Next Article in Journal
Topological Indices of Graphs from Vector Spaces
Next Article in Special Issue
Periodically Intermittent Control of Memristor-Based Hyper-Chaotic Bao-like System
Previous Article in Journal
On the Use of Variability Measures to Analyze Source Coding Data Based on the Shannon Entropy
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Constructions of Goethals–Seidel Sequences by Using k-Partition

School of Mathematical Sciences, University of Electronic Science and Technology of China, Chengdu 611731, China
*
Author to whom correspondence should be addressed.
Mathematics 2023, 11(2), 294; https://doi.org/10.3390/math11020294
Submission received: 15 December 2022 / Revised: 26 December 2022 / Accepted: 1 January 2023 / Published: 6 January 2023

Abstract

:
In this paper, we are devoted to finding Goethals–Seidel sequences by using k-partition, and based on the finite Parseval relation, the construction of Goethals–Seidel sequences could be transformed to the construction of the associated polynomials. Three different structures of Goethals–Seidel sequences will be presented. We first propose a method based on T-matrices directly to obtain a quad of Goethals–Seidel sequences. Next, by introducing the k-partition, we utilize two classes of 8-partitions to obtain a new class of polynomials still remaining the same (anti)symmetrical properties, with which a quad of Goethals–Seidel sequences could be constructed. Moreover, an adoption of the 4-partition together with a quad of four symmetrical sequences can also lead to a quad of Goethals–Seidel sequences.

1. Introduction

Hadamard matrices (HMs) are applied to many fields, such as cryptography, coding theory and signal processing, and many works are devoted to studying the properties and constructions of them. HMs H n of order n are the square matrices with entries ± 1 that satisfy H n H n T = n I n , where H n T is the transpose of H n and I n is the identity matrix. It is necessary that the order n > 2 is always divisible by 4. There are many references about HMs and how to construct them, such as [1,2,3,4,5,6,7,8]. According to the definition of HMs, we consider them as block structures. Let an HM be H n = [ ( H i j ) m ] , where n and m are the orders satisfying n = k m with i , j = 1 , 2 , , k , and the block matrices H i j satisfy the following conditions:
( i ) j H i 1 j H i 2 j T = n I m , i 1 = i 2 , ( ii ) j H i 1 j H i 2 j T = 0 , i 1 i 2 .
One alternative method to obtain these matrix blocks is making use of the circulant matrices, such as Williamson array, Goethals–Seidel (GS) array, Kharaghani array, and so on [6,9].
In the earlier works on constructing HMs using circulant matrix blocks, it was found that some skew-HMs did not exist, even if for small orders, such as 36 [6]. To overcome this difficulty, Goethals and Seidel [10] designed the GS array
G = A B R C R D R B R A D T R C T R C R D T R A B T R D R C T R B T R A ,
where A , B , C , D are four circulant matrices, and R denotes the back-diagonal identity matrix. Obviously the block matrices of GS array (2) satisfy the condition (1). Then, it is meaningful to investigate this array in the construction of HMs. In addition, compared with the Williamson array, the GS array has no requirements of symmetries, which will be relatively friendly to constructions. In this paper, we mainly focus on the construction of GS array and transform it into the construction of first rows of circulant matrices A , B , C , D through the finite Parseval relation [11]. In particular, we call first rows of these circulant matrices as a quad of GS sequences.
Goethals and Seidel [10] studied the HMs of GS type of order 4 n , where they obtained a quad of sequences with order n = 9 , 13 , and leave n = 23 the only unconstructed order less than 25. Later, Doković [12,13] searched them by computer and obtained the GS array of order 4 n , n = 37 , 43 , 49 , , 163 . More orders of GS array about Doković’s results can be referred to [14,15,16,17,18,19]. In [20], Fletcher et al. found a GS array of order 4 n = 36 . Furthermore, in [21,22], a family of GS sequences with order n = q + 1 was constructed by using the Parseval relation in Galois field G F ( q 2 ) theoretically, where q 3 ( m o d 8 ) is a prime power. In addition, using T-matrices or T-sequences can also lead to GS sequences. In [23], Yang discovered that, if a quad of Williamson sequences of order n and T-sequences of order m exist, then a quad of GS sequences of order m n also exists. Yang also presented some new infinite families of GS sequences by utilizing T-sequences and Williamson sequences in [24,25,26,27]. By the way, Williamson sequences can be considered as a special case of GS sequences, and Whiteman [11] obtained a quad of Williamson sequences of order ( q + 1 ) / 2 for q 1 ( m o d 4 ) being a prime power.
In the literature mentioned above, two sequences used in [11] are actually a 2-partition. Four sequences in [21,22] and Yang’s works are based on a 4-partition or the combination of 4-partitions. Thus, it is natural to study whether more GS sequences could be constructed by more partitions. In this paper, we mainly focus on the structures of GS sequences by using symmetrical and antisymmetrical k-partition, and eventually obtain three quads of GS sequences with different structures. There are two main reasons why we adopt symmetrical and antisymmetrical k-partition: it facilitates for us to reduce the range of searching k-partitions, and a k-partition with these properties is friendly to the construction. Then, the first method is based on the T-matrices, and a quad of GS sequences could be constructed directly, including the Williamson sequences as the special cases. The second way is utilizing two classes of 8-th partitions, where half are symmetrical and others are antisymmetrical, to generate a new class of polynomials. In this process, this class of polynomials would remain the same properties of symmetry and antisymmetry as two original classes of 8-partitions, and could lead to a quad of GS sequences too. Finally, we employ the 4-partition as well as a quad of Williamson sequences to obtain a quad of GS sequences. Since constructing Hadamard matrices of GS type are now transformed into finding GS sequences, we in this paper offer some alternative approaches, which include finding some appropriate 8-partitions or novel 4-partitions instead of T-matrix sequences.
The rest of this paper is organized as follows: In Section 2, we mainly introduce some notations and definitions used later. In Section 3, the definition of k-partition will be given, with which we rigorously prove that 8-partitions or the combination of 4-partition and Williamson sequences could lead to GS sequences. Some conclusions will be drawn in Section 4.

