Difference bases in cyclic groups
Taras Banakh, Volodymyr Gavrylkiv

TL;DR
This paper investigates the minimal size of difference bases in cyclic groups, establishing bounds and exact values for groups of small order, advancing understanding of their combinatorial structure.
Contribution
It provides explicit bounds for the difference size of cyclic groups and computes exact values for groups up to order 100, improving previous knowledge.
Findings
Bounds for difference size: (1+√(4|n|-3))/2 ≤ Δ[C_n] ≤ 1.5√n
Improved bounds for large n: Δ[C_n] ≤ (12/√73)√n for n ≥ 9 and n ≥ 2×10^15
Exact difference sizes calculated for all cyclic groups of order ≤ 100
Abstract
A subset of an Abelian group is called a difference basis of if each element can be written as the difference of some elements . The smallest cardinality of a difference basis is called the difference size of and is denoted by . We prove that for every the cyclic group of order has difference size . If (and ), then (and ). Also we calculate the difference sizes of all cyclic groups of cardinality .
| 2 | 2 | 3 | 3 | 3 | 3 | 4 | 4 | 4 | 4 | 4 | 5 | |
| 3 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 5 | 5 | 5 | 5 |
| 1 | 3 | 6 | 9 | 13 | 18 | 24 | 29 | 37 | 45 | 51 | 61 | |
| 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | |
| 4 | 3 | 2.666 | 2.777 | 2.769 | 2.722 | 2.666 | 2.793 | 2.703 | 2.688 | 2.824 | 2.770 |
| 4064 | 4713 | 5416 | 6166 | |
| 104 | 112 | 120 | 128 | |
| 2.6615 | 2.6616 | 2.6588 | 2.6572 |
| 2 | 3 | 4 | 5 | 7 | 8 | 9 | 11 | 13 | 16 | 17 | 19 | 23 | 25 | 27 | 29 | 31 | 32 | |
| 7 | 13 | 21 | 31 | 57 | 73 | 91 | 133 | 183 | 273 | 307 | 381 | 553 | 651 | 757 | 871 | 993 | 1057 | |
| 4 | 7 | 10 | 14 | 22 | 28 | 36 | 48 | 56 | 72 | 91 | 98 | 128 | 159 | 172 | 191 | 209 | 198 |
| 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | ||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 7 | 4 | 6 | 7 | - | - | - | - | - | - | - | - | - | - | - | - | - | - | |
| 13 | 7 | 10 | 12 | 13 | - | - | - | - | - | - | - | - | - | - | - | - | - | |
| 21 | 10 | 13 | 17 | 20 | 21 | - | - | - | - | - | - | - | - | - | - | - | - | |
| 31 | 14 | 19 | 25 | 28 | 30 | 31 | - | - | - | - | - | - | - | - | - | - | - | |
| 39 | 18 | 23 | 29 | 35 | 37 | 38 | 39 | - | - | - | - | - | - | - | - | - | - | |
| 57 | 22 | 29 | 38 | 43 | 50 | 54 | 56 | 57 | - | - | - | - | - | - | - | - | - | |
| 73 | 28 | 38 | 41 | 52 | 58 | 66 | 70 | 72 | 73 | - | - | - | - | - | - | - | - | |
| 91 | 36 | 46 | 51 | 62 | 71 | 74 | 84 | 88 | 90 | 91 | - | - | - | - | - | - | - | |
| 133 | 48 | 61 | 68 | 79 | 94 | 102 | 114 | 120 | 125 | 130 | 132 | 133 | - | - | - | - | - | |
| 183 | 56 | 69 | 92 | 98 | 109 | 128 | 139 | 153 | 158 | 168 | 175 | 180 | 182 | 183 | - | - | - | |
