Next Article in Journal
Existence and Uniqueness Results for Fuzzy Bipolar Metric Spaces
Next Article in Special Issue
A New Family of Multipartition Graph Operations and Its Applications in Constructing Several Special Graphs
Previous Article in Journal
Systematic Literature Review of Optimization Algorithms for P||Cmax Problem
Previous Article in Special Issue
Sequence of Bounds for Spectral Radius and Energy of Digraph
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Symmetric Point Sets with Few Intersection Numbers in PG(r,q)

by
Stefano Innamorati
Department of Industrial and Information Engineering and Economic, University of L’Aquila, Piazzale Ernesto Pontieri, 1, I-67100 L’Aquila, Italy
Symmetry 2025, 17(2), 179; https://doi.org/10.3390/sym17020179
Submission received: 2 January 2025 / Revised: 19 January 2025 / Accepted: 22 January 2025 / Published: 24 January 2025
(This article belongs to the Special Issue Symmetry in Combinatorics and Discrete Mathematics)

Abstract

:
In 1965, H. Retkin and E. Stein defined a symmetric point set as a set of points with the same intersection numbers. In this paper, we perform a detailed analysis of symmetric point sets of the finite projective space which shows that the class of symmetric sets is very broad including caps, two-character sets and transitive sets. We derive necessary conditions for the existence of such sets. Since the most studied sets are caps and two-character sets and not much seems to be known in the general case of sets, which are different from caps, with more than two intersection numbers, by using incidence-preserving group actions, symmetric point sets with few intersection numbers are provided. The results indicate that any finite projective space contains symmetric sets with few intersection numbers.

1. Introduction

Let 𝔽q be the finite field of order q, with q = ph, a prime power. In PG(r,q), the projective space of dimension r and order q, let K denote a k–set, i.e., a set of k points. A set K has class [m1,m2,…,ms] if every hyperplane meets K in mi points for some i = 1,…,s, cf. [1]. If each value mi occurs as an intersection number, then we say K has type (m1, m2,…,ms), and it is called an s-intersection set, cf. [2,3]. A cap is a set of points such that at most two of them are on a hyperplane. Caps and sets with few intersection numbers have been extensively studied due to their connections to coding theory, cf. [4,5,6,7,8]. A point P has type v m 1 P , v m 2 P , , v m s P if P is contained in v m i P hyperplanes meeting K in mi points. Following [9] p. 393, a k-set K is said to be symmetric if for each value of mi the value of v m i P is independent of P, in which case we use the symbol v m i . The fundamental objective of the research developed in this paper is to initiate the investigation of symmetric sets and discuss whether such sets exist in finite projective spaces. We derive necessary conditions for the existence of such sets and construct a class of symmetric sets by using the cyclic structure of a finite projective space. The paper is organized as follows. In Section 2, we observe that sets of class [0,m,n] are a broad class of symmetric point sets including caps and two-character sets. As they have been extensively studied, in Section 3, after showing that transitive point sets represent another large class of symmetric point sets, we generalize the construction of Brower [10] to get transitive few-intersection sets. In Section 4, we explicitly construct transitive sets with at most three intersection numbers in the smallest two projective planes PG(2,q) with q≡1 mod 3. In Section 5, we derive transitive sets with at most four intersection numbers in the smallest projective space PG(3,q) with q≡1 mod 4.

2. Sets of Class [0,m,n] are Symmetric

In this section, we show that symmetric sets are far from being trivial. Such sets include caps, i.e., sets of type (0,1,2), and two-character sets and these have been subject to extensive study since the middle of the last century, c.f. [11]. Let K denote a k-set of class [0,m,n] in PG(r,q). We prove the following:
Lemma 1.
A set of class [0,m,n] is symmetric.
Proof. 
Double counting the number of hyperplanes through a point PK, the number of pairs (P,H), where PK and H is a hyperplane through P, we get the following equations on the integers vi(P):
v m P + v n P = q r 1 + + q + 1 m 1 v m P + n 1 v n P = k 1 q r 2 + + q + 1
By Cramer’s rule, the system of linear equations has one solution.
Therefore, the integers vi = vi(P) do not depend on the point P and we have that
v m = n k n m q r 2 + + q + 1 + n 1 n m q r 1 v n = k m n m q r 2 + + q + 1 m 1 n m q r 1
and the set K is symmetric. □
The existence of a set of class [0,m,n] requires particular admissible parameters. Here, we provide a necessary condition for the existence of a set K of class [0,m,n] which involves the number of external hyperplanes through any point not belonging to the set. Let Q denote a point with QK. For each integer i such that 0 ≤ iqr−1+…+q + 1, let us denote by ui(Q) the number of hyperplanes of PG(r,q) through Q meeting K in exactly i points.
Lemma 2.
A necessary condition for the existence of a k-set K of class [0,m,n] is that
n m   |   n u 0 Q q r 1 n m   |   m u 0 Q q r 1 .
Proof. 
Double counting the number of hyperplanes through a point QK, the number of pairs (Q,H), where QK and H is a hyperplane through Q, we get the following equations on the integers ui(Q):
u 0 Q + u m Q + u n Q = q r 1 + + q + 1 m u m Q + n u n Q = k q r 2 + + q + 1
Thus, we have that
u m Q = n k n m q r 2 + + q + 1 + n n m q r 1 n n m u 0 Q u n Q = k m n m q r 2 + + q + 1 m n m q r 1 + m n m u 0 Q
Therefore,
u m Q =   v m n u 0 Q q r 1 n m u n Q = v n + m u 0 Q q r 1 n m .
Since ui(Q) and vi are integers, the claimed necessary condition is established. □
Unfortunately, neither any general formula nor general divisibility properties are known for the existence of sets of class [0,m,n]. Sets of class [0,m,n] have been studied in a variety of contexts with some special cases forming individual fields of study, cf. [7]. The most interesting cases appear to be two intersection sets, cf. [12,13]. As they have been extensively studied, in the next section, we show another large class of symmetric point sets.

