinatorial Designs, second edition (ed. C. J. Colbourn and J. H.
Dinitz), CRC Press, 2007, 633–635.
[10] H. D. O. F. Gronau and R. C. Mullin, On super-simple 2(v, 4, λ)
designs, J. Combin. Math. Combin. Comput., 11 (1992), 113–121.
[11] A. Khodkar, Various super-simple designs with block size four, Aus-
tralas. J. Combin., 9(1994), 201–210.
[12] E. S. Mahmoodian and M. A. Shokrollahi, Open problems at the com-
binatorics workshop of AIMC25 (Tehran, 1994), in combinatorics Ad-
vances, C. J. Colbourn and E. S. Mahmoodian, eds., Mathematics and
Its Applications, Kluwer Academic Publisher, Dordrecht, Boston, Lon-
don, (1995), 321–324.
[13] E. S. Mahmoodian and N. Soltankhah, Intersections of 2(v, 4,1)
Directed Designs, J. Combin. Math. and Combin. Comput. (JCMCC),
20 (1996), 225–236.
[14] E. S. Mahmoodian, N. Soltankhah and A. P. Street, On Defining Sets
of Directed Designs, Australas. J. Combin., 19 (1999), 179–190.
[15] R. C. Mullin and H. D. O. F. Gronau, PBDs and GDDs: the basics,
in The CRC Handbook of Combinatorial Designs, second edition (ed.
C. J. Colbourn and J. H. Dinitz), CRC Press, 2007.
[16] N. Soltankhah, On directed trades, Australas. J. Combin., 11 (1995),
59–66.
[17] D. J. Street and J. R. Seberry, All DBIBDs with block size four exist,
Utilitas Math., 18 (1980), 27–34.
14
... (see [1,20,22,23,25,29,39]), which are listed below. For more results on super-simple designs we refer the reader to [2,6,9,15,16,24,28,30,36,39,41] and references therein. ...
... A super-simple TD λ (4, v) exists if and only if λ ≤ v and (λ, v) is neither (1,2) nor (1,6). (5, v) exists for any (1) n ≥ 5; (2) n(n − 1)h 2 ≡ 0 (mod 5); (3) if h is odd, then n is odd. ...
... Lemma 2.6. ( [16]) There exists a super-simple (5, 4)-GDD of type g u if and only if u ≥ 5, g(u −2) ≥ 12 and u(u −1)g 2 ≡ 0 (mod 5) with some possible exceptions: (a) (g, u) = (3, 6), (b) g = 10, u ∈ {18, 23,24,27, 28} and (c) g = 15, u ∈ {12, 14, 18, 24, 28, 34}. We shall use the following basic recursive constructions in proofing Theorems 1.4-1.5. ...
Article
A design is said to be super-simple if the intersection of any two blocks has at most two elements. In statistical planning of experiments, super-simple designs provide samples with intersection as small as possible. Super-simple group divisible designs are also useful for constructing other types of super-simple designs, which can be applied to codes and designs. In this article, we investigate the existence of a super-simple (4,λ)-GDD of group type gu, where λ=7,8, and show that such a design exists if and only if u≥4, g(u−2)≥2λ, g(u−1)≡0(mod3) and λu(u−1)g2≡0(mod12).
... The necessary conditions for the existence of a super-simple (v, 5, λ)-BIBD are known to be sufficient for λ ∈ {2, 3, 4, 5} (see [1,[15][16][17]23,25,33]). For more results on super-simple design we refer the reader to [2,5,8,13,18,28,32,33] and references therein. ...
... □ When t ≡ 2 (mod 3), then 4t + 1 ≡ 9 (mod 12). We can obtain a super-simple ({3, 4}, 5)-GDD of group type (4t) 3 (9) 1 by deleting 4t − 9 points from the last group of the super-simple TD 5 Proof. For v ≡ 6 (mod 12) and v ≥ 222, it can be written as v = 12t + 6, where t ≥ 18. ...
... For v ≡ 10 (mod 12) and v > 34, it can be written as v = 12t + 10, where t > 2.When t ≡ 0, 1 (mod 3), then 4t ≡ 0, 4 (mod 12). We can obtain a super-simple ({3, 4}, 5)-GDD of group type (4t) 3 (10) 1 by deleting 4t − 10 points from the last group of the super-simple TD5 (4, 4t) coming from Lemma 2.1. By Lemma 2.3 with η = 0, we get a super-simple (v, {3, 4}, 5)-PBD, where the input super-simple (4t, {3, 4}, 5)-PBD and (10, {3, 4}, 5)-PBD come from Lemma 4.6 and Lemma 3.2 respectively. ...