2. Preliminaries

2.1. Parseval Relation

Let a = ( a 0 , a 1 , , a n 1 ) be a sequence, whose periodic autocorrelation function R a ( τ ) is defined as
R a ( τ ) = i = 0 n 1 a i a ¯ i + τ , τ = 0 , 1 , , n 1 ,
where a ¯ i is the conjugate of a i , and the subscript i + τ is evaluated modulo-n. The polynomial
Φ a ( ξ ) = a 0 + a 1 ξ + a 2 ξ 2 + + a n 1 ξ n 1
is called the associated polynomial of sequence a , where ξ is the n-th root of unity. The finite Parseval relation [11] (also the Wiener–Khinchin theorem [28,29]) between R a ( τ ) and Φ a ( ξ ) is presented in the following identity:
R a ( τ ) = 1 n j = 0 n 1 Φ a ( ξ j ) 2 ξ j τ , τ = 0 , 1 , , n 1 ,
with its inverse form
Φ a ( ξ j ) 2 = Φ a ( ξ j ) Φ a ( ξ j ) ¯ = τ = 0 n 1 R a ( τ ) ξ j τ , j = 0 , 1 , , n 1 .
For the HMs in the form of GS type (2), four circulant matrices therein have the following property.
Lemma 1 
([11]). Let A , B , C and D denote the four circulant matrices of order n whose first rows are four sequences a = { a i } i = 0 n 1 , b = { b i } i = 0 n 1 , c = { c i } i = 0 n 1 and d = { d i } i = 0 n 1 , respectively. Then, A A T + B B T + C C T + D D T = 4 n I n if and only if
Φ a ( ξ j ) 2 + Φ b ( ξ j ) 2 + Φ c ( ξ j ) 2 + Φ d ( ξ j ) 2 = 4 n ,
where ξ is the n-th root of unity and j = 0 , 1 , , n 1 .
Proof. 
It follows immediately from (1) and (6). □
Remark 1. 
By Lemma 1, the relationship is now transformed from the circulant matrices to the associated polynomials.
In the later statements, when the polynomials, sequences, and the coefficients appear in the same place, without special clarifications, we denote by the capital letter, for example, F i ( ξ ) polynomials, the bold letter f i sequences and the lower case letter f i j corresponding coefficients, where i and j rely on different cases.

2.2. GS Sequences

Definition 1 
(GS sequences, [23]). Four ± 1 sequences q i = ( q i 0 , q i 1 , , q i , n 1 ) , i = 1 , 2 , 3 , 4 are said to be a quad of GS sequences, if their associated polynomials Q i ( ξ ) = q i 0 + q i 1 ξ + + q i , n 1 ξ n 1 satisfy
i = 1 4 Q i ( ξ ) 2 = 4 n ,
where ξ is the n-th root of unity.
As a special case of the GS sequences, a quad of Williamson sequences will be defined analogously, which requires the symmetry additionally.
Definition 2 
(Williamson sequences). Four ± 1 sequences w i = ( w i 0 , , w i , n 1 ) , i = 1 , 2 , 3 , 4 , are said to be a quad of Williamson sequences, if their associated polynomials W i ( ξ ) = w i 0 + w i 1 ξ + + w i , n 1 ξ n 1 satisfy
i = 1 4 W i ( ξ ) 2 = 4 n , W i ( ξ ) ¯ = W i ( ξ ) ,
where ξ is the n-th root of unity.

2.3. T-Matrices and T-Matrix Sequences

Definition 3 
(T-matrices, [6,30]). Four ( 0 , 1 , 1 ) circulant matrices T 1 , T 2 , T 3 and T 4 of order m are T-matrices if they satisfy the following conditions:
( i ) T i * T j T = 0 , i j a n d i , j = 1 , 2 , 3 , 4 , ( ii ) i = 1 4 T i T i T = m I m .
where * denotes the Hadamard product.
According to the definition of T-matrices, we define a quad of T-matrix sequences.
Definition 4 
(T-matrix sequences). Let t i = ( t i 0 , t i 1 , , t i , m 1 ) , i = 1 , 2 , 3 , 4 be the first rows of T-matrices satisfying i = 1 4 | t i j | = 1 , j = 0 , , m 1 . Then, we call t i as a quad of T-matrix sequences (TMS).
We give two examples of TMS: for n = 6 ,
t 1 = ( 1 , 0 , 0 , 0 , 0 , 0 ) , t 2 = ( 0 , 0 , 1 , 0 , 1 , 0 ) , t 3 = ( 0 , 0 , 0 , 0 , 0 , 0 ) , t 4 = ( 0 , 1 , 0 , 1 , 0 , 1 ) ,
and for n = 8 ,
t 1 = ( 1 , 0 , 0 , 0 , 1 , 0 , 0 , 0 ) , t 2 = ( 0 , 1 , 0 , 1 , 0 , 1 , 0 , 1 ) , t 3 = ( 0 , 0 , 1 , 0 , 0 , 0 , 1 , 0 ) , t 4 = ( 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 ) .
Let T i ( ξ ) = t i 0 + t i 1 ξ + + t i , m 1 ξ m 1 be the associated polynomials of sequences t i , i = 1 , 2 , 3 , 4 , with ξ being the m-th root of unity. Then, from Definition 3 and Definition 4, we have the following results.
Lemma 2. 
Let polynomials T 1 ( ξ ) , T 2 ( ξ ) , T 3 ( ξ ) and T 4 ( ξ ) be the associated polynomials of TMS t i = ( t i 0 , , t i , m 1 ) , i = 1 , 2 , 3 , 4 , and then the sum i = 1 4 ( ± T i ( ξ ) ) determines a polynomial whose coefficients are ± 1 . Moreover, it holds that
T 1 ( ξ ) 2 + T 2 ( ξ ) 2 + T 3 ( ξ ) 2 + T 4 ( ξ ) 2 = m ,
where ξ is the m-th root of unity.
Proof. 
It follows directly from the Parseval relation (5) and Definition 4 of TMS. □
A quad of GS sequences could be obtained directly by utilizing TMS.
Lemma 3 
([31]). Let T 1 ( ξ ) , T 2 ( ξ ) , T 3 ( ξ ) , T 4 ( ξ ) of order m be the associated polynomials of TMS t 1 , t 2 , t 3 , t 4 . The coefficients of the following four polynomials:
Q 1 ( ξ ) = T 1 ( ξ ) + T 2 ( ξ ) + T 3 ( ξ ) + T 4 ( ξ ) , Q 2 ( ξ ) = T 1 ( ξ ) T 2 ( ξ ) + T 3 ( ξ ) + T 4 ( ξ ) , Q 3 ( ξ ) = T 1 ( ξ ) + T 2 ( ξ ) T 3 ( ξ ) + T 4 ( ξ ) , Q 4 ( ξ ) = T 1 ( ξ ) + T 2 ( ξ ) + T 3 ( ξ ) T 4 ( ξ ) ,
i.e., q 1 , q 2 , q 3 , q 4 , are a quad of GS sequences, where ξ is the m-th root of unity.
Now, we present a novel construction of a quad of GS sequences different from the method by using TMS in Lemma 3.
Theorem 1. 
For four sequences h i = ( h i 0 , h i 1 , , h i , n 1 ) consisting of { 0 , 1 , 1 } with associated polynomials H i ( ξ ) = h i 0 + h i 1 ξ + h i , n 1 ξ n 1 , where ξ is the n-th root of unity, i = 1 , 2 , 3 , 4 , if they satisfy the conditions
( i ) h i 0 = 0 , i = 1 , 2 , 3 , 4 , ( ii ) | h 1 k | + | h 2 k | + | h 3 k | + | h 4 k | = 1 , k = 1 , , n 1 , ( iii ) i = 1 4 H i ( ξ ) + 1 2 2 = n ,
then there exists a quad of GS sequences whose associated polynomials Q 1 ( ξ ) , , Q 4 ( ξ ) satisfy
Q 1 ( ξ ) = 1 H 1 ( ξ ) + H 2 ( ξ ) + H 3 ( ξ ) + H 4 ( ξ ) , Q 2 ( ξ ) = 1 + H 1 ( ξ ) H 2 ( ξ ) + H 3 ( ξ ) + H 4 ( ξ ) , Q 3 ( ξ ) = 1 + H 1 ( ξ ) + H 2 ( ξ ) H 3 ( ξ ) + H 4 ( ξ ) , Q 4 ( ξ ) = 1 + H 1 ( ξ ) + H 2 ( ξ ) + H 3 ( ξ ) H 4 ( ξ ) .
Proof. 
It is easy to verify that the coefficients of Q i ( ξ ) are ± 1 due to conditions (i) and (ii) of (11). Additionally, from (iii) of (11), we have
Q 1 ( ξ ) 2 + Q 2 ( ξ ) 2 + Q 3 ( ξ ) 2 + Q 4 ( ξ ) 2 = 4 + 4 ( H 1 ( ξ ) 2 + H 2 ( ξ ) 2 + H 3 ( ξ ) 2 + H 4 ( ξ ) 2 ) + 2 H 1 ( ξ ) ¯ + H 2 ( ξ ) ¯ + H 3 ( ξ ) ¯ + H 4 ( ξ ) ¯ + 2 H 1 ( ξ ) + H 2 ( ξ ) + H 3 ( ξ ) + H 4 ( ξ ) = 4 ( H 1 ( ξ ) + 1 2 2 + H 2 ( ξ ) + 1 2 2 + H 3 ( ξ ) + 1 2 2 + H 4 ( ξ ) + 1 2 2 ) = 4 n ,
which completes the proof. □
Two examples of GS sequences as Definition 1 are presented to verify this theorem: for n = 11 ,
h 1 = ( 0 , 1 , 1 , 1 , 0 , 0 , 0 , 0 , 0 , 0 , 0 ) , h 2 = ( 0 , 0 , 0 , 0 , 1 , 0 , 0 , 1 , 0 , 1 , 0 ) , h 3 = ( 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 1 , 0 , 1 ) , h 4 = ( 0 , 0 , 0 , 0 , 0 , 1 , 1 , 0 , 0 , 0 , 0 ) ,
which lead to the GS sequences of order 11
q 1 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , q 2 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , q 3 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , q 4 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) ,
and for n = 13 ,
h 1 = ( 0 , 1 , 1 , 1 , 1 , 1 , 0 , 0 , 0 , 0 , 0 , 0 , 0 ) , h 2 = ( 0 , 0 , 0 , 0 , 0 , 0 , 1 , 1 , 0 , 0 , 0 , 0 , 0 ) , h 3 = ( 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 1 , 0 , 1 , 0 , 1 ) , h 4 = ( 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 1 , 0 , 1 , 0 ) ,
with which we obtain the GS sequences of order 13
q 1 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , q 2 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , q 3 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , q 4 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) .
Remark 2. 
Theorem 1 provides a method to construct GS sequences by using T-matrix type sequences. Note that, for i = 1 , 2 , 3 , 4 , if polynomials H i ( ξ ) in (12) are symmetrical, the coefficients of Q i ( ξ ) are actually a quad of Williamson sequences as Definition 2.

