Next Article in Journal
Using High-Frequency Entropy to Forecast Bitcoin’s Daily Value at Risk
Next Article in Special Issue
Centroid-Based Clustering with αβ-Divergences
Previous Article in Journal
Quantifying Data Dependencies with Rényi Mutual Information and Minimum Spanning Trees
Previous Article in Special Issue
Quaternion Entropy for Analysis of Gait Data
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

New Construction of Maximum Distance Separable (MDS) Self-Dual Codes over Finite Fields

Department of Mathematical Sciences, Xi’an University of Technology, Xi’an 710054, China
*
Author to whom correspondence should be addressed.
Entropy 2019, 21(2), 101; https://doi.org/10.3390/e21020101
Submission received: 24 December 2018 / Revised: 10 January 2019 / Accepted: 17 January 2019 / Published: 22 January 2019
(This article belongs to the Special Issue Information Theory Applications in Signal Processing)

Abstract

:
Maximum distance separable (MDS) self-dual codes have useful properties due to their optimality with respect to the Singleton bound and its self-duality. MDS self-dual codes are completely determined by the length n , so the problem of constructing q-ary MDS self-dual codes with various lengths is a very interesting topic. Recently X. Fang et al. using a method given in previous research, where several classes of new MDS self-dual codes were constructed through (extended) generalized Reed-Solomon codes, in this paper, based on the method given in we achieve several classes of MDS self-dual codes.

1. Introduction

Let F q be the finite field with q elements. A q-ary [ n , k , d ] linear code C is a k-dimensional subspace of F q n with minimum (Hamming) distance d . If the parameters [ n , k , d ] satisfy k + d = n + 1 , the code is called an MDS (maximum distance separable) code. A self-dual code is a linear code satisfying C = C . A linear complementary-dual code is a linear code satisfying C C = { 0 } .
The study of MDS self-dual codes has attracted a great deal of attention in recent years due to its theoretical and practical importance. The center of the study of MDS codes includes the existence of MDS codes [1], classification of MDS codes [2], balanced MDS codes [3], non-Reed-Solomon MDS codes [4], complementary-dual MDS codes [5,6], and lowest density MDS codes [7].
As the parameters of an MDS self-dual code are completely determined by the code’s length n, the main interest here is to determine the existence and give the construction of q-ary MDS self-dual codes for various lengths. The problem is completely solved for the case where q is even [8]. Many MDS self-dual codes over finite fields of odd characteristics were constructed [9,10,11,12,13,14].
In [11], Jin and Xing constructed several classes of MDS self-dual code from generalized Reed-Solomon code. Yan generalized Jin and Xing’s method and constructed several classes of MDS self-dual codes via generalized Reed-Solomon codes and extended generalized Reed-Solomon codes [14]. In [12], Ladad, Liu and Luo produced more classes of MDS self-dual codes based on [11] and [14]. In [9], based on the [11,12,14] more new parameter MDS self-dual codes were presented. Based on the method raised in [9], we present some classes of MDS self-dual codes.

