Next Article in Journal
Novel Entropy for Enhanced Thermal Imaging and Uncertainty Quantification
Next Article in Special Issue
On the Dimensions of Hermitian Subfield Subcodes from Higher-Degree Places
Previous Article in Journal
A Personalized Collaborative Filtering Recommendation System Based on Bi-Graph Embedding and Causal Reasoning
Previous Article in Special Issue
Evaluating the Gilbert–Varshamov Bound for Constrained Systems
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Optimal Quaternary Hermitian LCD Codes

Fundamentals Department, Air Force Engineering University, Xi’an 710051, China
*
Author to whom correspondence should be addressed.
Entropy 2024, 26(5), 373; https://doi.org/10.3390/e26050373
Submission received: 29 February 2024 / Revised: 17 April 2024 / Accepted: 22 April 2024 / Published: 28 April 2024
(This article belongs to the Special Issue Discrete Math in Coding Theory)

Abstract

:
Linear complementary dual (LCD) codes, which are a class of linear codes introduced by Massey, have been extensively studied in the literature recently. It has been shown that LCD codes play a role in measures to counter passive and active side-channel analyses on embedded cryptosystems. In this paper, tables are presented of good quaternary Hermitian LCD codes and they are used in the construction of puncturing, shortening and combination codes. The results of this, including three tables of the best-known quaternary Hermitian LCD codes of any length n 25 with corresponding dimension k, are presented. In addition, many of these quaternary Hermitian LCD codes given in this paper are optimal and saturate the lower or upper bound of Grassl’s code table, and some of them are nearly optimal.

1. Introduction

Let q be a power of a prime p, F q be a finite field with q elements, and F q n be an n-dimensional vector space over F q . A q-ary [ n , k , d ] q linear code over F q is a k-dimensional subspace of F q n with Hamming distance d. For a given [ n , k ] q linear code, the code C = { x | x · c = 0 , c C } is called the dual code of C . A q-ary linear code C is called a linear complementary dual (LCD) code if it meets its dual trivially, that is, C C = { 0 } , which was given by Massey [1,2]. In addition to their applications in data storage, communication systems, and consumer electronics, LCD codes have recently been employed in cryptography and quantum error correcting. Carlet and Guilley in ref. [3] showed that LCD codes play an important role in armoring implementations against side-channel attacks and presented several constructions of LCD codes.
In [4], according to finite geometry theory, Lu et al. proposed the radical codes  R ( C ) of C and C , which are R ( C ) = C C . If R ( C ) = C C = { 0 } , then C is called a zero radical code, which is the same as the LCD code presented in [2]. Using these zero radical codes, they constructed families of maximal entanglement entanglement-assisted quantum error-correcting codes, which can help to engineer more reliable quantum communication schemes and quantum computers. Furthermore, constructions of Hermitian zero radical BCH codes were discussed in [5], which are also called reversible codes in [1] or LCD cyclic codes in [6]. Güneri et al. studied quasi-cyclic LCD codes and introduced Hermitian LCD codes [7]. Moreover, for the Euclidean case, the question of when cyclic codes are LCD codes is answered affirmatively by Yang and Massey in [8]. Ding et al. investigated LCD cyclic codes in [6], in which several families of LCD cyclic codes were constructed. It is shown that some LCD cyclic codes are optimal linear codes or have the best possible parameters for cyclic codes. Shi et al. constructed a lot of good LCD codes [9,10,11,12]. Moreover, many works have focused on the construction of LCD codes with good parameters, see [13,14,15,16,17,18,19,20,21].
Recently, Carlet, Mesnager, Tang, Qi and Pellikaan in [22] have shown that any [ n , k , d ] -linear code over F q 2 is equivalent to an [ n , k , d ] -linear Hermitian LCD code over F q 2 for q > 2 . Araya, Harada and Saito in [23] gave some conditions for the nonexistence of quaternary Hermitian linear complementary dual codes with large minimum weights. Inspired by these works and extending our previous work in [4], we study constructions of linear Hermitian LCD codes over F 4 . Then, some families of linear Hermitian LCD codes with good parameters are constructed from the known optimal codes via puncturing, extending, shortening and the combination method. Compared with the tables of best known linear codes (referred to as the Database later) maintained by Markus Grassl in [24], some of our codes presented in this paper saturate the lower bound of Grassl’s code table.
In this paper, an optimal quaternary Hermitian LCD code [ 18 , 7 , 9 ] is given, which improves the minimal distance of the codes in [4,25,26]. According to classification codes in [27], there exist some optimal quaternary Hermitian LCD codes [ 15 , 4 , 8 ] , [ 16 , 4 , 9 ] , [ 17 , 4 , 10 ] , [ 19 , 4 , 13 ] , [ 23 , 5 , 14 ] and [ 15 , 6 , 7 ] . According to [24], the following quaternary Hermitian LCD codes we give in this section are also optimal linear codes: [ n , k , d ] for 21 n 25 and 16 k 18 ; [ n i , 15 i , d ] for 21 n 24 and 0 i 2 ; and [ 23 , 4 , 15 ] , [ 24 , 5 , 15 ] , [ 21 , 6 , 12 ] , [ 22 , 6 , 12 ] , [ 23 , 6 , 13 ] , [ 24 , 6 , 14 ] , and [ 25 , 8 , 12 ] .
This paper is organized as follows. In Section 2, we provide some required basic knowledge on Hermitian LCD codes. We derive constructions of Hermitian LCD codes in Section 3. In Section 4, we discuss Hermitian LCD codes with good parameters.