3. Transitive Sets of PG(r,q) Are Symmetric

In this section, inspired by the work of [10], we continue to investigate symmetric sets by using the cyclic structure of a finite projective space. We derive a sufficient condition for the existence of such sets and determine enough information on intersection numbers. We report the cyclic construction here for the sake of completeness, cf. [14]. The projective space PG(r,q) admits a collineation σ such that the group σ generated by σ is transitive and hence regular on the set of all points. The cyclic group σ is, up to conjugacy, the only cyclic subgroup of PGL(r + 1,q) acting regularly on the set of points of PG(r,q), and it is called the Singer group of PG(r,q), cf. [14]. In order to show this, let us consider F q r + 1 as a r dimensional projective space PG(r,q) over 𝔽q. The points of PG(r,q) are represented as elements of F q r + 1 / F q , with two elements v1 and v2 of F q r + 1 representing the same point if and only if v2 = λv1 for some λ in F q . Let ω be a primitive element of F q r + 1 . To construct the cyclic Singer group, we consider the following:
F q r + 1 = x 0 + x 1 ω + + x r ω r   |   x i F q ,
F q = ω q r + q r 1 + + q + 1 i   |   1 i q 1 ,
By the function f : F q r + 1 * P G ( r , q ) defined as f x 0 + x 1 ω + + x r ω r x 0 , x 1 , , x r , the points of PG(r,q) can be represented by the set Ω = ω i   |   0 i q r + q r 1 + + q . Consider the action of the map σ over Ω , where σ : Ω Ω   σ ω i = ω · ω i = ω i + 1 m o d   q r + q r 1 + + q + 1 . The map σ is a collineation of PG(r,q), as σ x 0 , x 1 , , x r = f ω x 0 + x 1 ω + + x r ω r . If p x = x r + 1 a 0 a 1 x a r x r is the minimum polynomial of ω over 𝔽q, then f ω x 0 + x 1 ω + + x r ω r = f x 0 ω + x 1 ω 2 + + x r ω r + 1 = a 0 x r , x 0 + a 1 x 2 , , x r 1 + a r x r . Thus, σ is defined by the companion matrix of p x ,   C p = 0 1 0 0 0 0 1 a 0 a 1 a 2 a r , as
σ x 0 , x 1 , , x r = x 0 , x 1 , , x r 0 1 0 0 0 0 1 a 0 a 1 a 2 a r = a 0 x r , x 0 + a 1 x r , , x r 1 + a r x r
the group σ is the cyclic Singer group of PG(r,q). A transitive k-set of PG(r,q) is a k-set which consists of a point orbit under a subgroup of the cyclic Singer group of PG(r,q), cf. [13,14,15,16]. Then, k divides q r + q r 1 + + q + 1 and the group σ q r + q r 1 + + q + 1 k has q r + q r 1 + + q + 1 k orbits on points of PG(r,q), each of size k. Since the orbits are projectively equivalent, we will assume, without loss of generality, that K = 1 , ω q r + q r 1 + + q + 1 k , , ω k 1 q r + q r 1 + + q + 1 k . It is well-known that, under the action of the cyclic Singer group σ of PG(r,q), the point set and the hyperplane set have the same cyclic structure. Thus, the group σ q r + q r 1 + + q + 1 k also has q r + q r 1 + + q + 1 k orbits on hyperplanes of PG(r,q), each of size k. Therefore, the set K = 1 , ω q r + q r 1 + + q + 1 k , , ω k 1 q r + q r 1 + + q + 1 k is a set with at most q r + q r 1 + + q + 1 k intersection numbers, mi, i 1 , , q r + q r 1 + + q + 1 k , with respect to the hyperplanes.
Since the points of PG(r,q) can be represented by the set Ω = ω i   |   0 i q r + q r 1 + + q , any point ω i can be identified with the exponent i, 0 i q r + q r 1 + + q and so with the elements of the additive cyclic group Z q r + q r 1 + + q + 1 . By using this identification, the set K = 0 , q r + q r 1 + + q + 1 k , , k 1 q r + q r 1 + + q + 1 k .
The principal Lemma relevant to the present discussion is the following:
Lemma 3.
A transitive k-set is symmetric.
Proof. 
Let i , j 0 , q r + q r 1 + + q + 1 k , , k 1 q r + q r 1 + + q + 1 k be two points of a transitive k-set K. Therefore, there is an integer a, with a 1 , 2 , , k 1 , such that j = i + a q r + q r 1 + + q + 1 k . Thus, if H is an m-hyperplane through i, then σ a q r + q r 1 + + q + 1 k H is an m-hyperplane through j and the assertion follows. □
If q r + q r 1 + + q + 1 k is small, the set K is a k-set with few intersection numbers. Now, we give an explicit construction of a transitive k-set in the projective space PG(r,q) which has few intersection numbers. We prove
Lemma 4.
If q≡1 mod r + 1, then the projective space PG(r,q) contains a transitive ( q r + q r 1 + + q + 1 r + 1 )-set with at most r + 1 intersection numbers.
Proof. 
If q ≡ 1 mod r + 1, then r + 1 is a divisor of q r + q r 1 + + q + 1 , the number of points of PG(r,q). The orbit of the point 0 of PG(r,q) under the unique subgroup Z r + 1 of order r + 1 of the additive cyclic group Z q r + q r 1 + + q + 1 is the transitive q r + q r 1 + + q + 1 r + 1 -set 0 , r + 1 , 2 r + 1 , , q r + q r 1 + + q + 1 r + 1 1 r + 1 with at most r + 1 intersection numbers. □
By using the above identification, the set K is the set K 0 = 0 , r + 1 , 2 r + 1 , , q r + q r 1 + + q + 1 r + 1 1 r + 1 and any hyperplane H0 is a (v,k,λ)-difference set, i.e., a k-set H0 = {d1,…,dk} of residues modulo v such that for any residue α≠0 modulo v the congruence didjα (mod v) has exactly λ solutions pairs (di,dj) with di and dj in H0, where v = q r + q r 1 + + q + 1 , k = q r 1 + q r 2 + + q + 1 and λ = q r 2 + q r 3 + + q + 1 , cf. [17]. Let us arrange the hyperplane indexes in such a way that H i = σ i H 0 and put K i = σ i K 0 . It follows that if mi denotes the intersection number |K0Hi|, we have that mi is also |K-iH0|, i.e., the number of dj in H0 satisfying dj≡-i mod r+1. Thus, we have that the family of intersection numbers {mi}, i 0 , q r + q r 1 + + q + 1 r + 1 , , r q r + q r 1 + + q + 1 r + 1 , is a multiset, cf. [18], in which the elements mi are allowed to occur more than once, cf. [19]. The multiset is specified by its underlying set {mi} and, for each element m∈{mi}, its multiplicity μ(m), a nonnegative integer. Then, we have that v i = μ m i m i for any i 0 , q r + q r 1 + + q + 1 r + 1 , , r q r + q r 1 + + q + 1 r + 1 .
Therefore, we have proved that
Lemma 5.
The at most r + 1 intersection numbers, mi, i 0 , q r + q r 1 + + q + 1 r + 1 , , r q r + q r 1 + + q + 1 r + 1 ,  with respect to the hyperplanes of a transitive ( q r + q r 1 + + q + 1 r + 1 )-set K are related to the at most r + 1 intersection numbers of any point P∈K by the equations  v i = μ m i m i .
Now, we prove that
Lemma 6.
The r + 1 intersection numbers, mi i 0 , q r + q r 1 + + q + 1 r + 1 , , r q r + q r 1 + + q + 1 r + 1 ,  with respect to the hyperplanes, satisfy the system of equations
i = 0 r m i = q r 1 + q r 2 + + q + 1 i = 0 r m i 2 = + q r 1 + q r 2 + q r 3 + + q + 1 q r + q r 1 + + q + 1 r + 1 i = 0 r m i m i j   m o d   r + 1 = q r 2 + q r 3 + + q + 1 q r + q r 1 + + q + 1 r + 1   j = 1 , , r
Proof. 
Since any hyperplane H0 is a (v,k,λ)-difference set H0={d1,…,dk}, with v = q r + q r 1 + + q + 1 , k = q r 1 + q r 2 + + q + 1 and λ = q r 2 + q r 3 + + q + 1 , then its Hall polynomial θ x = x d 1 + + x d k satisfies the congruence θ x θ x 1 k λ + λ 1 + x + + x v 1 modulo x v 1 . Thus, for any divisor r + 1 of v, one gets
θ x m 0 + m 1 x + + m r x r modulo   x r + 1 1
and
θ x θ x 1 k λ + λ v r + 1 1 + x + + x r modulo   x r + 1 1
where mi is the number of dj in H0 satisfying dj≡i mod r+1. By comparing the coefficients, the equations yield. □