2. Preliminaries

In this section we introduce some basic notations of generalized Reed-Solomon codes and extended generalized Reed-Solomon codes. For more details, the reader is referred to [15].
Throughout this paper, q is a prime power, F q is the finite fields with q elements and let n be a positive integer with 1 < n q . For any x F q 2 , we denote by x ¯ the conjugation of x . Given an [ n , k , d ] linear code C , its Euclidean dual code (resp. Hermitian dual code) is denoted by C (resp. C H ). The codes C and C H are defined by
C = { x = ( x 1 , x 2 , , x n ) F q n : i = 1 n x i y i = 0 , y = ( y 1 , y 2 , , y n ) C } ,
C H = { x = ( x 1 , x 2 , , x n ) F q 2 n : i = 1 n x i y i ¯ = 0 , y = ( y 1 , y 2 , , y n ) C } ,
respectively. In this paper, we only consider the Euclidean inner product.
Let a = ( α 1 , α 2 , , α n ) , where α 1 , α 2 , , α n are n distinct elements of F q . Fix n nonzero elements v 1 , v 2 , , v n of F q ( v i are not necessarily distinct), put v = ( v 1 , v 2 , , v n ) . For 1 k n , the k-dimensional generalized Reed-Solomon code (GRS for short) of length n associated with a and v is defined to be
GRS k ( a , v ) = { ( v 1 f ( α 1 ) , v 2 f ( α 2 ) , , v n f ( α n ) ) : f ( x ) F q [ x ] , deg ( f ( x ) ) k 1 } .
It is well known that the code GRS k ( a , v ) is a q-ary [ n , k , n k + 1 ] MDS code and the dual of a GRS code is again a GRS MDS code; indeed
GRS k ( a , v ) = GRS n k ( a , v )
for some v = ( v 1 , v 2 , , v n ) with v i 0 for all 1 i n (e.g., see [15]).
Furthermore, the extended generalized Reed-Solomon code GRS k ( a , v , ) given by
GRS k ( a , v , ) = { ( v 1 f ( α 1 ) , v 2 f ( α 2 ) , , v n f ( α n ) , f k 1 ) : f ( x ) F q [ x ] , deg ( f ( x ) ) k 1 } ,
where f k 1 stands for the coefficient of x k 1 in f ( x ) . It is also well known that GRS k ( a , v , ) is a q-ary [ n + 1 , k , n k + 2 ] MDS code and the dual code is also a GRS MDS code (e.g., see [15]).
Put a = ( α 1 , α 2 , , α n ) and denote by A a the matrix
1 1 1 α 1 α 2 α n α 1 2 α 2 2 α n 2 α 1 n 2 α 2 n 2 α n n 2
Lemma 1
([11]). The solution space of the equation system A a X T = 0 has dimension 1 and { u = ( u 1 , u 2 , , u n ) } is a basis of this solution space, where u i = 1 j n , j i ( α i α j ) 1 . Furthermore, for any two polynomials f ( x ) , g ( x ) F q [ x ] with deg ( f ) k 1 and deg ( g ) n k 1 , one has i = 1 n f ( α i ) ( u i g ( α i ) ) = 0 .
We define
L a ( α i ) = 1 j n , j i ( α i α j ) .
The conclusion of the following lemma is straightforward. For completeness, we provide its proof.
Lemma 2
([11]). Let n be an even number, if there exists λ F q * such that λ L a ( α i ) is square element for all i = 1 , 2 , , n , then the code GRS n / 2 ( a , v ) defined in (1) is MDS self-dual code of length n.
Proof. 
Let f ( x ) , g ( x ) F q [ x ] with deg ( f ) n 2 1 and deg ( g ) n 2 1 . By Lemma 1, we have i = 1 n f ( α i ) ( u i g ( α i ) ) = 0 , where u i = 1 j n , j i ( α i α j ) 1 for i = 1 , 2 , , n . Hence,
0 = λ i = 1 n f ( α i ) ( u i g ( α i ) ) = i = 1 n f ( α i ) ( λ u i g ( α i ) ) = i = 1 n ( v i f ( α i ) ) ( v i g ( α i ) ) ( since λ u i = v i 2 ) .
This implies that GRS n / 2 ( a , v ) = GRS n / 2 ( a , v ) .  □
H. Yan [14] observed the following two results.
Lemma 3
([14]). Let n be an even integer and k = n 2 . If L a ( α i ) is square element for all i = 1 , 2 , , n 1 , then the code GRS k ( a , v , ) defined in (2) is MDS self-dual code of length n .
Lemma 4
([14]). Let m q 1 be a positive integer and let α F q be a primitive m-th root of unity. Then for any 1 i m , we have
1 j m , j i ( α i α j ) = m α i .