2. Preliminary

In this section, we introduce some basic concepts on quaternary linear codes. Let F 4 = { 0 , 1 , ω , ϖ } be the Galois field with four elements, with ϖ = 1 + ω = ω 2 , ω 3 = 1 . Denote the n-dimensional space over F 4 by F 4 n ; we call a k-dimensional subspace C of F 4 n a k-dimensional linear code of length n and denote it as C = [ n , k ] . A matrix G whose rows form the basis of C is called a generator matrix of C . If the minimum distance of C is d, then C can be denoted as C = [ n , k , d ] . A code C = [ n , k , d ] is an optimal code if there is no [ n , k , d + 1 ] code. An optimal code is denoted [ n , k , d o ( n , k ) ] in this paper. For a given code [ n , k , d ] , if d is the largest value present known, then C is called the best-known code and also denoted as [ n , k , d o ( n , k ) ] . Denote d l ( n , k ) = m a x { d | as an [ n , k , d ] LCD code}. If a C = [ n , k , d l ( n , k ) ] LCD code saturates the lower or upper bound of Grassl’s code table [24], we call C an optimal LCD code and can say d l ( n , k ) = d o ( n , k ) . If d l ( n , k ) = d o ( n , k ) 1 , we call C a nearly optimal LCD code.
Define the Hermitian inner product of u, v   F 4 n as
( u , v ) h = u v 2 = u 1 v 1 ¯ + u 2 v 2 ¯ + · · · + u n v n ¯ .
The Hermitian dual code of C = [ n , k ] is C h = { x ( x , y ) h = 0 , y C } , and C h = [ n , n k ] . A generator matrix H = H ( n k ) × n of C h is called a parity check matrix of C . If C C h , C is called a weakly self-orthogonal code. If C is a self-orthogonal code, then each generator matrix G of C must satisfy rank( G G ) = 0 , where G is the conjugate transpose of G.
If C C h = { 0 } , then C (or C h ) is called a quaternary Hermitian LCD code, and each generator matrix G of C must satisfy k = rank( G G ) , see refs. [2,4].
In the following sections, we will discuss the construction of Hermitian quaternary LCD code C = [ n , k , d ] , where d is as large as possible for a given n and k 5 . Firstly, we present some notation for later use.
Let 1 n = ( 1 , 1 , , 1 ) 1 × n and 0 n = ( 0 , 0 , , 0 ) 1 × n denote an all-one vector and an all-zero vector of length n, respectively.
Construct
S 2 = 0 1 1 1 1 1 0 1 ω ϖ = ( α 1 , , α 5 ) ,
S 3 = S 2 0 2 × 1 S 2 S 2 S 2 0 5 1 1 5 ω 1 5 ϖ 1 5 = ( β 1 , β 2 , , β 21 ) ,
S 4 = S 3 0 3 × 1 S 3 S 3 S 3 0 21 1 1 21 ω 1 21 ϖ 1 21 = ( γ 1 , γ 2 , , γ 85 ) .
S 5 = S 4 0 4 × 1 S 4 S 4 S 4 0 85 1 1 85 ω 1 85 ϖ 1 85 = ( ζ 1 , ζ 2 , , ζ 341 ) .
S k = S k 1 0 k 1 × 1 S k 1 S k 1 S k 1 0 4 ( k 1 ) 3 1 1 4 ( k 1 ) 3 ω 1 4 ( k 1 ) 3 ϖ 1 4 ( k 1 ) 3
It is well known that the matrix S 2 generates the [ 5 , 2 , 4 ] simplex code with weight polynomial 1 + 15 y 4 . S 3 generates the [ 21 , 3 , 16 ] simplex code with weight polynomial 1 + 63 y 16 . S 4 generates the [ 85 , 4 , 64 ] simplex code with weight polynomial 1 + 255 y 64 , S 5 generates the [ 341 , 5 , 256 ] simplex code with weight polynomial 1 + 1023 y 256 , and S k S k = 0 for k = 2 , 3 , 4 , 5 , , see ref. [28].

Notation

In the following sections, the conjugation is defined by x ¯ = x 2 for x F 4 . We use 2 and 3 to represent ω and ϖ in each generator matrix of linear codes, respectively. An [ n , k , d ] 4 code is denoted as [ n , k , d ] for short.