| 273 | 72 | 92 | 110 | 130 | 145 | 152 | 175 | 184 | 202 | 218 | 226 | 242 | 258 | 266 | 270 | 272 | 273 | |
| 307 | 91 | 106 | 128 | 152 | 160 | 171 | 199 | 214 | 232 | 245 | 261 | 276 | 284 | 292 | 300 | 304 | 306 |
| 72 | 84 | 87 | 105 | 109 | 135 | 138 | 142 | 145 | 200 | 204 | 208 | 211 | 251 | 256 | 262 | 268 | |
| 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 |
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | |
| 1 | 2 | 2 | 3 | 3 | 3 | 3 | 4 | 4 | 4 | 4 | 4 | 4 | 5 | 5 | 5 | 5 | 5 | 5 | 5 | |
| 1 | 2 | 2 | 3 | 3 | 3 | 3 | 4 | 4 | 4 | 4 | 4 | 4 | 5 | 5 | 5 | 5 | 5 | 5 | 6 | |
| 2 | 2 | 4 | 3 | 4 | 4 | 5 | 8 | |||||||||||||
| 1 | 2 | 2 | 3 | 3 | 3 | 3 | 4 | 4 | 4 | 4 | 4 | 4 | 5 | 5 | 5 | 5 | 5 | 5 | 6 | |
| 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | 32 | 33 | 34 | 35 | 36 | 37 | 38 | 39 | 40 | |
| 5 | 6 | 6 | 6 | 6 | 6 | 6 | 6 | 6 | 6 | 6 | 7 | 7 | 7 | 7 | 7 | 7 | 7 | 7 | 7 | |
| 5 | 6 | 6 | 6 | 6 | 6 | 6 | 6 | 7 | 7 | 6 | 7 | 7 | 7 | 7 | 7 | 7 | 8 | 7 | 8 | |
| 5 | 7 | 6 | ||||||||||||||||||
| 6 | 6 | 6 | 6 | 6 | 6 | 6 | 7 | 7 | 7 | 7 | 7 | 7 | 7 | 7 | 7 | 7 | 8 | 8 | 8 | |
| 41 | 42 | 43 | 44 | 45 | 46 | 47 | 48 | 49 | 50 | 51 | 52 | 53 | 54 | 55 | 56 | 57 | 58 | 59 | 60 | |
| 7 | 7 | 7 | 8 | 8 | 8 | 8 | 8 | 8 | 8 | 8 | 8 | 8 | 8 | 8 | 8 | 8 | 9 | 9 | 9 | |
| 8 | 8 | 8 | 8 | 8 | 8 | 8 | 8 | 8 | 8 | 8 | 9 | 9 | 9 | 9 | 9 | 8 | 9 | 9 | 9 | |
| 10 | 9 | 8 | ||||||||||||||||||
| 8 | 8 | 8 | 8 | 8 | 8 | 8 | 8 | 8 | 9 | 9 | 9 | 9 | 9 | 9 | 9 | 9 | 9 | 9 | 10 | |
| 61 | 62 | 63 | 64 | 65 | 66 | 67 | 68 | 69 | 70 | 71 | 72 | 73 | 74 | 75 | 76 | 77 | 78 | 79 | 80 | |
| 9 | 9 | 9 | 9 | 9 | 9 | 9 | 9 | 9 | 9 | 9 | 9 | 9 | 10 | 10 | 10 | 10 | 10 | 10 | 10 | |
| 9 | 9 | 9 | 9 | 9 | 10 | 10 | 10 | 10 | 10 | 10 | 10 | 9 | 10 | 10 | 10 | 10 | 10 | 10 | 11 | |
| 10 | 9 | 12 | ||||||||||||||||||
| 10 | 10 | 10 | 10 | 10 | 10 | 10 | 10 | 10 | 10 | 10 | 10 | 10 | 10 | 10 | 11 | 11 | 11 | 11 | 11 | |
| 81 | 82 | 83 | 84 | 85 | 86 | 87 | 88 | 89 | 90 | 91 | 92 | 93 | 94 | 95 | 96 | 97 | 98 | 99 | 100 | |
| 10 | 10 | 10 | 10 | 10 | 10 | 10 | 10 | 10 | 10 | 10 | 11 | 11 | 11 | 11 | 11 | 11 | 11 | 11 | 11 | |
| 11 | 11 | 11 | 11 | 11 | 11 | 11 | 11 | 11 | 11 | 10 | 11 | 12 | 12 | 12 | 12 | 12 | 12 | 12 | 12 | |
| 10 | ||||||||||||||||||||
| 11 | 11 | 11 | 11 | 11 | 11 | 11 | 11 | 11 | 11 | 11 | 12 | 12 | 12 | 13 | 13 | 13 | 13 | 13 | 13 |
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Difference bases in cyclic groups
Taras Banakh and Volodymyr Gavrylkiv
Ivan Franko National University of Lviv (Ukraine), and
Institute of Mathematics, Jan Kochanowski University in Kielce (Poland)
Vasyl Stefanyk Precarpathian National University, Ivano-Frankivsk, Ukraine
Abstract.