3. Main Result

Let q = r 2 , where r is odd prime power, F q be the finite fields with q elements. Suppose m q 1 , α is a primitive m-th root of unity and H = < β > is the cyclic group generated by β .
Theorem 1.
Let q = r 2 , where r is an odd prime power, r 1 ( mod 4 ) . Suppose that m ( q 1 ) and q 1 m is even, m 0 ( mod 4 ) . If 1 t 2 ( r + 1 ) g c d ( 2 ( r + 1 ) , m ) . Then there exists an [ n = t m , n 2 ] -MDS self-dual code.
Proof. 
Let α be a primitive m-th root of unity and H = < β > is the cyclic group of order 2 ( r + 1 ) . By the theorem of group homomorphism,
( H × α ) / α H / ( H α ) .
Let i 1 , i 2 , , i t be t distinct elements, such that 0 i 1 < i 2 < < i t < 2 ( r + 1 ) . Denote I = { i 1 , i 2 , , i t } , A = i 1 + i 2 + + i t and B = { β i 1 , β i 2 , , β i t } be a set of coset representatives of ( H × α ) / α . Let
a = ( α β i 1 , , α m β i 1 , α β i 2 , , α m β i 2 , , α β i t , , α m β i t ) .
Then the entries of a are distinct in F q * .
It is known that x m y m = j = 1 m ( x α j y ) . By the statement of Lemma 3, we get
L a ( β z α k ) = 1 j m , j k ( β z α k β z α j ) l I , l z j = 1 m ( β z α k β l α j ) = β z ( m 1 ) 1 j m , j k ( α k α j ) l I , l z [ ( β z α k ) m β l m ] = β z ( m 1 ) m α k l I , l z ( β z m β l m ) .
Let v = l I , l z ( β z m β l m ) , then
v r = l I , l z ( β z m r β l m r ) ( since β 2 ( r + 1 ) = 1 , β r = β 1 ) = l I , l z [ ( β 1 ) z m ( β 1 ) l m ] = l I , l z [ ( β 1 ) z m ( β 1 ) l m ] = l I , l z ( β 1 ) z m + l m ( β l m β z m ) = ( 1 ) t 1 β ( A + ( t 2 ) z ) m v
So v r 1 = ( 1 ) t 1 β ( A + ( t 2 ) z ) m .
Let g be a generator of F q * , then α = g q 1 m , β = g r 1 2 , 1 = g r 2 1 2 , v = g r + 1 2 ( t 1 ) ( A + ( t 2 ) z ) m 2 + i ( r + 1 ) . Note that β , m and α are square elements of F q * , we take λ = g r + 1 2 ( t 1 ) , then λ L a ( β z α k ) is a square element of F q * .
This implies there exists a q-ary [ n , n 2 ] MDS self-dual code. □
Example 1.
Let r = 173 , q = 173 2 , r 1 ( mod 4 ) , m = 4 × 43 , q 1 m = 174 is even. For 1 t 2 ( r + 1 ) g c d ( 2 ( r + 1 ) , m ) = 87 , we choose t = 81 . By Theorem 1, there exists the MDS self-dual code with length n = m t = 13,932.
Theorem 2.
Let q = r 2 , where r is an odd prime power. Suppose that m is odd, m ( q 1 ) and q 1 m is even. If 1 t m i n { r + 1 g c d ( 2 ( r + 1 ) , m ) , r + 1 2 } and t is odd, then there exists a q-ary [ n = t m + 1 , n 2 ] MDS self-dual code over F q .
Proof. 
Let α and β be the same as in Theorem 1, we choose t distinct even number i 1 , i 2 , , i t , 0 i 1 < i 2 < < i t < 2 ( r + 1 ) . Denote I = { i 1 , i 2 , , i t } , A = i 1 + i 2 + + i t . Suppose all i j 2 ( mod 4 ) , j = 1 , 2 , , t . The proof is as similar as in Theorem 1. We get
L a ( β z α k ) = β z ( m 1 ) m α k l I , l z ( β z m β l m ) .
Let v = l I , l z ( β z m β l m ) , then we get
v r 1 = ( 1 ) t 1 β ( A + ( t 2 ) z ) m , v = g r + 1 2 ( t 1 ) ( A + ( t 2 ) z ) m 2 + i ( r + 1 ) ,
