Identifying Complex Hadamard Submatrices of the Fourier Matrices via Primitive Sets

TL;DR

This paper presents a method to identify Hadamard submatrices within Fourier matrices using primitive sets, enabling classification into equivalence classes and aiding in their construction.

Contribution

It introduces a criterion based on primitive sets to determine Hadamard submatrices and develops compatibility graphs for their classification.

Findings

Primitive sets determine Hadamard submatrix existence

Partitioning into equivalence classes simplifies analysis

Compatibility graphs facilitate submatrix construction

Abstract

For a given selection of rows and columns from a Fourier matrix, we give a number of tests for whether the resulting submatrix is Hadamard based on the primitive sets of those rows and columns. In particular, we demonstrate that whether a given selection of rows and columns of a Fourier matrix forms a Hadamard submatrix is exactly determined by whether the primitive sets of those rows and columns are compatible with respect to the size of the Fourier matrix. This allows the partitioning of all submatrices into equivalence classes that will consist entirely of Hadamard or entirely of non-Hadamard submatrices and motivates the creation of compatibility graphs that represent this structure. We conclude with some results that facilitate the construction of these graphs for submatrix sizes 2 and 3.