A subset of an Abelian group is called a difference basis of if each element can be written as the difference of some elements . The smallest cardinality of a difference basis is called the difference size of and is denoted by . We prove that for every the cyclic group of order has difference size . If (and ), then (and ). Also we calculate the difference sizes of all cyclic groups of cardinality .
Key words and phrases:
finite group, cyclic group, difference basis, difference characteristic
1991 Mathematics Subject Classification:
05B10, 05E15, 20D60
1. Introduction
A subset of a group is called a difference basis for a subset if each element can be written as for some . If the group operation of is denoted by , then the element is written as the difference (which justifies the choice of the terminology).
The smallest cardinality of a difference basis for is called the difference size of and is denoted by . For example, the set is a difference basis for the interval witnessing that . In Proposition 2.2(4) we shall prove that the difference size is subadditive in the sense that for any non-empty subsets of a group .
The definition of a difference basis for a set in a group implies that and hence . The fraction
[TABLE]
is called the difference characteristic of . The difference characteristic is submultiplicative in the sense that for any normal subgroup of a finite group , see [3, 1.1].
In this paper we are interested in evaluating the difference characteristics of finite cyclic groups. In fact, this problem has been studied in the literature (see [5], [10], [15]). In particular, Kozma and Lev [15] proved (using the classification of finite simple groups) that each finite group has difference characteristic . In this paper we shall show that for finite cyclic groups this upper bound can be improved to . Moreover, if a finite cyclic group has cardinality (resp. ), then (resp. ). It is an open problem if . However, many subsequences of the sequence indeed converge to . In particular, using known results on (relative) difference sets, we shall prove that
[TABLE]
where runs over prime numbers and runs over prime powers. A number is called a prime power if is equal to the power of some prime number .
To derive an upper bound for the difference sizes of arbitrary finite cyclic group, we shall use known information on the difference sizes of order intervals in the group of integer numbers. Here we exploit the approach first used by Rédei and Rényi [17] and then developed by Leech [14] and Golay [12].
For a model of a cyclic group of order we take the multiplicative group
[TABLE]
of complex -th roots of .
2. Known results
In this section we recall some known results on difference bases in finite groups. The following important fact was proved by Kozma and Lev [15] (using the classification of finite simple groups).
Theorem 2.1** (Kozma, Lev).**
Each finite group has difference characteristic .
For cyclic groups the upper bound can be improved to , which will be done in Theorem 4.14.
For a real number we put
[TABLE]
The first three statements of the following proposition were proved in [3, 1.1].
Proposition 2.2**.**
Let be a finite group. Then
- (1)
\frac{1+\sqrt{4|G|-3}}{2}\leq\Delta[G]\leq\big{\lceil}\frac{|G|+1}{2}\big{\rceil}, 2. (2)
* and for any normal subgroup ;* 3. (3)
* for any subgroup ;* 4. (4)
* for any non-empty sets .*
Proof.
Since (1)–(3) are proved in [3, 1.1], we shall explain (4). Given non-empty sets , find difference bases and for the sets of cardinality and . Taking any point and replacing by its shift , we can assume that the unit of the group belongs to . By the same reason, we can assume that . The union is a difference basis for , witnessing that
[TABLE]
∎
Finite groups with \Delta[G]=\big{\lceil}\frac{|G|+1}{2}\big{\rceil} were characterized in [3] as follows.