Article
Super-simple pairwise balanced designs with block sizes 3 and 4
... In [7] Grannell, Griggs and Quinn have shown that for each admissible value of v, there exists a simple 2-(v, 3, 1)DD whose smallest defining sets have at least a half of the blocks. In [3] Amirzade and Soltankhah have proved that for all v ≡ 1 (mod 3) there exists a super-simple 2-(v, 4, 1)DD whose smallest defining sets have at least a half of blocks. The necessary condition for the existence of a super-simple 2-(v, 5, 1) directed design is that v ≡ 1, 5 (mod 10) and v = 5, 15. ...
... Here we will construct super-simple DGDDs with block size 5 and index 1 of type (g u ) for (g, u) = (5, 5), (5, 7), (5,9), (6,5), (6,6), (10,5), (4,5), (4,6), (4,10), (4,11), (4,16), (2,6), (10,6) and super-simple DGDDs of type (g u m 1 ) for (g, u, m) = (6,5,8), (4,8,6), (1,20,3), (1,16,5). Now we can construct a super-simple DGDD of type (5 5 ) on the set V = {0, 1, 2, 3, 4} × Z 5 with groups {(i, 0), (i, 1), (i, 2), (i, 3), (i, 4)}; i = 0, 1, 2, 3, 4 by developing the second coordinate of the following 10 base blocks modulo 5 or simply by (−, +1 mod 5). ...
... ((0, 0), (1, 1), (2,4), (3,4) ...
Article
In this paper we investigate the spectrum of super-simple 2-$(v,5,1)$ directed designs (or simply super-simple 2-$(v,5,1)$DDs) and also the size of their smallest defining sets. We show that for all $v\equiv1,5\ ({\rm mod}\ 10)$ except $v=5,15$ there exists a super-simple $(v,5,1)DD$. Also for these parameters, except possibly $v=11,91$, there exists a super-simple 2-$(v,5,1)$DD whose smallest defining sets have at least a half of the blocks.
... In [12], it has been shown that for each admissible value of v, there exists a simple (v, 3, 1)DD whose smallest defining sets have at least a half of the blocks. In [16], it has been shown that the necessary and sufficient condition for the existence of a super-simple (v, 4, 1)DD is v ≡ 1 (mod 3) and for these values of v except v = 7, there exists a super-simple (v, 4, 1)DD whose smallest defining sets have at least a half of the blocks. Also, in [17], it has been shown that for all v ≡ 1, 5 (mod 10) except v = 5, 15, there exists a super-simple (v, 5, 1)DD such that their smallest defining sets have at least a half of the blocks. ...
... Starting from a T D(5, 7) coming from Lemma 1 and applying Construction 1 by using a super-simple (5, 2)-DGDD of type 5 5 with d ≥ (1,0,5,22,15) (13,2,0,16,21) (0, 1,20,9,16) This super-simple DGDD has 96 blocks, each of two columns has 24 disjoint directed trades of volume 2. Therefore each defining set for this super-simple DGDD contains at least 24 × 2 = 48 blocks. So d ≥ 1 2 . ...
Preprint
In this paper, we show that for all $v\equiv 0,1$ (mod 5) and $v \geq 15$, there exists a super-simple $(v,5,2)$ directed design. Also, we show that for these parameters there exists a super-simple $(v,5,2)$ directed design such that its smallest defining sets contain at least half of its blocks.
... In [11], it has been shown that for each admissible value of v, there exists a simple 2 − (v, 3, 1)DD whose smallest defining sets have at least a half of the blocks. In [15], it has been shown that the necessary and sufficient condition for the existence of a super-simple 2 − (v, 4, 1)DD is v ≡ 1 (mod 3) and for these values of v except v = 7, there exists a super-simple 2 − (v, 4, 1)DD whose smallest defining sets have at least a half of the blocks. Also, in [16], it has been shown that for all v ≡ 1, 5 (mod 10) except v = 5, 15, there exists a super-simple 2 − (v, 5, 1)DD such that their smallest defining sets have at least a half of the blocks. ...