4. Symmetric Sets with at Most Three Intersection Numbers

The simplest case we can consider is the plane PG(2,q) with q≡1 mod 3. Since 3 divides q2+q+1, the group σ 3 has three orbits 0 , 3 , , q 2 + q 2 , 1 , 4 , , q 2 + q 1 and 2 , 5 , , q 2 + q on PG(2,q). Since the three orbits are projectively equivalent, we will assume without loss of generality that K = 0 , 3 , , q 2 + q 2 .
Since 4 is the smallest order with q≡1 mod 3, let us consider PG(2,4). The equations of Lemma 3.4 with r = 2 and q = 4 become
x + y + z = 5 x 2 + y 2 + z 2 = 11 x y + z x + y = 7
Since z is an integer such that z 2 < 11 , it follows that z 0 , 1 , 2 , 3 .
z = 3 x + y = 2 x 2 + y 2 = 2 x y + 3 x + y = 7 x + y = 2 x 2 + y 2 = 2 x y = 1 x = 1 y = 1 z = 3 .
z = 2 x + y = 3 x 2 + y 2 = 7 x y + 2 x + y = 7 x + y = 3 x 2 + y 2 = 7 x y = 1 n o   i n t e g e r   s o l u t i o n s .
z = 1 x + y = 4 x 2 + y 2 = 10 x y + x + y = 7 x + y = 4 x 2 + y 2 = 10 x y = 3 x = 1 y = 3 z = 1 .
z = 0 x + y = 5 x 2 + y 2 = 11 x y = 7 n o   i n t e g e r   s o l u t i o n s .
Therefore, the multiset {1,1,3} is the unique solution. We explicitly construct the transitive two-intersection sets of type (1,3) of PG(2,4). The elements of the field 𝔽4 are {0,1,α,α2} where α is a root of the irreducible polynomial t2 + t + 1 in 𝔽2[t]. In order to write the cyclic structure of ℙ2: = PG(2,4), let ω be a primitive element of 𝔽43 over 𝔽4 and let p x = x 3 a 0 a 1 x a 2 x 2 be its minimal polynomial over 𝔽4. The companion matrix C p of f is given by C p = 0 1 0 0 0 1 a 0 a 1 a 2 , and it induces a Singer cycle σ of PG(2,4), cf. [14]. Let us consider a primitive polynomial of degree 3 over 𝔽4, p(x) = x3 + α + αx + αx2. The companion matrix C(p) is 0 1 0 0 0 1 5 4 0 . Let us consider the point ω 0 = x 0 , x 1 , x 2 = 1 , 0 , 0 . We get:
ω 1 = ω 0   C ( p ) = 1 , 0 , 0 0 1 0 0 0 1 α α α = 0 , 1 , 0 ,
ω 2 = ω 0   C ( f ) 2 = ω 1   C ( f ) = 0 , 1 , 0 0 1 0 0 0 1 α α α = 0 , 0 , 1 ,
ω 3 = ω 0   C ( f ) 3 = ω 2   C ( f ) = 0 , 0 , 1 0 1 0 0 0 1 α α α = α , α , α = α 1 , 1 , 1 = 1 , 1 , 1
By continuing in this way and by denoting the points represented by ωi simply by i, the points of PG(2,4) can be identified with the elements of the additive cyclic group Z 21 , the integers modulo 21 as shown in Table 1.
Since under the action of a cyclic collineation group of a finite projective plane, the point set and the line set have the same cyclic structure, select any line, for example, we choose the line 𝓁0:= x0 = 0, which contains the 5-set of points written in Table 2.
The remaining lines of the plane are found by adding 1 to each point of the preceding line beginning with 𝓁0 and using addition modulo 21.
The lines of PG(2,4) consist of the points in the vertical columns of Table 3.
The cyclic structure of PG(2,4) is well represented in Figure 1, reprinted from [20].
Let us consider the orbit of the point 0 under the action of σ 3 ,
K = 0 , 3 , 6 , 9 , 12 , 15 , 18 .
Since lines of an orbit have the same intersection number, in order to compute the intersection numbers of K, it is sufficient to consider only the three base lines written in Table 4:
Thus, the set K = 0 , 3 , 6 , 9 , 12 , 15 , 18 is a 7-set with 14 1-lines and 7 3-lines, i.e., K is of type (1,3), a Baer subplane, as was noted by Brouwer in [10], see also [6]. By the projective equivalence of points and lines, K is a symmetric 7-set, i.e., any point P has v1(P) = 2·1 = 2 and v3(P) = 1·3 = 3.
