Binary extremal self-dual codes of length $60$ and related codes
Masaaki Harada

TL;DR
This paper classifies specific self-dual codes of length 60, constructs new extremal codes with previously unknown weight enumerators, and explores restrictions on weight enumerators of related codes.
Contribution
It provides a classification of four-circulant singly even self-dual codes of length 60 and constructs new extremal codes with novel weight enumerators.
Findings
New extremal singly even self-dual [60,30,12] codes constructed.
Existence of extremal codes with previously unknown weight enumerators.
Restrictions on weight enumerators of certain self-dual codes with shadow of minimum weight 1.
Abstract
We give a classification of four-circulant singly even self-dual codes for and . These codes are used to construct extremal singly even self-dual codes with weight enumerator for which no extremal singly even self-dual code was previously known to exist. From extremal singly even self-dual codes, we also construct extremal singly even self-dual codes with weight enumerator for which no extremal singly even self-dual code was previously known to exist. Finally, we give some restriction on the possible weight enumerators of certain singly even self-dual codes with shadow of minimum weight .
| Code | |||
| 0 | |||
| 0 | |||
| 0 | |||
| 0 | |||
| 0 | |||
| 0 | |||
| 0 | |||
| 0 | |||
| 10 | |||
| 10 | |||
| 10 | |||
| 10 | |||
| 10 | |||
| Code | ||
|---|---|---|
| Code | Code | Code | |||
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Binary extremal self-dual codes of length and related codes
Masaaki Harada
Research Center for Pure and Applied Mathematics, Graduate School of Information Sciences, Tohoku University, Sendai 980–8579, Japan. email: [email protected].
Abstract
We give a classification of four-circulant singly even self-dual codes for and . These codes are used to construct extremal singly even self-dual codes with weight enumerator for which no extremal singly even self-dual code was previously known to exist. From extremal singly even self-dual codes, we also construct extremal singly even self-dual codes with weight enumerator for which no extremal singly even self-dual code was previously known to exist. Finally, we give some restriction on the possible weight enumerators of certain singly even self-dual codes with shadow of minimum weight .
Keywords: extremal self-dual code; weight enumerator; neighbor.
MSC 2010 Codes: 94B05
1 Introduction
Let be a (binary) singly even self-dual code. All codes in this note are binary. Let denote the subcode of consisting of codewords having weight . The shadow of is defined to be . Shadows for self-dual codes were introduced by Conway and Sloane [3] in order to derive new upper bounds for the minimum weight of singly even self-dual codes, and to provide restrictions on the weight enumerators of singly even self-dual codes. By considering shadows, the largest possible minimum weights of singly even self-dual codes of lengths up to are given in [3, Table I]. In this note, we say that a singly even self-dual code with the largest possible minimum weight given in [3, Table I] is extremal.
The possible weight enumerators of extremal singly even self-dual codes are given in [3] for lengths up to and length (see also [7] for length ). It is a fundamental problem to find which weight enumerators actually occur for the possible weight enumerators (see [3]). The possible weight enumerators of extremal singly even self-dual codes are known as follows:
[TABLE]
where is an integer. If there is an extremal singly even self-dual code with weight enumerator , then [8]. For and , an extremal singly even self-dual code with weight enumerator was found in [12], [2], [13], [5] and [7], respectively. An extremal singly even self-dual code with weight enumerator was found in [3].
One of the main aims of this note is to show the following:
Proposition 1**.**
There is an extremal singly even self-dual code with weight enumerator for .
These codes are constructed from four-circulant singly even self-dual codes for and by considering self-dual neighbors. It remains to determine whether there is an extremal singly even self-dual code with weight enumerator for .
The possible weight enumerators of extremal singly even self-dual codes are known as follows:
[TABLE]
where are integers [3]. If there is an extremal singly even self-dual code with weight enumerator , then [8]. An extremal singly even self-dual code with weight enumerator is known for [11]. An extremal singly even self-dual code with weight enumerator is known for
[TABLE]
The following proposition is one of the main results of this note.
Proposition 2**.**
There is an extremal singly even self-dual code with weight enumerator for
[TABLE]
These codes are constructed from extremal singly even self-dual codes constructed in this note by subtracting and their self-dual neighbors. Finally, we give some restriction on the possible weight enumerators of certain singly even self-dual codes with shadow of minimum weight (Proposition 5). As a consequence, it is shown that for the possible weight enumerator (Corollary 6). All self-dual codes in this note are singly even. From now on, we omit the term singly even.
All computer calculations in this note were done with the help of Magma [1].
2 Extremal four-circulant
self-dual codes
An circulant matrix has the following form:
[TABLE]
so that each successive row is a cyclic shift of the previous one. Let and be circulant matrices. Let be a code with generator matrix of the following form:
[TABLE]
where denotes the identity matrix of order and denotes the transpose of . It is easy to see that is self-dual if . The codes with generator matrices of the form (1) are called four-circulant.
In this section, we give a classification of extremal four-circulant self-dual codes. Two codes are equivalent if one can be obtained from the other by a permutation of coordinates. Our exhaustive search found all distinct extremal four-circulant self-dual codes, which must be checked further for equivalence to complete the classification. This was done by considering all pairs of circulant matrices and satisfying the condition that , the sum of the weights of the first rows of and is congruent to and the sum of the weights is greater than or equal to . Since a cyclic shift of the first rows gives an equivalent code, we may assume without loss of generality that the last entry of the first row of is . Then our computer search shows that the above distinct extremal four-circulant self-dual codes are divided into inequivalent codes.
Proposition 3**.**
Up to equivalence, there are extremal four-circulant self-dual codes.
We denote the codes by . For the codes , the first rows (resp. ) of the circulant matrices (resp. ) in generator matrices (1) are listed in Table 1. We verified that the codes have weight enumerator , where are also listed in Table 1.
3 Extremal self-dual neighbors
Two self-dual codes and of length are said to be neighbors if . Any self-dual code of length can be reached from any other by taking successive neighbors (see [3]). It is known that a self-dual code of length has self-dual neighbors. These neighbors are constructed by finding subcodes of codimension in containing the all-one vector. A computer program written in Magma, which was used to find self-dual neighbors, can be obtained electronically from http://www.math.is.tohoku.ac.jp/~mharada/Paper/neighbor.txt. In this section, we construct extremal self-dual codes by considering self-dual neighbors.
For , by finding all self-dual neighbors of , we determined the equivalence classes among extremal self-dual neighbors of . Our computer search shows that the code has inequivalent extremal self-dual neighbors, which are equivalent to none of the codes , where are given by
[TABLE]
We denote the extremal self-dual codes by . These codes are constructed as
[TABLE]
where and the support of are listed in Table 2. We verified that the codes have weight enumerator , where in Table 2 indicates the values in the weight enumerator . The code has the following weight enumerator:
[TABLE]
We verified that there is no pair of equivalent codes among the codes and the codes .
We continue the search to find extremal self-dual codes by considering self-dual neighbors. We found all inequivalent extremal self-dual neighbors of , which are equivalent to none of the extremal self-dual codes previously obtained in this note. For the codes , and are listed in Table 3. In the table, indicates the values in the weight enumerator . By continuing this process, we found all inequivalent extremal self-dual neighbors of , which are equivalent to none of the extremal self-dual codes previously obtained in this note. Finally, we verified that there is no extremal self-dual neighbor of , which are equivalent to none of the extremal self-dual codes previously obtained in this note.
4 Extremal four-circulant
self-dual codes and self-dual neighbors
Using an approach similar to that given in Section 2, our exhaustive search found all distinct four-circulant self-dual codes. Then our computer search shows that the distinct four-circulant self-dual codes are divided into inequivalent codes.
Proposition 4**.**
Up to equivalence, there are four-circulant self-dual codes.
We denote the codes by . For the codes , the first rows (resp. ) of the circulant matrices (resp. ) in generator matrices (1) are listed in Table 4. The first rows for the all codes can be obtained from http://www.math.is.tohoku.ac.jp/~mharada/Paper/60-4cir-d10.txt.
In addition, we found extremal self-dual codes by considering self-dual neighbors of . Using a method similar to that given in [4], we completed the classification of extremal self-dual neighbors of . Our computer search shows that there is an extremal self-dual neighbor for and that there is no extremal self-dual neighbor for . The codes are constructed as , where and are listed in Table 5 and indicates the values in the weight enumerator . We verified that there are the following equivalent codes among :
[TABLE]
where means that and are equivalent.
Similar to Section 3, by continuing this process, we completed a classification of extremal self-dual neighbors (resp. , ), which are equivalent to none of the extremal self-dual codes previously obtained in this note, of (resp. , ). Finally, we verified that there is no extremal self-dual neighbor of , which are equivalent to none of the codes in Tables 1, 2, 3 and 5. We remark that there is no pair of equivalent codes among the following codes:
[TABLE]
The codes and (see Tables 2 and 5) establish Proposition 1. The code has the following weight enumerator:
[TABLE]
5 Extremal self-dual codes
An extremal self-dual code gives an extremal self-dual code by subtracting two coordinates. We found all the extremal self-dual codes by subtracting from the inequivalent extremal self-dual codes given in Sections 2, 3 and 4. The only extremal self-dual code gives extremal self-dual codes with weight enumerator for which no extremal self-dual code was previously known to exist. More precisely, the codes by subtracting and have weight enumerator for and , where are listed in Table 6. We verified that there are the following equivalent codes:
[TABLE]
where and are inequivalent.
Similar to Sections 3 and 4, we continue the search to find extremal self-dual codes with weight enumerator for which no extremal self-dual code was previously known to exist, by considering self-dual neighbors of . These codes are constructed as
[TABLE]
where and are listed in Table 7. We verified that the codes have weight enumerator , where in Table 7 indicates the values in the weight enumerator . By continuing this process, we found more extremal self-dual codes with weight enumerator for which no extremal self-dual code was previously known to exist. The results are listed in Table 7. From Tables 6 and 7, we have Proposition 2.
6 Weight enumerator
In this section, we give a remark on the possible weight enumerator . First, we discuss a general case including .
Proposition 5**.**
Let be a self-dual code with shadow of minimum weight . Let and denote the numbers of vectors of weight in and , respectively. Suppose that and . Then .
Proof.
Let be the vector of weight and let be a vector of weight in . Since , has weight . Thus, .
Now let be a codeword of weight in and let be the vector of weight in . Then we have . From the assumption that , the weight of is congruent to by Theorem 5 in [3]. Hence, from the assumption that , has weight . Thus, . The result follows. ∎
For example, Proposition 5 can be applied to the following parameters:
[TABLE]
- •
:
The possible weight enumerator of the shadow of an extremal self-dual code with weight enumerator is as follows [3]:
[TABLE]
By Proposition 5, we have
[TABLE]
Since there is an extremal self-dual code with weight enumerator for [11], we have the following:
Corollary 6**.**
There is an extremal self-dual code with weight enumerator if and only if .
- •
:
The weight enumerator in [6, p. 2039] is the possible weight enumerator of an extremal self-dual code with shadow of minimum weight . By Proposition 5, we have for in [6, p. 2039]. The weight enumerators of such a code and its shadow are as follows:
[TABLE]
respectively. It is still unknown whether there is an extremal self-dual code (with shadow of minimum weight ).
- •
:
The weight enumerator in [6, p. 2041] is the unique weight enumerator for an extremal self-dual code with shadow of minimum weight . The weight enumerators of such a code and its shadow are as follows:
[TABLE]
respectively. It is still unknown whether there is an extremal self-dual code (with shadow of minimum weight ).
Acknowledgment. This work was supported by JSPS KAKENHI Grant Number 15H03633. The author would like to thank the anonymous reviewers for the useful comments.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] W. Bosma, J. Cannon and C. Playoust, The Magma algebra system I: The user language, J. Symbolic Comput. 24 (1997), 235–265.
- 2[2] S. Bouyuklieva, R. Russeva and N. Yankov, On the structure of binary self-dual codes having an automorphism of order a square of an odd prime, IEEE Trans. Inform. Theory 51 (2005), 3678–3686.
- 3[3] J.H. Conway and N.J.A. Sloane, A new upper bound on the minimal distance of self-dual codes, IEEE Trans. Inform. Theory 36 (1990), 1319–1333.
- 4[4] N. Chigira, M. Harada and M. Kitazume, Extremal self-dual codes of length 64 64 64 through neighbors and covering radii, Des. Codes Cryptogr. 42 (2007), 93–101.
- 5[5] R. Dontcheva and M. Harada, Some extremal self-dual codes with an automorphism of order 7, Appl. Algebra Engrg. Comm. Comput. 14 (2003), 75–79.
- 6[6] S.T. Dougherty, T.A. Gulliver and M. Harada, Extremal binary self-dual codes, IEEE Trans. Inform. Theory 43 (1997), 2036–2047.
- 7[7] T.A. Gulliver and M. Harada, Weight enumerators of extremal singly-even [ 60 , 30 , 12 ] 60 30 12 [60,30,12] codes, IEEE Trans. Inform. Theory 42 (1996), 658–659.
- 8[8] M. Harada and A. Munemasa, Some restrictions on weight enumerators of singly even self-dual codes, IEEE Trans. Inform. Theory 52 (2006), 1266–1269.