Theorem 2.3** (Banakh, Gavrylkiv, Nykyforchyn).**
For a finite group
- (i)
\Delta[G]=\big{\lceil}\frac{|G|+1}{2}\big{\rceil}>\frac{|G|}{2}* if and only if is isomorphic to one of the groups:*
, , , , , , , ; 2. (ii)
* if and only if is isomorphic to one of the groups: , , , , .*
In this theorem by we denote the dihedral group of cardinality and by the 8-element group of quaternion units. In [3] the difference sizes was calculated for all groups of cardinality .
Observing that for each cyclic group of cardinality the difference size coincides with the lower bound \big{\lceil}\frac{1+\sqrt{4n-3}}{2}\big{\rceil} given in Proposition 2.2(1), the authors of [3] posed the following problem.
Problem 2.4** (Banakh, Gavrylkiv, Nykyforchyn).**
Is \Delta[C_{n}]=\big{\lceil}\frac{1+\sqrt{4n-3}}{2}\big{\rceil} for any finite cyclic groups ?
Using the results of computer calculations we shall give a negative answer to Problem 2.4. On the other hand, we shall observe that the classical difference sets of Singer [19] witness that for any number where is a prime power.
3. Difference sizes of some special cyclic groups
In this section we collect some upper bounds on the difference size of a cyclic group whose order has some special arithmetic properties, for example, is equal to or for a prime power or to for a prime number . To derive such upper bounds, we shall use known information on (relative) difference sets.
A subset of a group is called a difference set if each non-idempotent element can be uniquely written as for some elements . This definition implies and hence . The following fundamental result was proved by Singer [19] in 1938.
Theorem 3.1** (Singer).**
For any prime power the cyclic group contains a difference set of cardinality and hence has difference size .
Singer’s Theorem implies that for any prime power the cyclic group of cardinality has difference size
[TABLE]
So, for infinitely many numbers the lower bound for (given in Proposition 2.2(1)) is attained.
The converse result to Singer’s Theorem is known in Algebraic Combinatorics as PPC (abbreviated from the Prime Power Conjecture).
Conjecture 3.2** ().**
If for a natural number some Abelian group of order contains a difference set , then is a prime power.
In [13] is confirmed for all numbers . The Prime Power Conjecture implies the following converse to Theorem 3.1.
Proposition 3.3**.**
Let be an Abelian group of difference size . If holds, then is a prime power and .
Proof.
Let . The equality implies that |G|=\frac{1}{4}\big{(}(2q+1)^{2}+3\big{)}=q^{2}+q+1. Let be a difference basis of cardinality in . Consider the surjective map , . The equality implies that for each the preimage is a singleton, which means that is a difference set in . Now implies that is a prime power. ∎
Singer derived his theorem studying properties of projective planes over finite fields. A corresponding result for affine planes was obtained by Bose [6] and Chowla [7].
Theorem 3.4** (Bose-Chowla).**
For any prime power the set has a difference basis of cardinality in the cyclic group . Consequently, and
[TABLE]
We shall also need the following theorem essentially proved by Rusza [18].
Theorem 3.5** (Rusza).**
For any prime number the set has a difference basis of cardinality in the group . Consequently, and
[TABLE]
Proof.
Given a prime number , consider the field of residues modulo . It is known that the multiplicative group of this field is isomorphic to the cyclic group . Since the numbers and are relatively prime, the product is isomorphic to the cyclic group . So, instead of the group , we can consider the group (which is the direct product of the additive and multiplicative groups of the field ).
We claim that the set is a difference basis for the set in the group . Given any pair we need to find two elements such that . Since , the element is invertible in the field , so we can consider the elements and . The pairs and belong to the set and their difference in the group is equal to , witnessing that is a difference basis for the set .
Observe that the complement of the set in the group is equal to the union of two subgroups and . Therefore,
[TABLE]
The upper bound combined with the equality
[TABLE]
implies that .
By Proposition 2.2(4),
[TABLE]
∎
4. Difference sizes of number intervals
In this section we apply known information on difference sizes of number intervals to evaluating the difference sizes of finite cyclic groups. For integer numbers by we shall denote the order-interval in the group of integer numbers.
For a natural number by we shall denote the difference size of the interval . It is equal to the difference size of the intervals and . Also we put .