The second order q with q≡1 mod 3 is 7; let us consider the field 𝔽7. The equations of Lemma 3.2 with k − 1 = 2 and q = 7 become
x + y + z = 8 x 2 + y 2 + z 2 = 26 x y + x z + y z = 19
Since z is an integer such that z 2 < 26 , it follows that z 0 , 1 , 2 , 3 , 4 , 5 .
z = 5 x + y = 3 x 2 + y 2 = 1 x y + 5 x + y = 19 x + y = 3 x 2 + y 2 = 1 x y = 4 n o   i n t e g e r   s o l u t i o n s .
z = 4 x + y = 4 x 2 + y 2 = 10 x y + 4 x + y = 19 x + y = 4 x 2 + y 2 = 10 x y = 3 x = 1 y = 3 z = 4 .
z = 3 x + y = 5 x 2 + y 2 = 17 x y + 3 x + y = 19 x + y = 5 x 2 + y 2 = 17 x y = 4 x = 1 y = 4 z = 3 .
z = 2 x + y = 6 x 2 + y 2 = 22 x y + 2 x + y = 19 x + y = 6 x 2 + y 2 = 22 x y = 7 n o   i n t e g e r   s o l u t i o n s .
z = 1 x + y = 7 x 2 + y 2 = 25 x y + x + y = 19 x + y = 7 x 2 + y 2 = 25 x y = 12 x = 3 y = 4 z = 1 .
z = 0 x + y = 8 x 2 + y 2 = 26 x y = 19 n o   i n t e g e r   s o l u t i o n s .
Therefore, (1,3,4) is the unique solution. We explicitly construct the transitive three-intersection sets of type (1,3,4) of PG(2,7). In 𝔽7, 3 is a root of x2+x+2. In order to write the cyclic structure of ℙ2 :=PG(2,7), let ω be a primitive element of 𝔽73 over 𝔽7 and let p x = x 3 a 0 a 1 x a 2 x 2 be its minimal polynomial over 𝔽7. The companion matrix C p of f is given by C p = 0 1 0 0 0 1 a 0 a 1 a 2 , and it induces a Singer cycle σ of PG(2,7), cf. [14]. Let us consider a primitive polynomial with minimal weight, i.e., the minimal number of nonzero coefficients, among all primitives of that degree over 𝔽7, p(x) = x3 + 2 + 3x, cf. [21]. The companion matrix C(p) is 0 1 0 0 0 1 5 4 0 . Let us consider the point ω 0 = x 0 , x 1 , x 2 = 1 , 0 , 0 .
We get:
ω 1 = ω 0   C ( f ) =   1 , 0 , 0 0 1 0 0 0 1 5 4 0 = 0 , 1 , 0 ,
ω 2 = ω 0   C ( f ) 2 = ω 1   C ( f ) = 0 , 1 , 0 0 1 0 0 0 1 5 4 0 = 0 , 0 , 1 ,
ω 3 = ω 0   C ( f ) 3 = ω 2   C ( f ) = 0 , 0 , 1 0 1 0 0 0 1 5 4 0 = 5 , 4 , 0 = 5 1 , 5 , 0 = 1 , 5 , 0
By continuing in this way and by denoting the points represented by ωi simply by i, we obtain the cyclic structure of PG(2,7), as shown in Table 5. Thus, the Singer group is isomorphic to the additive group Z57, the integers modulo 57.
Since under the action of a cyclic collineation group of a finite projective plane the point set and the line set have the same cyclic structure, select any line. For example, we choose the line 𝓁0: = x0 = 0, which contains the 8-set of points written in Table 6.
The remaining lines of the plane are found by adding 1 to each point of the preceding line beginning with 𝓁0 and using addition modulo 57. Let us consider the orbit of the point 0 under the action of σ 3 ,
K = 0 , 3 , 6 , 9 , 12 , 15 , 18 , 21 , 24 , 27 , 30 , 33 , 36 , 39 , 42 , 45 , 48 , 51 , 54 .
Since lines of an orbit have the same intersection number, in order to compute the intersection numbers of K, it is sufficient to consider only the three base lines written in Table 7:
Thus, the set
K = 0 , 3 , 6 , 9 , 12 , 15 , 18 , 21 , 24 , 27 , 30 , 33 , 36 , 39 , 42 , 45 , 48 , 51 , 54
is a 19-set with 19 1-lines, 19 3-lines and 19 4-lines, i.e., K is of type (1,3,4), the largest minimal blocking set of PG(2,7). By the projective equivalence of points and lines, K is a symmetric 19-set, i.e., any point P has v1(P) = 1, v3(P) = 3 and v4(P) = 4.