3. Main Results

Even if the conditions of a quad of GS sequences are weaker than Williamson sequences, actually it is still not easy to construct GS sequences directly. In this section, two indirect methods will be proposed, which utilize the properties of symmetry and antisymmetry together with some sequences known beforehand.
Definition 5 
(Symmetry and antisymmetry). Let F i ( ξ ) be a polynomial with coefficients f i = ( f i 0 , , f i , n 1 ) . F i ( ξ ) being symmetrical (or antisymmetrical) if they satisfy
F i ( ξ ) ¯ = F i ( ξ ) ( o r F i ( ξ ) ¯ = F i ( ξ ) ) ,
where ξ is the n-th root of unity. That is, the coefficients ( f i 0 , , f i , n 1 ) satisfy f i j = f i , n j ( f i j = f i , n j , respectively), j = 1 , 2 , , n 1 .
Lemma 4. 
Given polynomials G i ( ξ ) = g i 0 + g i 1 ξ + . . . + g i , n 1 ξ n 1 for i = 1 , , 8 , G 1 ( ξ ) , G 2 ( ξ ) , G 3 ( ξ ) , G 4 ( ξ ) are symmetrical, and G 5 ( ξ ) , G 6 ( ξ ) , G 7 ( ξ ) , G 8 ( ξ ) are antisymmetrical, where ξ is the n-th root of unity. Then, there exist four polynomials F 1 ( ξ ) , F 2 ( ξ ) , F 3 ( ξ ) , F 4 ( ξ )
F 1 ( ξ ) = G 1 ( ξ ) G 2 ( ξ ) G 3 ( ξ ) G 4 ( ξ ) + G 5 ( ξ ) + G 6 ( ξ ) + G 7 ( ξ ) + G 8 ( ξ ) , F 2 ( ξ ) = G 1 ( ξ ) + G 2 ( ξ ) G 3 ( ξ ) G 4 ( ξ ) + G 5 ( ξ ) G 6 ( ξ ) G 7 ( ξ ) + G 8 ( ξ ) , F 3 ( ξ ) = G 1 ( ξ ) G 2 ( ξ ) + G 3 ( ξ ) G 4 ( ξ ) G 5 ( ξ ) + G 6 ( ξ ) G 7 ( ξ ) + G 8 ( ξ ) , F 4 ( ξ ) = G 1 ( ξ ) G 2 ( ξ ) G 3 ( ξ ) + G 4 ( ξ ) G 5 ( ξ ) G 6 ( ξ ) + G 7 ( ξ ) + G 8 ( ξ ) ,
satisfying
i = 1 4 F i ( ξ ) 2 = i = 1 4 F i ( ξ ) F i ( ξ ) ¯ = 4 i = 1 8 G i ( ξ ) 2 .
Proof. 
Duo to the properties of symmetry and antisymmetry, we have G i ( ξ ) = G i ( ξ ) ¯ , i = 1 , 2 , 3 , 4 and G i ( ξ ) = G i ( ξ ) ¯ , i = 5 , 6 , 7 , 8 , which leads to the result after some tedious calculation. □
Lemma 4 implies that the construction of polynomials { F i } i = 1 4 could be changed to find some appropriate polynomials { G i } i = 1 8 . In order to obtain new GS sequences, we now introduce the definition of a k-partition which is a special case of L-matrices ([6], Definition 4.15), i.e., the appropriate polynomials { G i } i = 1 8 we find.
Definition 6 
(k-partition). If polynomials G i ( ξ ) = g i 0 + g i 1 ξ + + g i , n 1 ξ n 1 , i = 1 , , k , satisfy
( i ) g i j { 0 , 1 , 1 } , i , j = 0 , 1 , , n 1 , ( ii ) i = 1 k | g i j | = 1 , j = 0 , 1 , , n 1 , ( iii ) i = 1 k G i ( ξ ) 2 = n ,
where ξ is the n-th root of unity, then we call { G i ( ξ ) } i = 1 k a k-partition of sequences with ± 1 , abbreviation as k-partition without confusion.
We show some examples of 8-partition as Definition 6: for n = 10
g 1 = ( 1 , 0 , 0 , 0 , 1 , 0 , 1 , 0 , 0 , 0 ) , g 2 = ( 0 , 0 , 1 , 0 , 0 , 0 , 0 , 0 , 1 , 0 ) , g 3 = ( 0 , 0 , 0 , 0 , 0 , 1 , 0 , 0 , 0 , 0 ) , g 4 = ( 0 , 0 , 0 , 1 , 0 , 0 , 0 , 1 , 0 , 0 ) , g 5 = ( 0 , 1 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 1 ) , g 6 = ( 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 ) , g 7 = ( 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 ) , g 8 = ( 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 ) ,
and for n = 12
g 1 = ( 1 , 0 , 0 , 0 , 0 , 0 , 1 , 0 , 0 , 0 , 0 , 0 ) , g 2 = ( 0 , 0 , 0 , 1 , 0 , 0 , 0 , 0 , 0 , 1 , 0 , 0 ) , g 3 = ( 0 , 1 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 1 ) , g 4 = ( 0 , 0 , 1 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 1 , 0 ) , g 5 = ( 0 , 0 , 0 , 0 , 1 , 0 , 0 , 0 , 1 , 0 , 0 , 0 ) , g 6 = ( 0 , 0 , 0 , 0 , 0 , 1 , 0 , 1 , 0 , 0 , 0 , 0 ) , g 7 = ( 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 ) , g 8 = ( 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 ) .
Comparing Definition 4 with Definition 6, it is easy to see that actually T-matrices sequences are the special cases of 4-partitions. As an extension, we will next investigate how to utilize 8-th partitions to obtain more GS sequences with different structures and orders.