For example, the interval has difference size as witnessed by difference basis . It is clear that is a non-decreasing function of the integer parameter and , which implies that for all .
Difference bases for the order intervals were studied by Rédei and Rényi [17] who proved that the limit exists and is equal to . Moreover,
[TABLE]
These lower and upper bounds were improved by Leech [14] and Golay [12] who proved the following theorem.
Theorem 4.1** (Leech-Golay).**
For any natural number we get the lower and upper bounds:
[TABLE]
For small numbers the difference sizes of the intervals have been calculated by computer. The following table (taken from [14] and [20]) gives the values of for numbers such that .
This table shows that for the smallest value of the difference characteristic is attained for . Combining the difference basis for with the Singer difference sets, Leech [14] and Golay [12] have found four larger numbers with . These (four) numbers are presented in Table 3.
Difference sizes of the intervals yield upper bounds on the difference sizes of cyclic groups.
Proposition 4.2**.**
For any natural number we get
[TABLE]
where
Proof.
Given a natural number , consider the homomorphism , . For the number , choose a subset of cardinality such that contains the interval . Taking into account that , we conclude that , which means that the set is a difference basis for the group and hence .
Observe that k=\big{\lceil}\frac{n-1}{2}\big{\rceil}\leq\frac{n}{2} and hence
[TABLE]
∎
Proposition 4.2 and Theorem 4.1 allow us to evaluate lower and upper limits of the difference characteristics of cyclic groups.
Corollary 4.3**.**
For every natural number we have the lower and upper bounds:
[TABLE]
Corollary 4.3 implies that for all sufficiently large . In Theorem 4.12 we shall show that this upper bound holds for all .
At first we find some upper bounds of the difference sizes of the intervals , using the approach first exploited by Rédei and Rényi [17], and then developed by Leech [14] and Golay [12].
To write down these upper bounds, we shall need some information on the numbers , which are defined as follows. For a natural number and a non-negative number let be the largest integer number for which there exists a set of cardinality such that and . In [12] the numbers were denoted by and were calculated for all cyclic groups of order where is a prime power:
This table is completed by Table 5 giving the values of for positive . These values are found by computer.
Theorem 4.4**.**
For any natural number we get the lower bound
[TABLE]
If for some prime power , then
[TABLE]
Proof.
Given a natural number , fix a difference basis of cardinality . We lose no generality assuming that contains the unit of the group . It follows that . Let be a generator of the cyclic group . Let d=\big{\lfloor}\frac{m}{\Delta[C_{m}]}\big{\rfloor} and consider the interval , containing elements of .
Claim 4.5**.**
There exists such that .
Proof.
For a subset denote by its characteristic function (which means that ). Assume that for all and observe that
[TABLE]
which implies that and for all . The latter property of can be used to show that coincides with the subgroup generated by , which is not possible as . ∎
By Claim 4.5, there exists such that is empty. Then is a difference basis for with . It follows that the set has and witnesses that \delta_{0}[C_{m}]\geq d=\big{\lfloor}\frac{m}{\Delta[C_{m}]}\big{\rfloor}.
Now assume that for some prime power . In this case is a difference set of cardinality . We recall that is the multiplicative subgroup of the unit circle on the complex plane. Let be the upper half-circle and observe that .
Claim 4.6**.**
For some the set has cardinality .
Proof.
For a subset by we denote the characteristic function of the set in , which means that . Observe that each element has positive real part and the sum is a positive real number.
Now consider the complex number and observe that
[TABLE]
Then for some and hence the set has .
Let . Observe that
[TABLE]
Assuming that for all , we conclude that
[TABLE]
for all . Taking into account that each complex number has negative real part ,
[TABLE]
and
[TABLE]
we conclude that
[TABLE]
which implies that and hence . But this contradicts the definition of the constant . ∎
By Claim 4.6, there exists a complex number such that . Let and . Observe that the arc can be covered by disjoint copies of the arc . By the Pigeonhole Principle, for some the arc is disjoint with the set . Then the arc is disjoint with the set . It is easy to see that the arc contains at least consecutive points of the group . Therefore, the set contains consecutive points. Replacing by a suitable shift, we can assume that those consecutive points form the set where is the generator of the cyclic group . Then the set is disjoint with the set and witnesses that
[TABLE]
If , then
[TABLE]
If , then the equality follows from Table 5. ∎
The numbers are used in the following theorem giving an upper bound for the difference sizes of intervals.