... So the desired inequality follows. 15,22,9) This super-simple DGDD has 144 blocks, each of two columns of this supersimple DGDD contains 24 cyclical trades of volume 3, so each defining set for this super-simple DGDD contains at least 2( 3+1 2 × 24) = 96 blocks. So the desired inequality follows. ...
Article
Full-text available
In this paper, we show that for all v\pmod 1 (mod 3), there exists a super- simple (v, 4, 2) directed design. Also, we show that for these parameters there exists a super-simple (v, 4, 2) directed design whose each defining set has at least a half of the blocks.
... In [23] it is shown that for each admissible value of v, there exists a simple 2-(v, 3, 1)-DD whose smallest defining sets have at least half of the blocks. Moreover, for all v ≡ 1 (mod 3) there exist a super-simple 2-(v, 4, 1)-DD and a super-simple 2-(v, 4, 2)-DD whose smallest defining sets have at least half of the blocks [24], [25]. Also, the existence of super-simple 2-(v, 5, 1)-DDs and their smallest defining sets for all v ≡ 1, 5 (mod 10) is shown in [26]. ...
Preprint
LDPC codes based on multiple-edge protographs potentially have larger minimum distances compared to their counterparts, single-edge protographs. However, considering different features of their Tanner graph, such as short cycles, girth and other graphical structures, is harder than for Tanner graphs from single-edge protographs. In this paper, we provide a novel approach to construct the parity-check matrix of an LDPC code which is based on trades obtained from block designs. We employ our method to construct two important categories of LDPC codes; quasi-cyclic (QC) LDPC and spatially-coupled LDPC (SC-LDPC) codes. We use those trade-based matrices to define base matrices of multiple-edge protographs. The construction of exponent matrices corresponding to these base matrices has less complexity compared to the ones proposed in the literature. We prove that these base matrices result in QC-LDPC codes with smaller lower bounds on the lifting degree than existing ones. There are three categories of SC-LDPC codes: periodic, time-invariant and time-varying. Constructing the parity-check matrix of the third one is more difficult because of the time dependency in the parity-check matrix. We use a trade-based matrix to obtain the parity-check matrix of a time-varying SC-LDPC code in which each downwards row displacement of the trade-based matrix yields syndrome matrices of a particular time. Combining the different row shifts the whole parity-check matrix is obtained. Our proposed method to construct parity-check and base matrices from trade designs is applicable to any type of super-simple directed block designs. We apply our technique to directed designs with smallest defining sets containing at least half of the blocks. To demonstrate the significance of our contribution, we provide a number of numerical and simulation results.
... The necessary conditions for the existence of a super-simple (v, 5, λ)-BIBD are known to be sufficient for λ ∈ {2, 3, 4, 5} (see [1,16,17,18,25,36]). For more results on super-simple designs we refer the reader to [2,5,8,11,19,20,28,34,36] and references therein. ...
Article
Full-text available
In statistical planning of experiments, super-simple designs are the ones providing samples with maximum intersection as small as possible. Super-simple pairwise balanced designs are useful in constructing other types of super-simple designs which can be applied to codes and designs. In this paper, the super-simple pairwise balanced designs with block sizes 3 and 4 are investigated and it is proved that the necessary conditions for the existence of a super-simple (v, {3, 4}, λ)-PBD for λ = 7, 9 and λ = 2k, k ≥ 1, are sufficient with seven possible exceptions. In the end, several optical orthogonal codes and superimposed codes are given.
Article
Full-text available
Article
Non-uniform group divisible designs are instrumental in the constructions for other types of designs. Most of the progress for the existence of {4}-GDDs of type gum1 is on the case when gu is even, where the existence for small g has played a key role. In order to determine the spectrum for {4}-GDDs of type gum1 with gu being odd, we continue to investigate the small cases with g {7, 9, 21} in this paper. We show that, for each g {7, 9, 21}, the necessary conditions for the existence of a {4}-GDD of type gum1 are also sufficient. As the applications of these GDDs, we obtain a few pairwise balanced designs with minimum block size 4. Meanwhile, we also improve the existence result for frame self-orthogonal Mendelsohn triple systems of type hn by reducing an infinite class of possible exceptions, namely n = 9 and h 2 mod 6, to eight undetermined cases.
Article
Super-simple designs are useful in the construction of superimposed codes. The necessary condition for the existence of a super-simple balanced incomplete block design on v points with k=4 and λ=6, is that v⩾14. The condition is shown to be sufficient.