3. Hermitian LCD Linear Codes over F 4

In this subsection, we discuss the construction of [ n , k ] optimal Hermitian quaternary LCD codes. For k 5 and n 20 , there are some Hermitian quaternary LCD codes in [21, 23, 26, 28]. For 20 n 25 , there is no systematic discussion in the literature. The discussion is presented in four cases for n 25 .
Lemma 1 ([4,29]). 
If 21 n 25 and 1 k 5 , then d l ( 2 , 24 ) = 18 , d l ( 2 , 25 ) = 19 , d l ( 3 , 21 ) = 15 , d l ( 3 , 22 ) = 15 , d l ( 4 , 22 ) = 14 , d l ( 4 , 23 ) = 15 , d l ( 5 , 24 ) = 15 . All the other Hermitian quaternary LCD codes saturate the lower bound of Grassl’s code table [24].
Proof. 
Refs. [4,29] proved this lemma. □
Lemma 2 ([30]). 
There exist [ n , n 2 , 2 ] and [ n , n 3 , 2 ] quaternary Hermitian LCD codes.
Proof. 
(1) For when n is even, let G = I 2 | 1 1 × n 0 1 × n . If G is a check matrix of C with generator matrix H, then C = [ n , n 2 , 2 ] and r a n k ( H H h ) = n 2 . For when n is even, let G = I 2 ϖ ϖ | 1 1 × n 1 × n . If G is a check matrix of C with generator matrix H, then C = [ n , n 2 , 2 ] and r a n k ( H H h ) = n 2 .
(2) For when n is odd, let G = I 3 | 1 1 × n 0 2 × n . If G is a check matrix of C with generator matrix H, then C = [ n , n 3 , 2 ] and r a n k ( H H h ) = n 3 . For when n is even, let G = I 3 ϖ ϖ 0 | 1 1 × n 0 2 × n . If G is a check matrix of C with generator matrix H, then C = [ n , n 3 , 2 ] and r a n k ( H H h ) = n 3 . □
Theorem 1. 
If 21 n 25 and 13 k 18 , then there exist 29 optimal quaternary Hermitian LCD codes saturating the lower bound of Grassl’s code table [24], as in Table 1.
Proof. 
For 21 n 25 , calculating by Magma, one can obtain nine optimal LCD codes as follows: [ 21 , 14 , 5 ] , [ 21 , 15 , 5 ] , [ 21 , 16 , 4 ] , [ 22 , 16 , 4 ] , [ 24 , 16 , 6 ] , [ 22 , 18 , 3 ] , [ 23 , 18 , 4 ] , [ 25 , 18 , 5 ] , [ 23 , 19 , 3 ] .
And then, calculating by Magma, one can obtain another five optimal codes, [ 27 , 19 , 6 ] , [ 26 , 21 , 4 ] , [ 25 , 18 , 5 ] , [ 26 , 20 , 4 ] , [ 33 , 24 , 6 ] , which are not all quaternary Hermitian LCD codes.
Case 1. Construction of quaternary Hermitian LCD codes via puncturing. Puncturing C = [ 23 , 15 , 6 ] on coordinate sets { 16 } , { 1 , 19 } , one can obtain [ 22 , 15 , 5 ] and [ 21 , 17 , 3 ] Hermitian quaternary LCD codes. Puncturing C = [ 24 , 17 , 5 ] on coordinate sets { 1 , 18 } , one can obtain the [ 22 , 17 , 4 ] Hermitian quaternary LCD code. Puncturing C = [ 24 , 19 , 4 ] on coordinate sets { 1 , 4 , 8 } , one can obtain the [ 21 , 19 , 2 ] quaternary Hermitian LCD code.
Case 2. Construction of LCD codes via shortening. Shortening C = [ 27 , 19 , 6 ] on coordinate sets { 1 , 4 } , { 1 , 2 , 3 , 8 } , { 1 , 2 , 3 , 4 , 7 } , { 1 , 2 , 3 , 4 , 7 , 8 } , one can obtain the [ 25 , 17 , 6 ] , [ 23 , 15 , 6 ] , [ 22 , 14 , 6 ] , [ 21 , 13 , 6 ] Hermitian quaternary LCD codes, respectively. Shortening D = [ 26 , 21 , 4 ] on coordinate sets { 1 } , { 1 , 2 } obtain [ 25 , 20 , 4 ] and [ 24 , 19 , 4 ] Hermitian quaternary LCD codes, respectively. Shortening D = [ 25 , 18 , 5 ] on coordinate sets { 4 } and { 1 , 4 } , one can obtain the [ 24 , 17 , 5 ] , [ 23 , 16 , 5 ] Hermitian quaternary LCD codes. Shortening D = [ 26 , 20 , 4 ] on coordinate sets { 1 } , { 1 , 2 } , { 1 , 2 , 4 } , one can obtain the [ 25 , 19 , 4 ] , [ 24 , 18 , 4 ] , [ 23 , 17 , 4 ] Hermitian quaternary LCD codes. Shortening D = [ 33 , 24 , 6 ] on coordinate sets { 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 } and { 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 } , one can obtain the [ 25 , 16 , 6 ] and [ 24 , 15 , 6 ] Hermitian quaternary LCD codes. Shortening D = [ 33 , 24 , 6 ] on coordinate sets { 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 } and { 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 14 } , one can obtain the C = [ 23 , 14 , 6 ] and [ 22 , 13 , 6 ] Hermitian quaternary LCD codes. □
Remark 1. 
In Theorem 2, all of the codes are optimal quaternary Hermitian LCD codes. Since [ 21 , 3 , 16 ] is a simplex code, there is no [ 21 , 18 , 3 ] quaternary Hermitian LCD code. Hence, [ 21 , 18 , 2 ] is an optimal quaternary Hermitian LCD code. By shortening D = [ 33 , 24 , 6 ] on coordinate sets { 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 } , we can obtain C = [ 21 , 12 , 6 ] . This is a nearly optimal quaternary Hermitian LCD code with weight enumerator 1 + 279 z 6 + 1116 z 7 + 5739 z 8 + 22023 z 9 + 79815 z 10 + .
Theorem 2. 
If 21 n 25 and 6 k 8 , then d l ( 6 , 21 ) = 12 , d l ( 6 , 22 ) = 12 , d l ( 6 , 23 ) = 13 , d l ( 6 , 24 ) = 14 , d l ( 8 , 25 ) = 12 , d l ( 7 , 20 ) = 10 . All these codes are quaternary Hermitian LCD codes saturating the lower or upper bound of Grassl’s code table.
Proof. 