Theorem 4.7**.**
For any non-negative integer numbers with we get the upper bound
[TABLE]
Proof.
Fix a difference basis for the interval of cardinality .
By the definition of the number , there exists a set of cardinality such that and . Find two numbers with . It is clear that the set
[TABLE]
has cardinality
[TABLE]
We claim that the interval is contained in . Since the set is symmetric, it suffices to show that each positive number is contained in . Write as for some integer numbers such that and . By the choice of , there are numbers such that for some . Taking into account that , we conclude that and hence .
It follows that . If , then we can choose two numbers such that and conclude that
[TABLE]
So, we assume that . Then . Taking into account that , we conclude that , and hence . Then . Therefore and . ∎
Corollary 4.8**.**
Let be a natural number, is a prime power and . For any natural number we get the upper bound
[TABLE]
Proof.
By Theorem 3.1, the cyclic group of order has difference size . By Theorem 4.4, . By Theorem 4.7,
[TABLE]
∎
Applying Corollary 4.8 with we derive another corollary.
Corollary 4.9**.**
For any prime power and a natural number with we get the upper bound .
For a real number by we denote the smallest prime power, which is larger or equal than . It is easy to see that for real numbers and the inequality is equivalent to . This observation, combined with Corollary 4.9 yields the following upper bound for .
Corollary 4.10**.**
Each finite cyclic group of order has difference size \Delta[C_{n}]\leq 4+4q\big{(}\frac{-7+\sqrt{12n-131}}{12}\big{)}.
For a real number by we denote the smallest prime number greater or equal to . It is clear that . By [2], . The (still unproven) Andrica’s Conjecture [1] says that for all . This conjecture was confirmed by Imran Ghory [11] for all .
Corollary 4.11**.**
Each finite cyclic group of cardinality has difference size
[TABLE]
If the Andrica Conjecture is true, then this upper bound holds for all numbers .
Proof.
Given a number , consider the real number . The inequality implies that , so we can apply the result of Ghory [11], and conclude that . By Corollary 4.10,
[TABLE]
If the Andrica’s Conjecture is true, then the same argument works for all . ∎
Applying Corollary 4.10 with and the upper bound from [2], we get a more refined upper bound for .
Theorem 4.12**.**
For any prime power and a natural number we get the upper bound . Consequently, for any we get the upper bound
[TABLE]
If , then .
Proof.
By Theorem 3.1, the cyclic group of order has difference size . By Theorem 4.4, . Since , we can apply Theorem 4.7 and obtain the upper bound
[TABLE]
If , then the real number is well-defined and
[TABLE]
By [2], . Then
[TABLE]
Now assume that . In this case . By [8, 6.8], p(x)\leq x\big{(}1+\frac{1}{25\ln^{2}x}\big{)}. Then and we see that the inequality , follows from the inequality
[TABLE]
holding for all . ∎
Now we evaluate the difference sizes of intervals and cyclic groups for relatively small .
Applying Theorem 4.7 to the known values of the difference sizes , , , known values of difference sizes of cyclic groups given in Table 7, and known values of the numbers given in Tables 4 and 5, we obtain the upper bounds for the difference sizes of intervals of length , given in Table 6.
Now we can prove an upper bound for , holding for small .
Theorem 4.13**.**
The upper bound holds for all . Moreover, if , then .
Proof.
For the inequality can be verified using known values of , see Table 7.
For the upper bound follows from the upper bound \Delta[C_{n}]\leq\Delta[\big{\lceil}\frac{n-1}{2}\rceil\big{]} given in Proposition 4.2 and the upper bounds for the difference sizes given in Table 6. For example, let us prove the upper bound for the number .
By Proposition 4.2, . Looking at Table 2, we see that . Consequently, . For all other numbers by the same method we get the inequality .