Corollary 1. 
Let { G i ( ξ ) } be an 8-partition, where G 1 ( ξ ) , , G 4 ( ξ ) are symmetrical and G 5 ( ξ ) , . . . , G 8 ( ξ ) are antisymmetrical. Then, for F i ( ξ ) defined in (14), the coefficient sequences f i of associated polynomials F i ( ξ ) , i = 1 , 2 , 3 , 4 , make up a quad of GS sequences.
Proof. 
Since { G i ( ξ ) } is an 8-partition, together with the statement (ii) of (16), the coefficients of F i ( ξ ) consist of ± 1 . Combining the fact
i = 1 4 F i ( ξ ) 2 = 4 i = 1 8 G i ( ξ ) 2 = 4 n
with the definition of GS sequences (1), we arrive at the result. □
According to two 8-partition of length n = 10 (17) and n = 12 (18), we obtain two quad of GS sequences as Definition 1 based on the construction in Lemma 4. For n = 10 , the GS sequences are
f 1 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , f 2 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , f 3 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , f 4 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) ,
and, for n = 12 , the GS sequences are
f 1 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , f 2 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , f 3 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , f 4 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) .
Remark 3. 
Corollary 1 is a direct conclusion of Lemma 4, and indicates that we now could turn to find an 8-partition instead of a quad of GS sequences.
Theorem 2. 
Let two classes of polynomials { G i ( ξ m ) } i = 1 8 and { E i ( ξ n ) } i = 1 8 be symmetrical for i = 1 , 2 , 3 , 4 and antisymmetrical for i = 5 , 6 , 7 , 8 . Both { G i ( ξ m ) } and { E i ( ξ n ) } are 8-partitions, where ξ is the m n -th root of unity with ( m , n ) = 1 . Define eight polynomials { L i ( ξ ) } i = 1 8 by these two classes of polynomials as
L 1 ( ξ ) = G 1 ( ξ m ) E 1 ( ξ n ) + G 2 ( ξ m ) E 2 ( ξ n ) + G 3 ( ξ m ) E 3 ( ξ n ) + G 4 ( ξ m ) E 4 ( ξ n ) + G 5 ( ξ m ) E 5 ( ξ n ) + G 6 ( ξ m ) E 6 ( ξ n ) + G 7 ( ξ m ) E 7 ( ξ n ) + G 8 ( ξ m ) E 8 ( ξ n ) , L 2 ( ξ ) = G 1 ( ξ m ) E 2 ( ξ n ) G 2 ( ξ m ) E 1 ( ξ n ) + G 3 ( ξ m ) E 4 ( ξ n ) G 4 ( ξ m ) E 3 ( ξ n ) + G 5 ( ξ m ) E 6 ( ξ n ) G 6 ( ξ m ) E 5 ( ξ n ) + G 7 ( ξ m ) E 8 ( ξ n ) G 8 ( ξ m ) E 7 ( ξ n ) , L 3 ( ξ ) = G 1 ( ξ m ) E 3 ( ξ n ) G 2 ( ξ m ) E 4 ( ξ n ) G 3 ( ξ m ) E 1 ( ξ n ) + G 4 ( ξ m ) E 2 ( ξ n ) + G 5 ( ξ m ) E 7 ( ξ n ) G 6 ( ξ m ) E 8 ( ξ n ) G 7 ( ξ m ) E 5 ( ξ n ) + G 8 ( ξ m ) E 6 ( ξ n ) , L 4 ( ξ ) = G 1 ( ξ m ) E 4 ( ξ n ) + G 2 ( ξ m ) E 3 ( ξ n ) G 3 ( ξ m ) E 2 ( ξ n ) G 4 ( ξ m ) E 1 ( ξ n ) G 5 ( ξ m ) E 8 ( ξ n ) G 6 ( ξ m ) E 7 ( ξ n ) + G 7 ( ξ m ) E 6 ( ξ n ) + G 8 ( ξ m ) E 5 ( ξ n ) , L 5 ( ξ ) = G 1 ( ξ m ) E 5 ( ξ n ) G 2 ( ξ m ) E 6 ( ξ n ) G 3 ( ξ m ) E 7 ( ξ n ) + G 4 ( ξ m ) E 8 ( ξ n ) + G 5 ( ξ m ) E 1 ( ξ n ) G 6 ( ξ m ) E 2 ( ξ n ) G 7 ( ξ m ) E 3 ( ξ n ) + G 8 ( ξ m ) E 4 ( ξ n ) , L 6 ( ξ ) = G 1 ( ξ m ) E 6 ( ξ n ) + G 2 ( ξ m ) E 5 ( ξ n ) + G 3 ( ξ m ) E 8 ( ξ n ) + G 4 ( ξ m ) E 7 ( ξ n ) + G 5 ( ξ m ) E 2 ( ξ n ) + G 6 ( ξ m ) E 1 ( ξ n ) + G 7 ( ξ m ) E 4 ( ξ n ) + G 8 ( ξ m ) E 3 ( ξ n ) , L 7 ( ξ ) = G 1 ( ξ m ) E 7 ( ξ n ) G 2 ( ξ m ) E 8 ( ξ n ) + G 3 ( ξ m ) E 5 ( ξ n ) G 4 ( ξ m ) E 6 ( ξ n ) + G 5 ( ξ m ) E 3 ( ξ n ) G 6 ( ξ m ) E 4 ( ξ n ) + G 7 ( ξ m ) E 1 ( ξ n ) G 8 ( ξ m ) E 2 ( ξ n ) , L 8 ( ξ ) = G 1 ( ξ m ) E 8 ( ξ n ) + G 2 ( ξ m ) E 7 ( ξ n ) G 3 ( ξ m ) E 6 ( ξ n ) G 4 ( ξ m ) E 5 ( ξ n ) G 5 ( ξ m ) E 4 ( ξ n ) G 6 ( ξ m ) E 3 ( ξ n ) + G 7 ( ξ m ) E 2 ( ξ n ) + G 8 ( ξ m ) E 1 ( ξ n ) .
Then L i ( ξ ) are symmetrical for i = 1 , 2 , 3 , 4 and antisymmetrical for i = 5 , 6 , 7 , 8 , and the coefficients of the following polynomials:
Q 1 ( ξ ) = L 1 ( ξ ) L 2 ( ξ ) L 3 ( ξ ) L 4 ( ξ ) + L 5 ( ξ ) + L 6 ( ξ ) + L 7 ( ξ ) + L 8 ( ξ ) , Q 2 ( ξ ) = L 1 ( ξ ) + L 2 ( ξ ) L 3 ( ξ ) L 4 ( ξ ) + L 5 ( ξ ) L 6 ( ξ ) L 7 ( ξ ) + L 8 ( ξ ) , Q 3 ( ξ ) = L 1 ( ξ ) L 2 ( ξ ) + L 3 ( ξ ) L 4 ( ξ ) L 5 ( ξ ) + L 6 ( ξ ) L 7 ( ξ ) + L 8 ( ξ ) , Q 4 ( ξ ) = L 1 ( ξ ) L 2 ( ξ ) L 3 ( ξ ) + L 4 ( ξ ) L 5 ( ξ ) L 6 ( ξ ) + L 7 ( ξ ) + L 8 ( ξ ) ,
