arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2201.10057cs.ITmath.IT
收藏

特征为3的域上线性索引编码的最优性

On the Optimality of Linear Index Coding over the Fields with Characteristic Three

Arman Sharififar, Parastoo Sadeghi, Neda Aboutorab

摘要:

已经知道,线性编码在实现一般索引编码问题的最佳速率方面的不足是根源于其速率对字段大小的依赖性。然而,这种依赖性只通过两个著名的拟阵实例来描述,即Fano和非Fano拟阵,这反过来又将其范围限制在特征为2的字段上。在本文中,我们扩展了这一范围,以证明依赖的线性索引编码率字段的特征三。通过构造两个大小为29的索引编码实例,证明了对于第一个实例,线性编码仅在特征为3的域上是最优的,而对于第二个实例,线性编码在特征为3的域上永远不是最优的.然后,第二实例的变型被设计为大小为58的第三索引编码实例。对于这种情况下,它被证明,而线性编码在任何领域的特征三不能是最佳的,存在一个非线性码的领域的特征三,达到其最佳速率。以两种特定方式连接第一和第三索引编码实例,称为无路连接和双向连接,将导致大小为87和91的两个新索引编码实例,对于这两个新索引编码实例,线性编码优于非线性编码。本文的另一个主要贡献是减少了第一和第二索引编码实例的线性编码空间上的关键约束,每个索引编码实例的大小为29,到一个拟阵实例的基集大小为9,其线性表示性取决于特征为3的字段。本文通过这两个相对较小的拟阵实例给出的证明和讨论将揭示导致线性编码不足以解决一般索引编码问题的根本原因。

英文摘要:

It has been known that the insufficiency of linear coding in achieving the optimal rate of the general index coding problem is rooted in its rate's dependency on the field size. However, this dependency has been described only through the two well-known matroid instances, namely the Fano and non-Fano matroids, which, in turn, limits its scope only to the fields with characteristic two. In this paper, we extend this scope to demonstrate the reliance of linear index coding rate on fields with characteristic three. By constructing two index coding instances of size 29, we prove that for the first instance, linear coding is optimal only over the fields with characteristic three, and for the second instance, linear coding over any field with characteristic three can never be optimal. Then, a variation of the second instance is designed as the third index coding instance of size 58. For this instance, it is proved that while linear coding over any field with characteristic three cannot be optimal, there exists a nonlinear code over the fields with characteristic three, which achieves its optimal rate. Connecting the first and third index coding instances in two specific ways, called no-way and two-way connections, will lead to two new index coding instances of size 87 and 91, for which linear coding is outperformed by nonlinear codes. Another main contribution of this paper is the reduction of the key constraints on the space of the linear coding for the first and second index coding instances, each of size 29, into a matroid instance with the ground set of size 9, whose linear representability is dependent on the fields with characteristic three. The proofs and discussions provided in this paper through using these two relatively small matroid instances will shed light on the underlying reason causing the linear coding to become insufficient for the general index coding problem.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