5. Symmetric Sets with at Most Four Intersection Numbers

The second case we can consider is the space PG(3,q) with q ≡ 1 mod 4. Since 4 divides q3 + q2 + q + 1, the group σ 4 has four orbits 0 , 4 , , q 3 + q 2 + q 3 , 1 , 5 , , q 3 + q 2 + q 2 , 2 , 6 , , q 3 + q 2 + q 1 and 3 , 7 , , q 3 + q 2 + q on PG(3,q), each of size q 3 + q 2 + q + 1 4 . Since the four orbits are projectively equivalent, we will assume without loss of generality that K = 0 , 4 , , q 3 + q 2 + q 3 . Since 5 is the smallest order with q≡1 mod 4, let us consider PG(3,5). The equations of Lemma 3.3 with r = 3 and q = 5 become
x + z + y + t = 31 x 2 + y 2 + z 2 + t 2 = 259 x + z y + t = 234 2 x z + y t = 234   x + z = 13 z x 2 + t y 2 = 25 y + t = 18 x z + y t = 117   x + z = 13 z x = 3 y + t = 18 t y = 4   x = 5 y = 7 z = 8 t = 11
Therefore, (5,7,8,11) is the unique solution. We explicitly construct the transitive four-intersection sets of type (5,7,8,11) of PG(3,5). In order to write the cyclic structure of ℙ3: = PG(3,5), let ω be a primitive element of 𝔽54 over 𝔽5 and let p x = x 4 a 0 a 1 x a 2 x 2 a 3 x 3 be its minimal polynomial over 𝔽5. The companion matrix C p of f is given by C p = 0 1 0 0 0 0 1 0 0 a 0 0 a 1 0 a 2 1 a 3 , and it induces a Singer cycle σ of PG(3,5), cf. [14]. Let us consider a primitive polynomial with minimal weight, i.e., the minimal number of nonzero coefficients, among all primitives of that degree over 𝔽5, p(x) = x4 + 2 + 2x + x2, cf. [21]. The companion matrix C(p) is 0 1 0 0 0 0 1 0 0 3 0 3 0 4 1 0 .
Let us consider the point ω 0 = x 0 , x 1 , x 2 , x 3 = 1 , 0 , 0 , 0 . We get:
ω 1 = ω 0 C ( f ) = 1 , 0 , 0 , 0 0 1 0 0 0 0 1 0 0 3 0 3 0 4 1 0 = 0 , 1 , 0 , 0 ,
ω 2 = ω 0 C ( f ) 2 = ω 1 C ( f ) = 0 , 1 , 0 , 0 0 1 0 0 0 0 1 0 0 3 0 3 0 4 1 0 = 0 , 0 , 1 , 0 ,
ω 3 = ω 0 C ( f ) 3 = ω 2 C ( f ) = 0 , 0 , 1 , 0 0 1 0 0 0 0 1 0 0 3 0 3 0 4 1 0 = 0 , 0 , 0 , 1
ω 4 = ω 0 C ( f ) 4 = ω 3 C ( f ) = 0 , 0 , 0 , 1 0 1 0 0 0 0 1 0 0 3 0 3 0 4 1 0 = 3 , 3 , 4 , 0 = 2 3 , 3 , 4 , 0 = 1 , 1 , 3 , 0
By continuing in this way and by denoting the points represented by ωi simply by i, we obtain the cyclic structure of PG(3,5), as shown in Table 8. Thus, the Singer group is isomorphic to the additive group Z156, the integers modulo 156.
Since under the action of a cyclic collineation group of a finite projective space the point set and the plane set have the same cyclic structure, select any plane. For example, we choose the plane π0: = x0=0, which contains the 31-set of points written in Table 9.
The remaining planes of the space are found by adding 1 to each point of the preceding plane beginning with π0 and using addition modulo 156. Let us consider the orbit of the point 0 under the action of σ 4 ,
K = 0 , 4 , 8 , , 152 .
Since planes of an orbit have the same intersection number, in order to compute the intersection numbers of K, it is sufficient to consider only the four base planes written in Table 10:
Thus, the set
K = 0 , 4 , 8 , , 152
is a 39-set with 39 5-planes, 39 7-planes, 39 8-planes and 39 11-planes, i.e., K is of type (5,7,8,11). By the projective equivalence of points and planes, K is a symmetric 39-set, i.e., any point P has v5(P) = 5, v7(P) = 7, v8(P) = 8 and v11(P) = 11.