make up a quad of GS sequences.
Proof. 
Due to
G i ( ξ m ) ¯ = G i ( ξ m ) , E i ( ξ n ) ¯ = E i ( ξ n ) , i = 1 , 2 , 3 , 4 , G i ( ξ m ) ¯ = G i ( ξ m ) , E i ( ξ n ) ¯ = E i ( ξ n ) , i = 5 , 6 , 7 , 8 ,
correspondingly, we have
L i ( ξ ) ¯ = L i ( ξ ) , i = 1 , 2 , 3 , 4 , L i ( ξ ) ¯ = L i ( ξ ) , i = 5 , 6 , 7 , 8 .
Since { E i ( ξ n ) } and { G i ( ξ m ) } are 8-partitions, it holds that
i = 1 8 E i ( ξ n ) 2 = m and i = 1 8 G i ( ξ m ) 2 = n ,
which leads to the fact that L i ( ξ ) are an 8-partition satisfying
i = 1 8 L i ( ξ ) 2 = i = 1 8 G i ( ξ m ) 2 i = 1 8 E i ( ξ n ) 2 = m n ,
and the coefficients of Q i ( ξ ) are ± 1 . Then, we obtain
i = 4 4 Q i ( ξ ) 2 = 4 i = 1 8 L i ( ξ ) 2 = 4 i = 1 8 G i ( ξ m ) 2 i = 1 8 E i ξ n 2 = 4 m n .
Here, the tedious calculation in identities (21) and (22) is completed with the help of MATLAB. Using Corollary 1, we obtain that the coefficients q i of corresponding polynomials Q i ( ξ ) , i = 1 , 2 , 3 , 4 , are a quad of GS sequences. □
Theorem 2 provides another way to obtain the sequences { G i } i = 1 8 in Lemma 4, i.e., the sequences { L i } i = 1 8 in Theorem 2.
Remark 4. 
For now, it is not easy to find an 8-partition with two large n, as stated in Theorem 2, with the requirements of symmetry and antisymmetry. Hence, it could be reduced to find a 4-partition and a class of symmetrical sequences, which will lead to two known conclusions [32]. It could be considered as special cases of our results shown in Lemma 5 and Corollary 2.
Lemma 5. 
Given a class of symmetrical polynomials { G i ( ξ m ) } i = 1 4 and a 4-partition { E i ( ξ n ) } , and defining L ˜ i ( ξ ) as
L ˜ 1 ( ξ ) = G 1 ( ξ m ) E 1 ( ξ n ) ¯ + G 2 ( ξ m ) E 2 ( ξ n ) + G 3 ( ξ m ) E 3 ( ξ n ) + G 4 ( ξ m ) E 4 ( ξ n ) , L ˜ 2 ( ξ ) = G 1 ( ξ m ) E 2 ( ξ n ) ¯ G 2 ( ξ m ) E 1 ( ξ n ) + G 3 ( ξ m ) E 4 ( ξ n ) G 4 ( ξ m ) E 3 ( ξ n ) , L ˜ 3 ( ξ ) = G 1 ( ξ m ) E 3 ( ξ n ) ¯ G 2 ( ξ m ) E 4 ( ξ n ) G 3 ( ξ m ) E 1 ( ξ n ) + G 4 ( ξ m ) E 2 ( ξ n ) , L ˜ 4 ( ξ ) = G 1 ( ξ m ) E 4 ( ξ n ) ¯ + G 2 ( ξ m ) E 3 ( ξ n ) G 3 ( ξ m ) E 2 ( ξ n ) G 4 ( ξ m ) E 1 ( ξ n ) ,
then we have
i = 1 4 L ˜ i ( ξ ) 2 = i = 1 4 G i ( ξ m ) 2 i = 1 4 E i ( ξ n ) 2 ,
where ξ is the m n -th root of unity and ( m , n ) = 1 .
Proof. 
The symmetry of { G i ( ξ m ) } i = 1 4 yields
L ˜ 1 ( ξ ) 2 = L ˜ 1 ( ξ ) L ˜ 1 ( ξ ) ¯ = G 1 ( ξ m ) G 1 ( ξ m ) ¯ E 1 ( ξ n ) ¯ E 1 ( ξ n ) + G 1 ( ξ m ) G 2 ( ξ m ) ¯ E 1 ( ξ n ) ¯ E 2 ( ξ n ) ¯ + G 1 ( ξ m ) G 3 ( ξ m ) ¯ E 1 ( ξ n ) ¯ E 3 ( ξ n ) ¯ + G 1 ( ξ m ) G 4 ( ξ m ) ¯ E 1 ( ξ n ) ¯ E 4 ( ξ n ) ¯ + G 2 ( ξ m ) G 1 ( ξ m ) ¯ E 2 ( ξ n ) E 1 ( ξ n ) + G 2 ( ξ m ) G 2 ( ξ m ) ¯ E 2 ( ξ n ) E 2 ( ξ n ) ¯ + G 2 ( ξ m ) G 3 ( ξ m ) ¯ E 2 ( ξ n ) E 3 ( ξ n ) ¯ + G 2 ( ξ m ) G 4 ( ξ m ) ¯ E 2 ( ξ n ) E 4 ( ξ n ) ¯ + G 3 ( ξ m ) G 1 ( ξ m ) ¯ E 3 ( ξ n ) E 1 ( ξ n ) + G 3 ( ξ m ) G 2 ( ξ m ) ¯ E 3 ( ξ n ) E 2 ( ξ n ) ¯ + G 3 ( ξ m ) G 3 ( ξ m ) ¯ E 3 ( ξ n ) E 3 ( ξ n ) ¯ + G 3 ( ξ m ) G 4 ( ξ m ) ¯ E 3 ( ξ n ) E 4 ( ξ n ) ¯ + G 4 ( ξ m ) G 1 ( ξ m ) ¯ E 4 ( ξ n ) E 1 ( ξ n ) + G 4 ( ξ m ) G 2 ( ξ m ) ¯ E 4 ( ξ n ) E 2 ( ξ n ) ¯ + G 4 ( ξ m ) G 3 ( ξ m ) ¯ E 4 ( ξ n ) E 3 ( ξ n ) ¯ + G 4 ( ξ m ) G 4 ( ξ m ) ¯ E 4 ( ξ n ) E 4 ( ξ n ) ¯ ,
L ˜ 2 ( ξ ) 2 = L ˜ 2 ( ξ ) L ˜ 2 ( ξ ) ¯ = G 1 ( ξ m ) G 1 ( ξ m ) ¯ E 2 ( ξ n ) ¯ E 2 ( ξ n ) G 1 ( ξ m ) G 2 ( ξ m ) ¯ E 1 ( ξ n ) ¯ E 2 ( ξ n ) ¯ + G 1 ( ξ m ) G 3 ( ξ m ) ¯ E 2 ( ξ n ) ¯ E 4 ( ξ n ) ¯ G 1 ( ξ m ) G 4 ( ξ m ) ¯ E 2 ( ξ n ) ¯ E 3 ( ξ n ) ¯ G 2 ( ξ m ) G 1 ( ξ m ) ¯ E 2 ( ξ n ) E 1 ( ξ n ) + G 2 ( ξ m ) G 2 ( ξ m ) ¯ E 1 ( ξ n ) E 1 ( ξ n ) ¯ G 2 ( ξ m ) G 3 ( ξ m ) ¯ E 1 ( ξ n ) E 4 ( ξ n ) ¯ + G 2 ( ξ m ) G 4 ( ξ m ) ¯ E 1 ( ξ n ) E 3 ( ξ n ) ¯ + G 3 ( ξ m ) G 1 ( ξ m ) ¯ E 4 ( ξ n ) E 2 ( ξ n ) G 3 ( ξ m ) G 2 ( ξ m ) ¯ E 4 ( ξ n ) E 1 ( ξ n ) ¯ + G 3 ( ξ m ) G 3 ( ξ m ) ¯ E 3 ( ξ n ) E 3 ( ξ n ) ¯ G 3 ( ξ m ) G 4 ( ξ m ) ¯ E 4 ( ξ n ) E 3 ( ξ n ) ¯ G 4 ( ξ m ) G 1 ( ξ m ) ¯ E 3 ( ξ n ) E 2 ( ξ n ) + G 4 ( ξ m ) G 2 ( ξ m ) ¯ E 3 ( ξ n ) E 1 ( ξ n ) ¯ G 4 ( ξ m ) G 3 ( ξ m ) ¯ E 3 ( ξ n ) E 4 ( ξ n ) ¯ + G 4 ( ξ m ) G 4 ( ξ m ) ¯ E 3 ( ξ n ) E 3 ( ξ n ) ¯ ,