since A + ( t 2 ) z 2 is even, it implies that v is a square element of F q * . So L a ( β z α k ) is square element of F q * . By Lemma 3, there exists a q-ary [ n , n 2 ] MDS self-dual code. □
Example 2.
Let r = 67 , q = 67 2 , m = 11 , q 1 m = 408 is even. Since 2 ( r + 1 ) = 136 = 4 × 34 , for 1 t r + 1 g c d ( 2 ( r + 1 ) , m ) = 68 , we choose t = 27 . By Theorem 2, there exists the MDS self-dual code with length n = m t + 1 = 298 .
Theorem 3.
Let q = r 2 , where r is an odd prime power, r 1 ( mod 4 ) . Suppose that m is odd, m ( q 1 ) and q 1 m is even. If 1 t m i n { r + 1 g c d ( 2 ( r + 1 ) , m ) , r + 1 2 } and t is odd, then there exists a q-ary [ n = t m + 1 , n 2 ] MDS self-dual code over F q .
Proof. 
Let α and β be the same as in Theorem 1, we choose t distinct even number i 1 , i 2 , , i t , 0 i 1 < i 2 < < i t < 2 ( r + 1 ) . Denote I = { i 1 , i 2 , , i t } , A = i 1 + i 2 + + i t , and i j 2 ( mod 4 ) , j = 1 , 2 , , t . We define the generalized Reed -Solomon code GRS k ( a , v ) with
a = ( 0 , α β i 1 , , α m β i 1 , α β i 2 , , α m β i 2 , , α β i t , , α m β i t ) .
For any z I and 1 k m , we get
L a ( β z α k ) = β z α k 1 j m , j k ( β z α k β z α j ) l I , l z j = 1 m ( β z α k β l α j ) = β z m m l I , l z ( β z m β l m )
and
L a ( 0 ) = l I j = 1 m ( 0 β l α j ) = ( 1 ) m t α m ( m + 1 ) 2 ( l I β l ) m .
Since r 1 ( mod 4 ) , q 1 m is even, so α , β , m , 1 are square elements of F q * , we only need to consider v = l I , l z ( β z m β l m ) . As the calculation in the proof of Theorem 1, v = g r + 1 2 ( t 1 ) ( A + ( t 2 ) z ) m 2 + i ( r + 1 ) . Since all i j 2 ( mod 4 ) and t is odd, so ( A + ( t 2 ) z ) m 2 is even. L a ( β z α k ) , L a ( 0 ) are square elements of F q * . By Lemma 2, there exists a q-ary [ n , n 2 ] MDS self-dual code. □
Example 3.
Let r = 101 , r 1 ( mod 4 ) , q = 101 2 , m = 75 , q 1 m = 136 is even. Since 2 ( r + 1 ) = 204 = 4 × 51 , for 1 t r + 1 g c d ( 2 ( r + 1 ) , m ) = 34 , we choose t = 33 . By Theorem 2, there exists the MDS self-dual code with length n = m t + 1 = 2476 .
Theorem 4.
Let q = r 2 , where r is an odd prime power. Suppose that m ( q 1 ) , q 1 m is even. If 1 t 2 ( r + 1 ) g c d ( 2 ( r + 1 ) , m ) and t m is even, then there exists a q-ary [ n = t m + 2 , n 2 ] MDS self-dual code over F q .
Proof. 
Let α and β be the same as in Theorem 1. We define the extended generalized Reed -Solomon code GRS k ( a , v , ) with
a = ( 0 , α β i 1 , , α m β i 1 , α β i 2 , , α m β i 2 , , α β i t , , α m β i t ) .
For any z I and 1 k m , we get
L a ( β z α k ) = β z α k 1 j m , j k ( β z α k β z α j ) l I , l z j = 1 m ( β z α k β l α j ) = β z m m l I , l z ( β z m β l m )
and
L a ( 0 ) = l I j = 1 m ( 0 β l α j ) = ( 1 ) m t α m ( m + 1 ) 2 ( l I β l ) m .
Case 1: If m is even, t is odd.