A constacyclic code C = [ 21 , 15 , 5 ] is given in [21], where its generator polynomial is x 6 + ω ¯ x 5 + x 4 + ω ¯ x 2 + x + ω ¯ . The dual code of C is the code D = [ 21 , 6 , 12 ] with a generator matrix G 6 , 21 , and both C and D are quaternary Hermitian LCD codes.
Let
G 6 , 21 = 211210221102122100000 303201310313021010000 221130310133220001000 022113031013322000100 320131031302101000010 223203322032332000001 ,
G 6 , 24 = 100000111120112310122020 010000011112011231012202 001000201111101123201220 000100120111310112020122 000010112011231011202012 000001111201123101220201 ,
G 7 , 20 = 11011111101000000011 11120323003110000012 30132223011030000020 11202023001001100001 11000132002013010001 30220233000013001003 30313002003012000101 ,
G 8 , 25 = 1000000010012123213103310 0100000031322111123203311 0010000012120332303223021 0001000021122001100022303 0000100013332221330202233 0000010010113203310220220 0000001001011320330022022 0000000112200212232000033 ,
There exists a quaternary Hermitian LCD code [ 24 , 6 , 14 ] with generator matrix G 6 , 24 . Its weight enumerator is 1 + 207 z 14 + 378 z 15 + 630 z 16 + 360 z 17 + 495 z 18 + 1062 z 19 + 585 z 20 + 180 z 21 + 162 z 22 + 36 z 23 . Puncturing C = [ 24 , 6 , 14 ] on coordinate sets { 7 } , { 1 , 3 } , we can obtain two quaternary Hermitian LCD codes: [ 23 , 6 , 13 ] , [ 22 , 6 , 12 ] .
There exists a quaternary Hermitian LCD code [ 20 , 7 , 10 ] with generator matrix G 7 , 20 . Its weight enumerator is 1 + 210 z 10 + 594 z 11 + 969 z 12 + 1647 z 13 + 2703 z 14 + 3519 z 15 + 3060 z 16 + 2205 z 17 + 1107 z 18 + 291 z 19 + 78 z 20 .
There exists a quaternary Hermitian LCD code [ 25 , 8 , 12 ] with generator matrix G 8 , 25 . Its weight enumerator is 1 + 177 z 12 + 540 z 13 + 1365 z 14 + 2721 z 15 + 4836 z 16 + 8283 z 17 + 10938 z 18 + 11694 z 19 + 10983 z 20 + 7734 z 21 + 4185 z 22 + 1617 z 23 + 411 z 24 + 25 z 25 .
Shortening the [ 25 , 8 , 12 ] quaternary Hermitian LCD code on coordinate sets { 2 } , one can obtain [ 24 , 7 , 12 ] . Its weight enumerator is 1 + 102 z 12 + 267 z 13 + 561 z 14 + 1086 z 15 + 1764 z 16 + 2628 z 17 + 3144 z 18 + 2730 z 19 + 2226 z 20 + 1233 z 21 + 495 z 22 + 120 z 23 + 27 z 24 . We can deduce a submatrix G 7 , 25 from G 8 , 25 . Setting G 7 , 25 as a generator matrix, one can obtain [ 25 , 7 , 12 ] . □
Theorem 3. 
If 21 n 25 and 8 k 15 , then there exist 27 quaternary Hermitian LCD codes, as in Table 2.
Proof. 
Let
A 12 = 300201200120300312 330322220222130112 133330122212313132 213031312011331230 002033213230313123 021303131201133123 300300221033231123 330133122213123323 033110212131112103 203212121103311021 020120012030031210 002012001203003121 ,
A 11 = 1232130030332100 3202331131102132 1100132302320200 1321332231223320 0013213203303002 0131021110330323 1322322221212333 2203012112300132 3122131011102120 0002212120320222 3223331032121221 .
There exists a code [ 30 , 18 , 8 ] with generator matrix G 18 , 30 = I 18 | A 12 . It is not a quaternary Hermitian LCD code. Shortening C = [ 30 , 18 , 8 ] on coordinate sets { 1 , 3 , 6 , 12 , 13 , 17 } , { 1 , 2 , 3 , 4 , 5 , 6 , 8 } , { 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 } and { 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 10 } , one can obtain quaternary Hermitian LCD codes [ 24 , 12 , 8 ] , [ 23 , 11 , 8 ] , [ 22 , 10 , 8 ] and [ 21 , 9 , 8 ] , respectively.
There exists a code [ 27 , 16 , 7 ] with generator matrix G 16 , 27 = I 16 | A 11 . Shortening C = [ 27 , 16 , 7 ] on coordinate sets { 1 , 2 } , { 1 , 2 , 3 } , { 1 , 2 , 3 , 4 } , { 1 , 2 , 3 , 4 , 5 } and { 1 , 2 , 3 , 4 , 5 , 6 } , one can obtain quaternary Hermitian LCD codes [ 25 , 14 , 7 ] , [ 24 , 13 , 7 ] , [ 23 , 12 , 7 ] , [ 22 , 11 , 7 ] and [ 21 , 10 , 7 ] , respectively.
Let A 10 = 11121122120230323210 00000000000000000000 12103313201003102111 11331313032000112032 21133320003300003220 01030322203233222121 32210120113300112032 00331212201211230320 11321302013232210003 11010110212023020323 ,
B 16 = 0331200330200 0033102033020 2330330112100 0233003311210 1201303302020 2213122121000 0221321212100 0022113221210 1220202211020 2211033130300 0221130313030 3133123113000 0313331211300 0031313321130 3112111310210 1133222202120 ,
There exists a code [ 30 , 20 , 6 ] with generator matrix G 20 , 30 = I 20 | A 10 . Shortening C = [ 30 , 20 , 6 ] on coordinate sets { 1 , 2 , 3 , 4 , 5 } , { 1 , 2 , 3 , 4 , 5 , 6 } , { 1 , 2 , 3 , 5 , 6 , 7 , 13 } , { 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 } and { 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 } , one can obtain quaternary Hermitian LCD codes [ 25 , 15 , 6 ] , [ 24 , 14 , 6 ] , [ 23 , 13 , 6 ] , [ 22 , 12 , 6 ] and [ 21 , 11 , 6 ] , respectively.