and also for L ˜ i ( ξ ) 2 , i = 3 , 4 . Here, note that the 2nd, 5th, 12th, and 15th terms in L ˜ 1 ( ξ ) 2 will vanish when calculating L ˜ 1 ( ξ ) 2 + L ˜ 2 ( ξ ) 2 , and finally we have
i = 1 4 L ˜ i ( ξ ) 2 = G 1 ( ξ m ) 2 E 1 ( ξ n ) 2 + G 2 ( ξ m ) 2 E 2 ( ξ n ) 2 + G 3 ( ξ m ) 2 E 3 ( ξ n ) 2 + G 4 ( ξ m ) 2 E 4 ( ξ n ) 2 + G 1 ( ξ m ) 2 E 2 ( ξ n ) 2 + G 2 ( ξ m ) 2 E 1 ( ξ n ) 2 + G 3 ( ξ m ) 2 E 4 ( ξ n ) 2 + G 4 ( ξ m ) 2 E 3 ( ξ n ) 2 + G 1 ( ξ m ) 2 E 3 ( ξ n ) 2 + G 2 ( ξ m ) 2 E 4 ( ξ n ) 2 + G 3 ( ξ m ) 2 E 1 ( ξ n ) 2 + G 4 ( ξ m ) 2 E 2 ( ξ n ) 2 + G 1 ( ξ m ) 2 E 4 ( ξ n ) 2 + G 2 ( ξ m ) 2 E 3 ( ξ n ) 2 + G 3 ( ξ m ) 2 E 2 ( ξ n ) 2 + G 4 ( ξ m ) 2 E 1 ( ξ n ) 2 = i = 1 4 G i ( ξ m ) 2 i = 1 4 E i ( ξ n ) 2 .
This completes the proof. □
Furthermore, if the coefficient sequences g i of associated polynomials G i ( ξ m ) are a quad of Williamson sequences, we present the main result that a quad of GS sequences has been obtained, as shown in the following corollary:
Corollary 2. 
Let G 1 ( ξ m ) , , G 4 ( ξ m ) be four associated polynomials of a quad of Williamson sequences and E 1 ( ξ n ) , , E 4 ( ξ n ) be a 4-partition. Then, the coefficient sequences l ˜ i of polynomials { L ˜ i } defined as
L ˜ 1 ( ξ ) = G 1 ( ξ m ) E 1 ( ξ n ) + G 2 ( ξ m ) E 2 ( ξ n ) + G 3 ( ξ m ) E 3 ( ξ n ) + G 4 ( ξ m ) E 4 ( ξ n ) , L ˜ 2 ( ξ ) = G 1 ( ξ m ) E 2 ( ξ n ) G 2 ( ξ m ) E 1 ( ξ n ) + G 3 ( ξ m ) E 4 ( ξ n ) G 4 ( ξ m ) E 3 ( ξ n ) , L ˜ 3 ( ξ ) = G 1 ( ξ m ) E 3 ( ξ n ) G 2 ( ξ m ) E 4 ( ξ n ) G 3 ( ξ m ) E 1 ( ξ n ) + G 4 ( ξ m ) E 2 ( ξ n ) , L ˜ 4 ( ξ ) = G 1 ( ξ m ) E 4 ( ξ n ) + G 2 ( ξ m ) E 3 ( ξ n ) G 3 ( ξ m ) E 2 ( ξ n ) G 4 ( ξ m ) E 1 ( ξ n ) ,
are a quad of GS sequences, where ξ is the m n -th root of unity.
Proof. 
The coefficients of G 1 ( ξ m ) , G 2 ( ξ m ) , G 3 ( ξ m ) , G 4 ( ξ m ) are a quad of Williamson sequences and thus their coefficients belong to { 1 , 1 } . For E 1 ( ξ n ) , , E 4 ( ξ n ) being a 4-partition and from the condition (ii) of Definition 6, the coefficients of L i ˜ ( ξ ) are made up of ± 1 , too. Meanwhile, four polynomials G 1 ( ξ m ) , G 2 ( ξ m ) , G 3 ( ξ m ) , G 4 ( ξ m ) satisfy
G 1 ( ξ m ) 2 + G 2 ( ξ m ) 2 + G 3 ( ξ m ) 2 + G 4 ( ξ m ) 2 = 4 n .
Similar to Lemma 5, we obtain the desired result
i = 1 4 L ˜ i ( ξ ) 2 = i = 1 4 G i ( ξ m ) 2 i = 1 4 E i ( ξ n ) 2 = 4 m n .
This completes the proof. □
Next, we present an example to verify Corollary 2. For m = 7 , n = 6 , { g i } i = 1 4 is a quad of Williamson sequences of order n, and { e i } i = 1 4 is a 4-partition,
g 1 = ( 1 , 0 , 0 , 0 , 0 , 0 ) , g 2 = ( 0 , 0 , 0 , 1 , 0 , 0 ) , g 3 = ( 0 , 1 , 0 , 0 , 0 , 1 ) , g 4 = ( 0 , 0 , 1 , 0 , 1 , 0 ) ,
and
e 1 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , e 2 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , e 3 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , e 4 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 ) .
Thus, as the construction of (25), the GS sequences are
l ˜ 1 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , l ˜ 2 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , l ˜ 3 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) , l ˜ 4 = ( 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ) .
Remark 5. 
From Corollary 2, for now, the construction eventually becomes to find a quad of Williamson sequences and a 4-partition. It is meaningful since, in practice, some Williamson sequences have been constructed, for example, in [11], and a class of TMS itself is a 4-partition.