6. Conclusions

In this work, we investigated symmetric sets in finite projective spaces. Symmetric sets are interesting not only in themselves but also through becoming associated with objects in which an interest already exists, such as sets with few intersection numbers, cf. [7]. Symmetry of sets of class [0,m,n] means that the family of symmetric sets is very large, including caps and two-character sets widely studied in the literature to construct projective linear codes with two weights, cf. [4,5]. We provided necessary conditions for the existence of such sets which, to our knowledge, seem to be novel and involve the number of external hyperplanes through the points nonbelonging to the set. Since caps and two-character sets have been extensively studied, we directed our attention to transitive sets which represent another large class of symmetric sets. We generalized the construction of Brouwer [10], derived a sufficient condition for the existence of such sets and determined enough information on their intersection numbers. After reporting the cyclic construction for the sake of completeness, we reproduced and verified examples that can be calculated by hand. All these examples together seem to reveal a close and interesting relationship between several geometric and combinatorial structures. We expect that the computer-literate reader can extend the techniques used to larger cases.

Funding

This research received no external funding.

Data Availability Statement

Data are contained within the article.

Acknowledgments

This research was performed within the activity of GNSAGA of the Italian INDAM.

Conflicts of Interest

The author declares no conflicts of interest.

References

  1. Tallini Scafati, M. The k-set of PG(r,q) from the character point of view. In Finite Geometries; Baker, C.A., Batten, L.M., Eds.; Marcel Dekker Inc.: New York, NY, USA, 1985; pp. 321–326. [Google Scholar]
  2. Napolitano, V. Note on sets of hyperplane-type (m,n)r−1 in PG(r,q). Australas. J. Comb. 2022, 82, 197–200. Available online: https://ajc.maths.uq.edu.au/pdf/82/ajc_v82_p197.pdf (accessed on 3 January 2025).
  3. Hamilton, N.; Penttila, T. Sets of Type (a,b) From Subgroups of ΓL(1,pR). J. Algebr. Combin. 2001, 13, 67–76. [Google Scholar] [CrossRef]
  4. Delsarte, P. Weights of linear codes and strongly regular normed spaces. Discret. Math. 1972, 3, 47–64. [Google Scholar] [CrossRef]
  5. Calderbank, R.; Kantor, W.M. The geometry of two-weight codes. Bull. Lond. Math. Soc. 1986, 18, 97–122. [Google Scholar] [CrossRef]
  6. Coykendall, J.; Dover, J. Sets with few Intersection Numbers from Singer Subgroup Orbits. Eur. J. Combin. 2001, 22, 455–464. [Google Scholar] [CrossRef]
  7. Durante, N. On Sets with Few Intersection Numbers in Finite Projective and Affine Spaces. Electron. J. Comb. 2014, 21, P4.13. [Google Scholar] [CrossRef] [PubMed]
  8. Tondini, D. (11,3)-arc configurations in PG(2,5). J. Interdiscip. Math. 2023, 26, 1717–1731. [Google Scholar] [CrossRef]
  9. Retkin, H.; Stein, E. On the symmetric (15,3)-arcs of the finite projective plane of order 7. Rend. Mat. 1965, 24, 392–399. [Google Scholar]
  10. Brouwer, A.E. A series of separable designs with application to pairwise orthogonal Latin squares. Eur. J. Combin. 1980, 1, 137–146. [Google Scholar] [CrossRef]
  11. Hirschfeld, J.W.P. Projective Geometries over Finite Fields, 2nd ed.; Oxford University Press: Oxford, UK, 1998. [Google Scholar]
  12. Pavese, F. Geometric constructions of two-character sets. Discret. Math. 2015, 338, 202–208. [Google Scholar] [CrossRef]
  13. Innamorati, S. Classifying Sets of Type (4,n) in PG(3,q). Foundations 2024, 4, 263–272. [Google Scholar] [CrossRef]
  14. Singer, J.W. A theorem in finite projective geometry and some applications to number theory. Trans. Am. Math. Society 1938, 43, 377–385. [Google Scholar] [CrossRef]