There exists a code [ 29 , 16 , 8 ] with generator matrix G 16 , 29 = I 16 | B 16 . It is not a quaternary Hermitian LCD code. Shortening C = [ 29 , 16 , 8 ] on coordinate sets { 1 , 2 , 3 , 4 } , { 1 , 2 , 3 , 4 , 6 } , { 1 , 2 , 3 , 4 , 5 , 6 } , { 1 , 2 , 3 , 4 , 5 , 6 , 8 } , { 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 } , one can obtain quaternary Hermitian LCD codes [ 25 , 12 , 8 ] , [ 24 , 11 , 8 ] , [ 23 , 10 , 8 ] , [ 22 , 9 , 8 ] and [ 21 , 8 , 8 ] , respectively.
Let A 14 = 13000123002211 10211103211330 01021110321133 33231222301220 03323122230122 22110130001230 02211013000123 33112232033321 12200332312223 32113300102111 12300221101300 01230022110130 00123002211013 33321033112232 21110321133001 11111111111111 ,
A 15 = 102333010301203 210033301030123 122000202121223 013221122033211 202131213310100 220013121331010 223330023223223 123212100103011 113310121300022 212032120112232 322220031221030 033023210111210 000220311212022 000312113202231 000121022211230 000102331310131 ,
A 12 = 333231302012 122021230233 012231331333 110121333101 322213123331 123133323323 321131111021 301331322211 212033023133 021232013200 113202212020 000012102221 000001110222 ,
There exists a code [ 30 , 16 , 9 ] with generator matrix G 16 , 30 = I 16 | A 14 . It is not a quaternary Hermitian LCD code. Shortening C = [ 30 , 16 , 9 ] on coordinate sets { 1 , 2 , 3 , 4 , 5 } , { 1 , 2 , 3 , 4 , 5 , 11 } , { 1 , 2 , 3 , 4 , 5 , 6 , 11 } and { 1 , 2 , 3 , 4 , 5 , 6 , 11 } , one can obtain quaternary Hermitian LCD codes [ 25 , 11 , 9 ] , [ 24 , 10 , 9 ] , [ 23 , 9 , 9 ] and [ 22 , 8 , 9 ] , respectively.
There exists a code [ 31 , 16 , 10 ] with generator matrix G 16 , 31 = I 16 | A 15 . It is not a quaternary Hermitian LCD code. Shortening C = [ 31 , 16 , 10 ] on coordinate sets { 8 , 9 , 10 , 11 , 16 } , { 1 , 6 , 9 , 10 , 12 , 15 } , { 2 , 4 , 5 , 9 , 10 , 13 , 15 } and { 3 , 4 , 7 , 8 , 11 , 12 , 13 , 15 } , one can obtain quaternary Hermitian LCD codes [ 26 , 11 , 10 ] , [ 25 , 10 , 10 ] , [ 24 , 9 , 10 ] and [ 23 , 8 , 11 ] , respectively. Puncturing C = [ 23 , 8 , 11 ] on coordinate sets { 2 } and { 1 , 2 } , one can obtain quaternary Hermitian LCD codes [ 22 , 8 , 10 ] and [ 21 , 8 , 9 ] , respectively.
There exists a quaternary Hermitian LCD code [ 25 , 13 , 7 ] with generator matrix G 13 , 25 = I 13 | A 12 . □
Theorem 4. 
d l ( 7 , 21 ) = 10 , d l ( 7 , 22 ) = 11 , d l ( 7 , 22 ) = 11 , d l ( 7 , 23 ) = 12 , d l ( 7 , 25 ) = 13 , d l ( 6 , 25 ) = 14 , d l ( 12 , 25 ) = 8 , d l ( 13 , 25 ) = 7 , d l ( 7 , 18 ) = 9 and d l ( 7 , 19 ) = 9 are Hermitian quaternary LCD codes.
Proof. 
Let
G 7 , 24 = 022312110003211100000321 102231211000321110000032 110223121100032121000003 211022311110003232100000 121102232111000303210000 312110223211100000321000 231211020321110000032100 ,
G 7 , 25 = 1100000032031300131303121 0010000003203130213130311 0001000003123223332223321 0000100033111212020312021 0000010020112011011101111 0000001011212331312020201 0000000112322303222332331 ,
G 7 , 19 = 1130121231010000000 0133203320031200000 0303332320310100000 0131030320010011000 0130001110020230300 0301202120000230010 0302110010030220003 ,
G 6 , 25 = 3111201330100001230133122 1111110133030000122013312 2311100013323000012201331 1131131001312300001220133 1213130100101230003122013 1121313010000123003312201
There exists a code [ 24 , 7 , 13 ] with generator matrix G 7 , 24 . Its weight enumerator is 1 + 384 z 13 + 744 z 14 + 888 z 15 + 1746 z 16 + 2544 z 17 + 3156 z 18 + 2928 z 19 + 2118 z 20 + 1200 z 21 + 540 z 22 + 120 z 23 + 15 z 24 . It is not a quaternary Hermitian LCD code. Puncturing C 1 on coordinate sets { 1 } , { 1 , 3 } , { 1 , 2 , 7 } , we can obtain [ 23 , 7 , 12 ] , [ 22 , 7 , 11 ] , [ 21 , 7 , 10 ] quaternary Hermitian LCD codes.
There exists a quaternary Hermitian LCD code [ 25 , 7 , 13 ] with generator matrix G 7 , 25 . Its weight enumerator is 1 + 189 z 13 + 495 z 14 + 750 z 15 + 1179 z 16 + 1908 z 17 + 2577 z 18 + 2967 z 19 + 2667 z 20 + 1932 z 21 + 1092 z 22 + 495 z 23 + 117 z 24 + 15 z 25 .
There exists a quaternary Hermitian LCD code [ 25 , 6 , 14 ] with generator matrix G 6 , 25 . Its weight enumerator is 1 + 48 z 14 + 240 z 15 + 432 z 16 + 534 z 17 + 573 z 18 + 648 z 19 + 657 z 20 + 510 z 21 + 363 z 22 + 84 z 23 + 6 z 24 .
There exists an optimal quaternary Hermitian LCD code [ 19 , 7 , 9 ] with generator matrix G 7 , 19 . Its weight enumerator is 1 + 195 z 9 + 483 z 10 + 888 z 11 + 1479 z 12 + 2361 z 13 + 3165 z 14 + 3327 z 15 + 2508 z 16 + 1368 z 17 + 492 z 18 + 117 z 19 . Puncturing the quaternary Hermitian LCD code [ 19 , 7 , 9 ] on coordinate sets { 1 } , one can obtain the optimal quaternary Hermitian LCD code [ 18 , 7 , 9 ] with weight enumerator 1 + 393 z 9 + 666 z 10 + 1245 z 11 + 2193 z 12 + 3315 z 13 + 3597 z 14 + 2799 z 15 + 1554 z 16 + 504 z 17 + 117 z 18 . □

