Core Concepts
Nearly perfect covering codes are a class of codes that meet an improved sphere-covering bound, providing efficient coverage of the entire space with a small number of codewords.
Abstract
The content discusses the structure and properties of nearly perfect covering codes, which are a class of codes that meet an improved sphere-covering bound known as the van Wee bound. These codes have a covering radius of 1 and contain 2^(2r-r) codewords, where r is the length of the code.
The key insights and highlights are:
The codewords in a nearly perfect covering code can be partitioned into pairs, where the distance between the codewords in each pair is either 1 or 2, and no other codeword is within distance 2 of either codeword in the pair.
Based on this property, the nearly perfect covering codes can be classified into three families: type A, where the distance between the codewords in each pair is 1; type B, where the distance between the codewords in each pair is 2; and type C, which is a mix of the two.
Constructions for codes in each of the three families are presented, using perfect codes and their properties.
The weight distribution of the codes in the type A and type B families is shown to be unique, while the type C family can have varying weight distributions depending on the intersection of the underlying perfect codes.
A special class of balanced type A codes is discussed, where the codewords differ by exactly one coordinate.
Stats
The number of codewords in a (2r, 1)-nearly perfect covering code is 2^(2r-r).
The number of words in F_2^(2r) that are covered twice by the code is 2^(2r-r-1).
Quotes
"Nearly perfect packing codes are those codes that meet the Johnson upper bound on the size of error-correcting codes. This bound is an improvement to the sphere-packing bound. A related bound for covering codes is known as the van Wee bound. Codes that meet this bound will be called nearly perfect covering codes."
"If x is a word of length n, such that x ∉ C, then B_1(x) contains exactly one word that is covered by two codewords of C and no word that is covered by more than two codewords of C."