  15. Davydov, A.A.; Giulietti, M.; Marcugini, S.; Pambianco, F. On sharply transitive sets in PG(2,q). Innov. Incid. Geom. 2009, 6–7, 139–151. [Google Scholar] [CrossRef]
  16. Eberhard, S. Sharply transitive sets in PGL2(K). Adv. Geom. 2021, 21, 611–612. [Google Scholar] [CrossRef]
  17. Baumert, L.D. Cyclic Difference Sets; Volume 182 of Lecture Notes in Mathematics; Springer: Berlin/Heidelberg, Germany, 1971. [Google Scholar]
  18. Monro, G.P. The concept of multiset. Math. Log. Q. 1987, 33, 171–178. [Google Scholar] [CrossRef]
  19. Blizard, W.D. Multiset theory. Notre Dame J. Form. Log. 1989, 30, 36–66. [Google Scholar] [CrossRef]
  20. Wagner, A.W. Finite Projective Planes—PPT Download Slide 10/16. Available online: https://slideplayer.com/slide/3116502/ (accessed on 3 January 2025).
  21. Hansen, T.; Mullen, G.L. Primitive polynomials over finite fields. Math. Comp. 1992, 59, 639–643. [Google Scholar] [CrossRef]
Figure 1. The cyclic structure of PG(2,4).
Figure 1. The cyclic structure of PG(2,4).
Symmetry 17 00179 g001
Table 1. The cyclic structure of the points of PG(2,4).
Table 1. The cyclic structure of the points of PG(2,4).
ω0 = (1,0,0) = 0
ω1 = (0,1,0) = 1ω2 = (0,0,1) = 2ω3 = (1,1,1) = 3ω4 = (1,α, α) = 4
ω5 = (1,α2,α) = 5ω6 = (1,α2,0) = 6ω7 = (0,1,α2) = 7ω8 = (1,1,0) = 8
ω9 = (0,1,1) = 9ω10 = (1,1,α) = 10ω11 = (1,α2,α2) = 11ω12 = (1,0,α) = 12
ω13 = (1,α2,1) = 13ω14 = (1,α,α2) = 14ω15 = (1,0,α2) = 15ω16 = (1,0,1) = 16
ω17 = (1,α,1) = 17ω18 = (1,α,0) = 18ω19 = (0,1,α) = 19ω20 = (1,1,α2) = 20
Table 2. The starting line.
Table 2. The starting line.
𝓁0127919
Table 3. The lines of PG(2,4).
Table 3. The lines of PG(2,4).
12345678910111213141516171819200
23456789101112131415161718192001
78910111213141516171819200123456
91011121314151617181920012345678
19200123456789101112131415161718
Table 4. The base lines.
Table 4. The base lines.
𝓁0127919
𝓁12381020
𝓁2034911
Table 5. The cyclic structure of the points of PG(2,7).
Table 5. The cyclic structure of the points of PG(2,7).
ω0 = (1,0,0) = 0
ω1 = (0,1,0) = 1ω2 = (0,0,1) = 2ω3 = (1,5,0) = 3ω4 = (0,1,5) = 4
ω5 = (1,5,2) = 5ω6 = (1,3,4) = 6ω7 = (1,4,4) = 7ω8 = (1,4,3) = 8
ω9 = (1,6,4) = 9ω10 = (1,4,1) = 10ω11 = (1,1,5) = 11ω12 = (1,0,2) = 12
ω13 = (1,3,0) = 13ω14 = (0,1,3) = 14ω15 = (1,5,1) = 15ω16 = (1,1,1) = 16
ω17 = (1,1,3) = 17ω18 = (1,6,1) = 18ω19 = (1,1,4) = 19ω20 = (1,4,6) = 20
ω21 = (1,2,2) = 21ω22 = (1,3,3) = 22ω23 = (1,6,3) = 23ω24 = (1,6,6) = 24
ω25 = (1,2,3) = 25ω26 = (1,6,2) = 26ω27 = (1,3,2) = 27ω28 = (1,3,1) = 28
ω29 = (1,1,2) = 29ω30 = (1,3,5) = 30ω31 = (1,0,6) = 31ω32 = (1,2,0) = 32
ω33 = (0,1,2) = 33ω34 = (1,5,5) = 34ω35 = (1,0,3) = 35ω36 = (1,6,0) = 36
ω37 = (0,1,6) = 37ω38 = (1,5,4) = 38ω39 = (1,4,2) = 39ω40 = (1,3,6) = 40
ω41 = (1,2,5) = 41ω42 = (1,0,4) = 42ω43 = (1,4,0) = 43ω44 = (0,1,4) = 44
ω45 = (1,5,6) = 45ω46 = (1,2,6) = 46ω47 = (1,2,1) = 47ω48 = (1,1,6) = 48
ω49 = (1,2,4) = 49ω50 = (1,4,5) = 50ω51 = (1,0,1) = 51ω52 = (1,1,0) = 52
ω53 = (0,1,1) = 53ω54 = (1,5,3) = 54ω55 = (1,6,5) = 55ω56 = (1,0,5) = 56
Table 6. The starting line.
Table 6. The starting line.
𝓁01241433374453
Table 7. The base lines, with in bold the elements of K.
Table 7. The base lines, with in bold the elements of K.
𝓁01241433374453
𝓁12351534384554
𝓁23461635394655
Table 8. The cyclic structure of the points of PG(3,5).
Table 8. The cyclic structure of the points of PG(3,5).
ω0 = (1,0,0,0) = 0
ω1 = (0,1,0,0) = 1ω2 = (0,0,1,0) = 2ω3 = (0,0,0,1) = 3ω4 = (1,1,3,0) = 4ω5 = (0,1,1,3) = 5
ω6 = (1,1,2,4) = 6ω7 = (1,4,1,1) = 7ω8 = (1,3,1,2) = 8ω9 = (1,2,1,1) = 9ω10 = (1,3,2,2) = 10
ω11 = (1,2,1,2) = 11ω12 = (1,2,0,1) = 12ω13 = (1,3,2,0) = 13ω14 = (0,1,3,2) = 14ω15 = (1,1,4,3) = 15
ω16 = (1,0,2,1) = 16ω17 = (1,3,3,4) = 17ω18 = (1,4,2,4) = 18ω19 = (1,4,0,1) = 19ω20 = (1,3,1,0) = 20
ω21 = (0,1,3,1) = 21ω22 = (1,1,0,1) = 22ω23 = (1,3,0,0) = 23ω24 = (0,1,3,0) = 24ω25 = (0,0,1,3) = 25
ω26 = (1,1,3,4) = 26ω27 = (1,4,1,4) = 27ω28 = (1,4,0,3) = 28ω29 = (1,0,4,0) = 29ω30 = (0,1,0,4) = 30