For we shall apply the known fact (see the sequence https://oeis.org/A166968 on the On-line Encyclopedia of Integer Sequences) saying that for every the interval contains a prime number. Consider the real number and let be the smallest prime power which is greater or equal to . It follows that
[TABLE]
The inequality implies . By Corollary 4.10, . To prove that , it remains to check that . The elementary calculations show that this inequality holds for all . ∎
Theorem 4.13 and known values of for allow us to find the largest value of the difference characteristics .
Corollary 4.14**.**
.
Also we can establish some upper bounds for the difference sizes of cyclic groups whose cardinality has some special arithmetic properties.
Corollary 4.15**.**
For any prime power the cyclic group has difference size
[TABLE]
Proof.
For the upper bound
[TABLE]
can be verified using known values of the difference sizes given in Table 7.
So, assume that and hence . In this case Theorem 4.13 guarantees that . Applying Theorem 3.4, we conclude that the group has difference size
[TABLE]
∎
By analogy we can derive the following upper bound for from Theorem 3.5.
Corollary 4.16**.**
For any prime number the cyclic group has difference size
[TABLE]
In Table 7 we present the results of computer calculation of the difference sizes of cyclic groups of order . In this table
[TABLE]
is the lower bound given in Proposition 2.2(1) and
[TABLE]
is the upper bound for given in Theorems 3.1, 3.4 and 3.5. With the boldface font we denote the numbers , equal to for a prime power . For such numbers we know the exact value .
Remark 4.17**.**
For we get the strict inequality , which answers Problem 2.4 in negative.
The results of computer calculations suggest the following questions.
Question 4.18**.**
Is for every ?
Question 4.19**.**
Is ?
The following problem seems to be the most intriguing (see http://mathoverflow.net/questions/262317).
Problem 4.20**.**
Is as ? Equivalently, is
Theorem 3.1, Corollaries 4.15, 4.16, and Proposition 2.2(2) allows us to produce many subsequences of the sequence tending to the unit. In particular,
[TABLE]
where runs over prime powers and runs over prime numbers to infinity. On the other hand, we do not know the answers to the following problems (which are weaker versions of Problem 4.20).
Problem 4.21**.**
Is ?
Problem 4.22**.**
Let be a prime number. Is ?
Problem 4.23**.**
Is ?
5. Acknowledgment
The authors would like to express their sincere thanks to Oleg Verbitsky who turned their attention to the theory of difference sets and their relation with difference bases, to Alex Ravsky for valuable discussions on perfect rulers, to MathOverflow users Lucia, Seva, and Sean Eberhard for valuable comments to the questions asked by the first author on MathOverflow.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] D. Andrica, Note on a conjecture in prime number theory , Studia Univ. Babes–Bolyai Math. 31 :4 (1986) 44–48.
- 2[2] R.C. Baker, G. Harman, J. Pintz, The difference between consecutive primes, II . Proc. London Math. Soc. 83 :3 (2001), 532–562.
- 3[3] T. Banakh, V. Gavrylkiv, O. Nykyforchyn, Algebra in superextension of groups, I: zeros and commutativity , Algebra Discr. Math. 3 (2008), 1–29.
- 4[4] T. Banakh, V. Gavrylkiv, Algebra in the superextensions of twinic groups , Dissertationes Math. 473 (2010), 74 pp.
- 5[5] E. Bertram, M. Herzog, Bounds on character degrees and class numbers of finite nonabelian simple groups , Groups–St. Andrews 1989, Vol. 1, 46–51, London Math. Soc. Lecture Note Ser., 159, Cambridge Univ. Press, Cambridge, 1991.
- 6[6] R.C. Bose, An affine analogue of Singer s theorem , J. Indian Math. Soc. 6 (1942) 1–15.
- 7[7] R.C. Bose, S. Chowla, Theorems in the additive theory of numbers , Comment. Math. Helvetici 37 (1962-63) 141 -147.
- 8[8] P. Dusart, Estimates of ψ , θ 𝜓 𝜃 \psi,\theta for large values of x 𝑥 x without the Riemann hypothesis , Math. Comp. 85 :298 (2016) 875–888.