4. Discussion and Conclusions

This paper is dedicated to the construction of quaternary Hermitian LCD codes. For k n and n 25 , each [ n , k ] quaternary Hermitian LCD code is constructed. Some of these quaternary Hermitian LCD codes constructed in this paper are optimal codes which saturate the bound of the minimum distance of the code table in [24], and some of them are nearly optimal codes. According to weight enumerators for classification codes in [27], there exist some optimal codes, [ 15 , 4 , 9 ] , [ 16 , 4 , 10 ] , [ 17 , 4 , 11 ] , [ 19 , 4 , 14 ] , and [ 23 , 5 , 15 ] , which are not LCD codes. In addition, the number of these five optimal codes is one. Thus, the [ 15 , 4 , 8 ] , [ 16 , 4 , 9 ] , [ 17 , 4 , 10 ] , [ 19 , 4 , 13 ] , and [ 23 , 5 , 14 ] quaternary Hermitian LCD codes in this paper are optimal. In [27], all of the codes with parameters of [ 15 , 6 , 8 ] are self-orthogonal. Thus, the quaternary Hermitian LCD code in this paper, [ 15 , 6 , 7 ] , is also optimal. We emphasize that there are three quaternary Hermitian LCD codes, [ 18 , 7 , 9 ] , [ 19 , 7 , 9 ] and [ 20 , 7 , 10 ] , which are optimal.
According to ref. [24], the following quaternary Hermitian LCD codes constructed in this paper are also optimal codes with parameters of [ n , k , d ] for 21 n 25 and 16 k 18 ; [ n i , 15 i , d ] for 21 n 24 and 0 i 2 : [ 23 , 4 , 15 ] , [ 24 , 5 , 15 ] , [ 21 , 6 , 12 ] , [ 22 , 6 , 12 ] , [ 23 , 6 , 13 ] , [ 24 , 6 , 14 ] , [ 25 , 8 , 12 ] and [ 20 , 7 , 10 ] . Except for these codes mentioned above, the quaternary Hermitian LCD codes constructed in this paper do not reach the known upper or lower bounds of the minimum distance of a linear code. Nonetheless, the minimum distances of these codes appears to be the best possible. These codes are the best possible among those obtainable by our approach.
Combining the results in the previous subsections, we improved the table of lower and upper bounds on the minimum distance of quaternary Hermitian LCD codes for n 20 [4,25,30] in Table 3. In addition, many lower and upper bounds of the minimal distance of Hermitian LCD codes with a length of n 25 are listed. To make the bounds in Table 3 tighter, we need to choose other quaternary Hermitian LCD codes better than those given in this paper and investigate other code constructions to raise the lower bounds. We also plan to explore the construction of Hermitian LCD codes from a geometric aspect to decrease the upper bounds.
In [4,20,25], it has been shown that if there exists a quaternary Hermitian [ n , k , d ] code over F q 2 , then there exists a maximal entanglement entanglement-assisted quantum error correcting code (EAQECC) over F q with parameters [ [ n , 2 k n + c , d ; c ] ] , where c is the rank of the product of the parity check matrix and its conjugate. Moreover, a maximal entanglement EAQECC derived from an LCD code has the same minimum distance as the underlying classical code. Hence, all of the optimal quaternary Hermitian LCD codes can be used to construct optimal binary maximal entanglement EAQECCs. In addition, from the three quaternary Hermitian LCD codes [ 18 , 7 , 9 ] given in this paper, a maximal entanglement EAQECC [ [ 18 , 7 , 9 ; 11 ] ] can be constructed, which improves the minimal distance of the codes in [4,25]. The maximal entanglement EAQECCs [ [ 19 , 7 , 9 ; 12 ] ] and [ [ 20 , 7 , 10 ; 12 ] ] are optimal and are different to the codes constructed in [30].

Author Contributions

L.L. completed research design, data analysis, and the leading of writing of the article; R.L. provided important guidance in solving difficult or complex problems in the manuscript; Y.R. proofread all drafts. All authors contributed to the writing and revision of the article, providing constructive suggestions and improvements to ensure that the article accurately expresses complex research findings. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by National Natural Science Foundation of China: U21A20428; Natural Science Foundation of Shaanxi: 2023-JC-YB-003 and 2023-JC-QN-0033.

Data Availability Statement

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

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Massey, J.L. Reversible codes. Inf. Control 1964, 7, 369–380. [Google Scholar] [CrossRef]
  2. Massey, J.L. Linear codes with complementary duals. Discret. Math. 1992, 106–107, 337–342. [Google Scholar] [CrossRef]
  3. Carlet, C.; Guilley, S. Complementary dual codes for counter-measures to side-channel attacks. Adv. Math. Commun. 2016, 10, 131–150. [Google Scholar] [CrossRef]
  4. Lu, L.; Li, R.; Guo, L.; Fu, Q. Maximal entanglement entanglement-assisted quantum codes constructed from linear codes. Quantum Inf. Process. 2015, 14, 165–182. [Google Scholar] [CrossRef]
  5. Lv, L.; Li, R.; Fu, Q.; Li, X. Maximal entanglement entanglement-assisted quantum codes constructed from quaternary BCH codes. In Proceedings of the IEEE Advanced Information Technology, Electronic and Automation Control Conference, Chongqing, China, 19–20 December 2015. [Google Scholar]
  6. Ding, C.; Li, C.; Li, S. LCD cyclic codes over finite fields. IEEE Trans. Inf. Theory 2018, 63, 4344–4356. [Google Scholar]
  7. Güneri, C.; Özkaya, B.; Solé, P. Quasi-cyclic complementary dual codes. Finite Fields Their Appl. 2016, 42, 67–80. [Google Scholar] [CrossRef]
  8. Yang, X.; Massey, J.L. The necessary and sufficient condition for a cyclic code to have a complementary dual. Discret. Math. 1994, 126, 391–393. [Google Scholar] [CrossRef]
  9. Shi, M.; Özbudak, F.; Xu, L.; Solé, P. LCD codes from tridiagonal Toeplitz matrices. Finite Fields Their Appl. 2021, 75, 101892. [Google Scholar] [CrossRef]
  10. Shi, M.; Huang, D.; Sok, L.; Solé, P. Double circulant LCD codes over Z4. Finite Fields Their Appl. 2019, 58, 133–144. [Google Scholar] [CrossRef]
  11. Sok, L.; Shi, M.; Solé, P. Construction of optimal LCD codes over large finite fields. Finite Fields Their Appl. 2018, 50, 138–153. [Google Scholar] [CrossRef]
  12. Shi, M.; Li, S.; Kim, J.; Solé, P. LCD and ACD codes over a noncommutative non-unital ring with four elements. Cryptogr. Commun. 2022, 14, 627–640. [Google Scholar] [CrossRef]
  13. Hou, X.; Oggier, F. On LCD codes and lattices. In Proceedings of the IEEE International Symposium on Information Theory (ISIT), Barcelona, Spain, 10–15 July 2016; pp. 1501–1505. [Google Scholar] [CrossRef]
  14. Lina, E.R.L.; Nocon, E.G. On the construction of some LCD codes over finite fields. Manila J. Sci. 2016, 9, 67–82. [Google Scholar]
  15. Zhu, S.; Pang, B.; Sun, Z. The reversible negacyclic codes over finite fields. arXiv 2016, arXiv:1610.08206v1. [Google Scholar]
  16. Galvez, L.; Kim, J.; Lee, N.; Roe, Y.; Won, B. Some bounds on binary LCD codes. Cryptogr. Commun. 2018, 10, 719–728. [Google Scholar] [CrossRef]
  17. Araya, M.; Harada, M. On the classification of linear complementary dual codes. Discret. Math. 2019, 342, 270–278. [Google Scholar] [CrossRef]
  18. Araya, M.; Harada, M. On the minimum weights of binary linear complementary dual codes. Cryptogr. Commun. 2019, 12, 285–300. [Google Scholar] [CrossRef]
  19. Fu, Q.; Li, R.; Fu, F.; Rao, Y. On the Construction of Binary Optimal LCD Codes with Short Length. Int. J. Found. Comput. Sci. 2019, 30, 1237–1245. [Google Scholar] [CrossRef]
  20. Lai, C.; Brun, T.; Wilde, M. Duality in Entanglement-assisted quantum error correction. IEEE Trans. Inf. Theory 2013, 59, 4020–4024. [Google Scholar] [CrossRef]
  21. Dougherty, S.T.; Kim, J.-L.; Ozkaya, B.; Sok, L.; Solé, P. The combinatorics of LCD codes: Linear Programming bound and orthogonal matrices. Int. J. Inf. Coding Theory 2017, 4, 116–128. [Google Scholar] [CrossRef]
  22. Carlet, C.; Mesnager, S.; Tang, C.; Qi, Y.; Pellikaan, R. Linear Codes Over Fq Are Equivalent to LCD Codes for q > 3. IEEE Trans. Inf. Theory 2018, 64, 3010–3017. [Google Scholar] [CrossRef]
  23. Araya, M.; Harada, M.; Saito, K. Quaternary Hermitian linear complementary dual codes. IEEE Trans. Inf. Theory 2020, 66, 2751–2759. [Google Scholar] [CrossRef]
  24. Grassl, M. Bounds on the Minimum Distance of Linear Codes. 2024. Available online: http://www.codetables.de (accessed on 28 February 2024).
  25. Lai, C.; Ashikhmin, A. Linear Programming Bounds for Entanglement-Assisted Quantum Error-Correcting Codes by Split Weight Enumerators. IEEE Trans. Inf. Theory 2018, 64, 622–639. [Google Scholar] [CrossRef]
  26. Kschischang, F.; Pasupathy, S. Some ternary and quantum codes and associated sphere pachings. IEEE Trans. Inform. Theory 1992, 38, 227–246. [Google Scholar] [CrossRef]
  27. Bouyukliev, L.; Grassl, M.; Varbanov, Z. New bounds for n4(k, d) and classification of some optimal codes over GF(4). Discret. Math. 2004, 281, 43–66. [Google Scholar] [CrossRef]
  28. Li, R. Research on Additive Quantum Error Correcting Codes. Ph.D. Thesis, Northwest Poly-Technical University, Xi’an, China, 2004. [Google Scholar]
  29. Lu, L.; Li, R.; Guo, L. Entanglement-assisted quantum codes from quaternary codes of dimension five. Int. J. Quantum Inf. 2017, 14, 1750017. [Google Scholar] [CrossRef]
  30. Harada, M. Some optimal entanglement-assisted quantum codes constructed from quaternary Hermitian linear complementary dual codes. Int. Quantum Inf. 2019, 17, 1950053. [Google Scholar] [CrossRef]
Table 1.  Optimal quaternary Hermitian LCD codes with 21 n 25 and 13 k 19 . 
Table 1.  Optimal quaternary Hermitian LCD codes with 21 n 25 and 13 k 19 . 
n k 13141516171819
216554322
226654432
23 665443
24 66544
25 6654
Table 2.  Optimal quaternary Hermitian LCD codes with 21 n 25 and 8 k 15 . 
Table 2.  Optimal quaternary Hermitian LCD codes with 21 n 25 and 8 k 15 . 
n k 89101112131415
219876
22108876
231198876
24111098876
25 111098776
Table 3.  Lower and upper bounds on the minimum distance of quaternary Hermitian LCD codes. The bold entries represent improvements over prior work. The superscript * represents the codes that achieve bounds given in the Grassl table. 
Table 3.  Lower and upper bounds on the minimum distance of quaternary Hermitian LCD codes. The bold entries represent improvements over prior work. The superscript * represents the codes that achieve bounds given in the Grassl table. 
n k 123456789101112
33 *2 *
43 *21
55 *322 *
65 *4 *32 *1
77 *5 *4 *3 *2 *2 *
87 *6 *5 *4 *3 *2 *1
99 *66 *5 *4 *3 *2 *2 *
109 *76 *6 *5 *4 *3 *2 *11
1111 *8 *7 *6 *6 *5 *4 *3 *2 *2 *1
1211 *9 *8 *7 *6 *54 *4 *3 *2 *2 *1
1313 *10 *9 *8 *7 *6 *5 *4 *4 *3 *2 *2 *
1413 *10987–87 *6 *5 *4 *4 *3 *2 *
1515 *111098 *77 *6 *5 *4 *4 *3 *
1615 *12 *11109 *8 *7–86–76 *5 *4 *4 *
1717 *13 *12 *1110 *9 *7–87–86–76 *5 *4 *
1817 *14 *13 *11–1210 *9–109 *8 *7–86 *5–65 *
1919 *141312 *–1311 *10 *9 *8 *–98 *7 *6 *5–6
2019 *151413 *12 *11 *10 *9 *8 *–97–86–76 *
2121 *16 *1514 *1212 *10–119–108–97–96–86–7
2222 *17 *15141312 *–1311–12108–108–97–96–8
2323 *18 *16 *151413 *12 *–13119–118–108–97–9
2424 *1817 *16 *1514 *12–1311–1310–129–118–108–9
2525 *1918 *17 *1514–1513–1412 *–1311–1310–129–118–10
n k 131415161718192021222324
141
152 *2 *
163 *2 *1
173–43 *2 *2 *
184 *3 *3 *2 *1
195 *4 *3 *3 *2 *2 *
205–65 *4 *3 *22 *11
216 *5 *–65 *4 *3 *22 *2 *1
226 *–76 *5 *–64 *–54 *3 *2 *2 *21
236–86 *–76 *5 *–64 *–54 *2 *2 *2 *2 *1 *
247–96–86 *–76 *5 *–64 *–53 *3 *2 *2 *2 *1 *
257–97–96–86 *–76 *5 *–64 *4 *3 *2 *2 *2 *
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

Lu, L.; Li, R.; Ren, Y. Optimal Quaternary Hermitian LCD Codes. Entropy 2024, 26, 373. https://doi.org/10.3390/e26050373

AMA Style

Lu L, Li R, Ren Y. Optimal Quaternary Hermitian LCD Codes. Entropy. 2024; 26(5):373. https://doi.org/10.3390/e26050373

Chicago/Turabian Style

Lu, Liangdong, Ruihu Li, and Yuezhen Ren. 2024. "Optimal Quaternary Hermitian LCD Codes" Entropy 26, no. 5: 373. https://doi.org/10.3390/e26050373

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