ω31 = (1,1,1,0) = 31ω32 = (0,1,1,1) = 32ω33 = (1,1,0,2) = 33ω34 = (1,2,4,0) = 34ω35 = (0,1,2,4) = 35
ω36 = (1,1,1,1) = 36ω37 = (1,3,0,2) = 37ω38 = (1,2,1,0) = 38ω39 = (0,1,2,1) = 39ω40 = (1,1,0,4) = 40
ω41 = (1,4,1,0) = 41ω42 = (0,1,4,1) = 42ω43 = (1,1,0,3) = 43ω44 = (1,0,2,0) = 44ω45 = (0,1,0,2) = 45
ω46 = (1,1,4,0) = 46ω47 = (0,1,1,4) = 47ω48 = (1,1,1,3) = 48ω49 = (1,0,2,4) = 49ω50 = (1,4,3,1) = 50
ω51 = (1,3,1,1) = 51ω52 = (1,3,4,2) = 52ω53 = (1,2,1,4) = 53ω54 = (1,4,4,3) = 54ω55 = (1,0,4,1) = 55
ω56 = (1,3,3,3) = 56ω57 = (1,0,0,2) = 57ω58 = (1,2,3,0) = 58ω59 = (0,1,2,3) = 59ω60 = (1,1,2,3) = 60
ω61 = (1,0,2,3) = 61ω62 = (1,0,3,3) = 62ω63 = (1,0,3,2) = 63ω64 = (1,2,3,3) = 64ω65 = (1,0,1,2) = 65
ω66 = (1,2,3,1) = 66ω67 = (1,3,2,1) = 67ω68 = (1,3,4,4) = 68ω69 = (1,4,2,2) = 69ω70 = (1,2,2,2) = 70
ω71 = (1,2,0,2) = 71ω72 = (1,2,0,0) = 72ω73 = (0,1,2,0) = 73ω74 = (0,0,1,2) = 74ω75 = (1,1,3,1) = 75
ω76 = (1,3,0,1) = 76ω77 = (1,3,4,0) = 77ω78 = (0,1,3,4) = 78ω79 = (1,1,1,4) = 79ω80 = (1,4,1,3) = 80
ω81 = (1,0,4,4) = 81ω82 = (1,4,3,2) = 82ω83 = (1,2,2,3) = 83ω84 = (1,0,1,3) = 84ω85 = (1,0,3,4) = 85
ω86 = (1,4,3,4) = 86ω87 = (1,4,0,4) = 87ω88 = (1,4,0,0) = 88ω89 = (0,1,4,0) = 89ω90 = (0,0,1,4) = 90
ω91 = (1,1,3,3) = 91ω92 = (1,0,2,2) = 92ω93 = (1,2,3,2) = 93ω94 = (1,2,0,3) = 94ω95 = (1,0,1,0) = 95
ω96 = (0,1,0,1) = 96ω97 = (1,1,0,0) = 97ω98 = (0,1,1,0) = 98ω99 = (0,0,1,1) = 99ω100 = (1,1,3,2) = 100
ω101 = (1,2,4,3) = 101ω102 = (1,0,1,1) = 102ω103 = (1,3,3,2) = 103ω104 = (1,2,1,3) = 104ω105 = (1,0,1,4) = 105
ω106 = (1,4,3,3) = 106ω107 = (1,0,4,2) = 107ω108 = (1,2,3,4) = 108ω109 = (1,4,4,4) = 109ω110 = (1,4,0,2) = 110
ω111 = (1,2,2,0) = 111ω112 = (0,1,2,2) = 112ω113 = (1,1,4,2) = 113ω114 = (1,2,4,4) = 114ω115 = (1,4,4,2) = 115
ω116 = (1,2,2,4) = 116ω117 = (1,4,4,1) = 117ω118 = (1,3,1,3) = 118ω119 = (1,0,0,4) = 119ω120 = (1,4,3,0) = 120
ω121 = (0,1,4,3) = 121ω122 = (1,1,2,1) = 122ω123 = (1,3,0,4) = 123ω124 = (1,4,2,0) = 124ω125 = (0,1,4,2) = 125
ω126 = (1,1,4,4) = 126ω127 = (1,4,1,2) = 127ω128 = (1,2,2,1) = 128ω129 = (1,3,2,4) = 129ω130 = (1,4,2,1) = 130
ω131 = (1,3,1,4) = 131ω132 = (1,4,2,3) = 132ω133 = (1,0,4,3) = 133ω134 = (1,0,3,1) = 134ω135 = (1,3,3,1) = 135
ω136 = (1,3,4,1) = 136ω137 = (1,3,4,3) = 137ω138 = (1,0,0,1) = 138ω139 = (1,3,3,0) = 139ω140 = (0,1,3,3) = 140
ω141 = (1,1,2,2) = 141ω142 = (1,2,4,2) = 142ω143 = (1,2,0,4) = 143ω144 = (1,4,4,0) = 144ω145 = (0,1,4,4) = 145
ω146 = (1,1,1,2) = 146ω147 = (1,2,4,1) = 147ω148 = (1,3,2,3) = 148ω149 = (1,0,0,3) = 149ω150 = (1,0,3,0) = 150
ω151 = (0,1,0,3) = 151ω152 = (1,1,2,0) = 152ω153 = (0,1,1,2) = 153ω154 = (1,1,4,1) = 154ω155 = (1,3,0,3) = 155
Table 9. The starting plane.
Table 9. The starting plane.
π01235142124253032353942454759
7374788990969899112121125140145151153
Table 10. The base planes, with in bold the elements of K.
Table 10. The base planes, with in bold the elements of K.
π01235142124253032353942454759
7374788990969899112121125140145151153
π12346152225263133364043464860
74757990919799100113122126141146152154
π23457162326273234374144474961
757680919298100101114123127142147153155
π3045681724272833353842454850
62767781929399101102115124128143148154
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Innamorati, S. Symmetric Point Sets with Few Intersection Numbers in PG(r,q). Symmetry 2025, 17, 179. https://doi.org/10.3390/sym17020179

AMA Style

Innamorati S. Symmetric Point Sets with Few Intersection Numbers in PG(r,q). Symmetry. 2025; 17(2):179. https://doi.org/10.3390/sym17020179

Chicago/Turabian Style

Innamorati, Stefano. 2025. "Symmetric Point Sets with Few Intersection Numbers in PG(r,q)" Symmetry 17, no. 2: 179. https://doi.org/10.3390/sym17020179

APA Style

Innamorati, S. (2025). Symmetric Point Sets with Few Intersection Numbers in PG(r,q). Symmetry, 17(2), 179. https://doi.org/10.3390/sym17020179

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop