Mixed codes, which are error-correcting codes in the Cartesian product of different-sized spaces, model degrading storage systems well. While such codes have previously been studied for their algebraic properties (e.g., existence of perfect codes) or in the case of unbounded alphabet sizes, we focus on the case of finite alphabets, and generalize the Gilbert-Varshamov, sphere-packing, Elias-Bassalygo, and first linear programming bounds to that setting. In the latter case, our proof is also the first for the non-symmetric mono-alphabetic $q$-ary case using Navon and Samorodnitsky's Fourier-analytic approach.
翻译:混合代码是不同规模空间的笛卡尔产品中的错误更正代码,其模型储存系统非常优劣,虽然这些代码以前曾因其代数特性(如存在完美代码)或未限制字母大小而进行过研究,但我们侧重于限定字母,并推广Gilbert-Varshamov、球体包装、Elias-Bassalygo和第一个线性编程与该环境的界限。 在后一种情况下,我们的证据也是使用Navon和Samorodnitsky的四重分析方法的非对称性单体-负重美元案例的第一个。