β z m , m and L a ( 0 ) are square elements of F q * . Let v = l I , l z ( β z m β l m ) , as the calculation in Theorem 1, v = g r + 1 2 ( t 1 ) ( A + ( t 2 ) z ) m 2 + i ( r + 1 ) . So we only need to consider the parity of ( A + ( t 2 ) z ) m 2 .
  • i 1 , i 2 , , i t are even number, so A + ( t 2 ) z 0 ( mod 2 ) , v is a square element of F q * .
  • i 1 , i 2 , , i t are odd number, so A + ( t 2 ) z 0 ( mod 2 ) , v is a square element of F q * .
Case 2: If m and t are even, r 3 ( mod 4 ) , we assume A is an even integer. It follows that r + 1 2 ( t 1 ) ( A + ( t 2 ) z ) m 2 is an even integer.
Case 3: If m is odd, t is even.
  • t 0 ( mod 4 )
    (1)
    If r 1 ( mod 4 ) , all i 1 , i 2 , , i t are odd, and A 0 ( mod 4 ) , then then ( r + 1 ) ( t 1 ) ( A + ( t 2 ) z ) m 0 ( mod 4 ) , v is a square element of F q * .
    (2)
    If r 3 ( mod 4 ) , all i 1 , i 2 , , i t are even, and A 2 ( mod 4 ) , then ( r + 1 ) ( t 1 ) ( A + ( t 2 ) z ) m 0 ( mod 4 ) , v is a square element of F q * .
  • t 2 ( mod 4 ) .
    (1)
    If r 1 ( mod 4 ) , A 2 ( mod 4 ) , then ( r + 1 ) ( t 1 ) ( A + ( t 2 ) z ) m 0 ( mod 4 ) , v is square of F q * .
    (2)
    If r 3 ( mod 4 ) , A 0 ( mod 4 ) , then ( r + 1 ) ( t 1 ) ( A + ( t 2 ) z ) m 0 ( mod 4 ) , v is square of F q * .
 □
We can extend the Theorem 1 to a more general case.
Theorem 5.
Let q = r 2 , where r is an odd prime power. Suppose that m ( q 1 ) , q 1 m is even, s m , s r 1 and r 1 s is even. If 1 t s ( r + 1 ) g c d ( s ( r + 1 ) , m ) , then there exists a q-ary [ n = t m , n 2 ] MDS self-dual code over F q .
Proof. 
Let α be a primitive m-th root of unity and H = < β > is the cyclic group of order s ( r + 1 ) . By the theorem of group homomorphism,
( H × α ) / α H / ( H α ) ,
Let i 1 , i 2 , , i t be t distinct elements, such that 0 i 1 < i 2 < < i t < 2 ( r + 1 ) . Denote I = { i 1 , i 2 , , i t } , A = i 1 + i 2 + + i t and B = { β i 1 , β i 2 , , β i t } be a set of coset representatives of H × α . Let
a = ( α β i 1 , , α m β i 1 , α β i 2 , , α m β i 2 , , α β i t , , α m β i t ) .
Similar with Theorem 1, we get
L a ( β z α k ) = 1 j m , j k ( β z α k β z α j ) l I , l z j = 1 m ( β z α k β l α j ) = β z ( m 1 ) · m · α k l I , l z ( β z m β l m . )
Since β s ( r + 1 ) = 1 , then β r + 1 = ξ s , where ξ s is s-th primitive root of unity. So β r = ξ s β 1 . Let v = l I , l z ( β z m β l m ) . Since s m , then
v r = l I , l z ( ( β 1 ) z m ( β 1 ) l m ) = l I , l z β ( l + z ) m ( β l m β z m ) = ( 1 ) t 1 β ( A + ( t 2 ) z ) m v .
So v r 1 = ( 1 ) t 1 β ( A + ( t 2 ) z ) m .
Let g be a generator of F q * . It follows that β = g r 1 s and 1 = g r 2 1 2 . So
v = g ( r + 1 ) 2 ( t 1 ) [ A + ( t 2 ) z ] m s .
Case 1: If m odd and t even, we can take λ = g ( r + 1 ) 2 ( t 1 ) A · m s . Hence, we have λ L a ( β z α k ) is square element of F q * .
Case 2: If m even and 2 m s , we can take λ = g ( r + 1 ) 2 ( t 1 ) . Hence, we have λ L a ( β z α k ) is square element of F q * .
So there exists a q-ary MDS self-dual code with length n .  □

4. Conclusions

In this paper, based on the method from [9], we construct several classes of MDS self-dual code over finite fields with odd characteristics via the generalized Reed-Solomon code and extend the generalized Reed-Solomon code.

Author Contributions

Original ideas, writing, original draft preparation, A.Z.; review, Z.J.; funding acquisition, A.Z.

Funding

This research was funded by the National Natural Science Foundation of China under Grants 11401468.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Dau, S.H.; Song, W.; Yuen, C. On the existence of MDS codes over small fields with constrained generator matrices. In Proceedings of the 2014 IEEE International Symposium on Information Theory, Honolulu, HI, USA, 29 June–4 July 2014; pp. 1787–1791. [Google Scholar]
  2. Pedersen, J.P.; Dahl, C. Classification of pseudo-cyclic MDS codes. IEEE Trans. Inf. Theory 1991, 37, 365–370. [Google Scholar] [CrossRef]
  3. Dau, S.H.; Song, W.; Dong, Z.; Yuen, C. Balanced sparsest generator matrices for MDS codes. In Proceedings of the 2013 IEEE International Symposium on Information Theory, Istanbul, Turkey, 7–12 July 2013; pp. 1889–1893. [Google Scholar]
  4. Roth, R.M.; Lempel, A. A construction of non-Reed-Solomon type MDS codes. IEEE Trans. Inf. Theory 1989, 35, 655–657. [Google Scholar] [CrossRef]
  5. Chen, B.; Liu, H. New constructions of MDS codes with complementary dual. IEEE Trans. Inf. Theory 2018, 64, 5776–5782. [Google Scholar] [CrossRef]
  6. Carlet, C.; Mesnager, S.; Tang, C.; Qi, Y. Euclidean and hermitian LCD MDS codes. Des. Codes Cryptogr. 2018, 86, 2605–2618. [Google Scholar] [CrossRef]
  7. Blaum, M.; Roth, R.M. On lowest density MDS codes. IEEE Trans. Inf. Theory 1999, 45, 46–59. [Google Scholar] [CrossRef] [Green Version]
  8. Grassl, M.; Gulliver, T.A. On self-dual MDS codes. In Proceedings of the 2008 IEEE International Symposium on Information Theory, Toronto, ON, Canada, 6–11 July 2008; pp. 1954–1957. [Google Scholar]
  9. Fang, X.; Labad, K.; Liu, H.; Luo, J. New parameters on MDS self-dual codes over finite fields. arXiv, 2018; arXiv:1811.02802vl. [Google Scholar]
  10. Guenda, K. New MDS self-dual codes over finite fields. Des. Codes Cryptogr. 2012, 62, 31–42. [Google Scholar] [CrossRef]
  11. Jin, L.; Xing, C. New MDS self-dual codes from generalized Reed-Solomon codes. IEEE Trans. Inf. Theory 2017, 63, 1434–1438. [Google Scholar] [CrossRef]
  12. Labad, K.; Liu, H.; Luo, J. Construction of MDS self-dual codes over finite fields. arXiv, 2018; arXiv:1807.10625vl. [Google Scholar]
  13. Kim, J.L.; Lee, Y. Euclidean and Hermitian self-dual MDS codes over large finite fields. J. Comb. Theory Ser. A 2006, 105, 79–95. [Google Scholar] [CrossRef]
  14. Yan, H. A Note on the Construction of MDS Self-Dual Codes. Cryptogr. Commun. 2018. [Google Scholar] [CrossRef]
  15. MacWilliams, F.J.; Sloane, N.J.A. The Theory of Error-Correcting Codes; Elsevier: Amsterdam, The Netherland, 1977. [Google Scholar]

Share and Cite

MDPI and ACS Style

Zhang, A.; Ji, Z. New Construction of Maximum Distance Separable (MDS) Self-Dual Codes over Finite Fields. Entropy 2019, 21, 101. https://doi.org/10.3390/e21020101

AMA Style

Zhang A, Ji Z. New Construction of Maximum Distance Separable (MDS) Self-Dual Codes over Finite Fields. Entropy. 2019; 21(2):101. https://doi.org/10.3390/e21020101

Chicago/Turabian Style

Zhang, Aixian, and Zhe Ji. 2019. "New Construction of Maximum Distance Separable (MDS) Self-Dual Codes over Finite Fields" Entropy 21, no. 2: 101. https://doi.org/10.3390/e21020101

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