4. Conclusions

Due to the important application of Hadamard matrices, it is meaningful to find some novel methods to construct them. An alternative way is making use of GS sequences based on the combination of 4-partition from the existing works. It is noted, however, that, for some large n, there do not exist for 4-partitions but exist for 8-partitions, so that we are motivated to design GS sequences by utilizing 8-partitions as a natural generalization of 4-partition.
To this end, we first construct a quad of GS sequences based on the T-matrices directly. Next, we introduce the definitions of k-partition and utilize the properties of symmetry and antisymmetry. As a result, the construction process turns to finding some k-partitions. More specifically, this paper shows that a class of 8-partition with some (anti)symmetry properties could give a quad of GS sequences. As the special case k = 4 , one can use the existing results, a quad of Williamson sequences and a quad of TMS, to construct Hadamard matrices immediately with more orders, compared with only using a quad of Williamson sequences directly. In the future, the constructions of 8-partition will be investigated which could be used to construct GS sequences with more orders by using the methods proposed in this paper. Meanwhile, it has the potential to enhance k following the methods developed in this paper, to find more orders n for k-partitions and further GS sequences.

Author Contributions

Conceptualization, S.S. and X.Z.; methodology, S.S. and X.Z.; software, S.S. and X.Z.; validation, S.S. and X.Z.; writing—original draft preparation, S.S.; writing—review and editing, S.S. and X.Z.; visualization, S.S. and X.Z.; supervision, X.Z.; funding acquisition, X.Z.; All authors have read and agreed to the published version of the manuscript.

Funding

This research is financially supported by the National Natural Science Foundation of China (No.61771004).

Data Availability Statement

The authors confirm that the data supporting the findings of this study are available within the article.

Conflicts of Interest

We declare that we have no financial and personal relationships with other people or organizations that can inappropriately influence our work, there is no professional or other personal interest of any nature or kind in any product, service and/or company that could be construed as influencing the position presented in, or the review of, this paper.

References

  1. Álvarez, V.; Armario, J.; Falcón, R.; Frau, M.; Gudiel, F.; Güemes, M.; Osuna, A. On Cocyclic Hadamard Matrices over Goethals–Seidel Loops. Mathematics 2020, 8, 24. [Google Scholar] [CrossRef] [Green Version]
  2. Álvarez, V.; Armario, J.; Frau, M.; Gudiel, F.; Güemes, M.; Osuna, A. Hadamard Matrices with Cocyclic Core. Mathematics 2021, 9, 857. [Google Scholar] [CrossRef]
  3. Armario, J. Boolean Functions and Permanents of Sylvester Hadamard Matrices. Mathematics 2021, 9, 177. [Google Scholar] [CrossRef]
  4. Barrera Acevedo, S.; Ó Catháin, P.; Dietrich, H. Cocyclic two-circulant core Hadamard matrices. J. Algebraic Combin. 2022, 55, 201–215. [Google Scholar] [CrossRef]
  5. Horadam, K.J. Hadamard Matrices and Their Applications; Princeton University Press: Princeton, NJ, USA, 2007. [Google Scholar]
  6. Seberry, J. Orthogonal Designs; Hadamard Matrices, Quadratic forms and Algebras, Revised and Updated Edition of the 1979 Original [MR0534614]; Springer: Cham, Switzerland, 2017; pp. 1–5. [Google Scholar]
  7. Wallis, W.D.; Street, A.P.; Wallis, J.S. Combinatorics: Room Squares, Sum-Free Sets, Hadamard Matrices; Lecture Notes in Mathematics; Springer: Berlin, Germany; New York, NY, USA, 1972. [Google Scholar]
  8. Xia, T.; Xia, M.; Seberry, J. The construction of regular Hadamard matrices by cyclotomic classes. Bull. Iran. Math. Soc. 2021, 47, 601–625. [Google Scholar] [CrossRef]
  9. Kharaghani, H. Arrays for orthogonal designs. J. Combin. Des. 2000, 8, 166–173. [Google Scholar] [CrossRef]
  10. Goethals, J.; Seidel, J. A skew Hadamard matrix of order 36. J. Austral. Math. Soc. 1970, 11, 343–344. [Google Scholar] [CrossRef] [Green Version]
  11. Whiteman, A. An infinite family of Hadamard matrices of Williamson type. J. Comb. Theory A 1973, 14, 334–340. [Google Scholar] [CrossRef] [Green Version]
  12. Doković, D.Ž. Construction of some new Hadamard matrices. Bull. Austral. Math. Soc. 1992, 45, 327–332. [Google Scholar] [CrossRef] [Green Version]
  13. Doković, D.Ž. Skew Hadamard matrices of order 4 × 37 and 4 × 43. J. Combin. Theory Ser. A 1992, 61, 319–321. [Google Scholar] [CrossRef]
  14. Doković, D.Ž. Ten Hadamard matrices of order 1852 of Goethals-Seidel type. Eur. J. Combin. 1992, 13, 245–248. [Google Scholar] [CrossRef] [Green Version]
  15. Doković, D.Ž. Ten new orders for Hadamard matrices of skew type. Univ. Beograd. Publ. Elektrotehn. Fak. Ser. Mat. 1992, 3, 47–59. [Google Scholar]
  16. Doković, D.Ž. Two Hadamard matrices of order 956 of Goethals-Seidel type. Combinatorica 1994, 14, 375–377. [Google Scholar] [CrossRef]
  17. Doković, D.Ž. Skew-Hadamard matrices of orders 436, 580, and 988 exist. J. Combin. Des. 2008, 16, 493–498. [Google Scholar] [CrossRef] [Green Version]
  18. Doković, D.Ž.; Golubitsky, O.; Kotsireas, I. Some new orders of Hadamard and skew-Hadamard matrices. J. Combin. Des. 2014, 22, 270–277. [Google Scholar] [CrossRef] [Green Version]
  19. Doković, D.Ž.; Kotsireas, I. Goethals-Seidel difference families with symmetric or skew base blocks. Math. Comput. Sci. 2018, 12, 373–388. [Google Scholar] [CrossRef] [Green Version]
  20. Fletcher, R.; Koukouvinos, C.; Seberry, J. New skew-Hadamard matrices of order 4 × 59 and new D-optimal designs of order 2 × 59. Discrete Math. 2004, 286, 251–253. [Google Scholar] [CrossRef]
  21. Whiteman, A. Skew Hadamard matrices of Goethals—Seidel type. Discrete Math. 1972, 2, 397–405. [Google Scholar] [CrossRef] [Green Version]
  22. Xia, M.; Xia, T.; Seberry, J.; Wu, J. An infinite family of Goethals-Seidel arrays. Discrete Appl. Math. 2005, 145, 498–504. [Google Scholar] [CrossRef] [Green Version]
  23. Yang, C.H. Hadamard matrices, finite sequences, and polynomials defined on the unit circle. Math. Comp. 1979, 33, 688–693. [Google Scholar] [CrossRef]
  24. Yang, C.H. Hadamard matrices and δ-codes of length 3n. Proc. Amer. Math. Soc. 1982, 85, 480–482. [Google Scholar] [CrossRef]
  25. Yang, C.H. Lagrange identity for polynomials and δ-codes of lengths 7t and 13t. Proc. Amer. Math. Soc. 1983, 88, 746–750. [Google Scholar] [CrossRef]
  26. Yang, C. A composition theorem for δ-codes. Proc. Amer. Math. Soc. 1983, 89, 375–378. [Google Scholar] [CrossRef]
  27. Yang, C. On composition of four-symbol δ-codes and Hadamard matrices. Proc. Amer. Math. Soc. 1989, 107, 763–776. [Google Scholar]
  28. Doković, D.Ž.; Kotsireas, I. Compression of periodic complementary sequences and applications. Des. Codes Cryptogr. 2015, 74, 365–377. [Google Scholar] [CrossRef] [Green Version]
  29. Fletcher, R.; Gysin, M.; Seberry, J. Application of the discrete Fourier transform to the search for generalised Legendre pairs and Hadamard matrices. Australas. J. Combin. 2001, 23, 75–86. [Google Scholar]
  30. Zuo, G.; Xia, M.; Xia, T. Constructions of composite T-matrices. Linear Algebra Appl. 2013, 438, 1223–1228. [Google Scholar] [CrossRef]
  31. Cooper, J.; Wallis, J. A construction for Hadamard arrays. Bull. Austral. Math. Soc. 1972, 7, 269–277. [Google Scholar] [CrossRef] [Green Version]
  32. Xia, M.; Xia, T. A family of C-partitions and T-matrices. J. Combin. Des. 1999, 7, 269–281. [Google Scholar] [CrossRef]
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

Shen, S.; Zhang, X. Constructions of Goethals–Seidel Sequences by Using k-Partition. Mathematics 2023, 11, 294. https://doi.org/10.3390/math11020294

AMA Style

Shen S, Zhang X. Constructions of Goethals–Seidel Sequences by Using k-Partition. Mathematics. 2023; 11(2):294. https://doi.org/10.3390/math11020294

Chicago/Turabian Style

Shen, Shuhui, and Xiaojun Zhang. 2023. "Constructions of Goethals–Seidel Sequences by Using k-Partition" Mathematics 11, no. 2: 294. https://doi.org/10.3390/math11020294

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