Jump to content

Paley construction: Difference between revisions

From Wikipedia, the free encyclopedia
Content deleted Content added
added missing minus sign
Remove mention of Scarpis. His construction appears to be unrelated.
Line 1: Line 1:
In [[mathematics]], the '''Paley construction''' is a method for constructing [[Hadamard matrix|Hadamard matrices]] using [[finite field]]s. The construction was described in [[1933]] by the [[England|English]] [[mathematician]] [[Raymond Paley]], although a version of it was already known to Scarpis.
In [[mathematics]], the '''Paley construction''' is a method for constructing [[Hadamard matrix|Hadamard matrices]] using [[finite field]]s. The construction was described in [[1933]] by the [[England|English]] [[mathematician]] [[Raymond Paley]].


The Paley construction uses quadratic residues in a finite field ''GF''(''q'') where ''q'' is a power of an odd [[prime number]]. There are two versions of the construction depending on whether ''q'' is congruent to 1 or 3 (mod 4).
The Paley construction uses quadratic residues in a finite field ''GF''(''q'') where ''q'' is a power of an odd [[prime number]]. There are two versions of the construction depending on whether ''q'' is congruent to 1 or 3 (mod 4).
Line 74: Line 74:


==References==
==References==
* Scarpis, Umberto, Sui Determinants di Valore Massimo, Rend, della R. Istituto Lombardo di Scienze e Lettere (2), 31 (1898), 1441–1446.
* {{cite journal
* {{cite journal
| last = Paley
| last = Paley

Revision as of 02:12, 14 July 2008

In mathematics, the Paley construction is a method for constructing Hadamard matrices using finite fields. The construction was described in 1933 by the English mathematician Raymond Paley.

The Paley construction uses quadratic residues in a finite field GF(q) where q is a power of an odd prime number. There are two versions of the construction depending on whether q is congruent to 1 or 3 (mod 4).

Quadratic character and Jacobsthal matrix

The quadratic character χ(a) indicates whether the given finite field element a is a perfect square or not. Specifically, χ(0)=0, χ(a) = 1 if a=b2 for some non-zero finite field element b, and χ(a) = −1 if a is not the square of any finite field element. For example, in GF(7) the non-zero squares are 12 = 62 = 1, 22 = 52 = 4, and 32 = 42 = 2. Hence χ(0) = 0, χ(1) = χ(2) = χ(4) = 1, and χ(3) = χ(5) = χ(6) = −1.

The Jacobsthal matrix Q for GF(q) is the q×q matrix with rows and columns indexed by finite field elements such that the entry in row a and column b is χ(ab). For example, in GF(7), if the rows and columns of the Jacobsthal matrix are indexed by the field elements 0, 1, 2, 3, 4, 5, 6, then

The Jacobsthal matrix has the properties QQT = qIJ and QJ = JQ = 0 where I is the q×q identity matrix and J is the q×q all-1 matrix.

Paley construction I

If q is congruent to 3 (mod 4) then

is a Hadamard matrix of size q + 1. Here j is the all-1 column vector of length q−1.

Paley construction II

If q is congruent to 1 (mod 4) then the matrix obtained by replacing all 0 entries in

with the matrix

and all entries ±1 with the matrix

is a Hadamard matrix of size 2(q + 1).

The Hadamard conjecture

The size of a Hadamard matrix must be 1, 2, or a multiple of 4. The tensor product of two Hadamard matrices of sizes m and n is an Hadamard matrix of size mn. By forming tensor products of matrices from the Paley construction and the 2×2 matrix,

Hadamard matrices of every allowed size up to 100 except for 92 are produced. This led Paley to conjecture that Hadamard matrices exist for every size which is a multiple of 4. A matrix of size 92 was eventually constructed by Baumert, Golomb, and Hall, using a construction due to Williamson combined with a computer search. Currently, Hadamard matrices have been shown to exist for all for m < 668.

References

  • Paley, R.E.A.C. (1933). "On orthogonal matrices". J. Math. Phys. 12: 311–320.
  • L. D. Baumert, S. W. Golomb, M. Hall Jr., Discovery of an Hadamard matrix of order 92, Bull. Amer. Math. Soc. 68 (1962) 237–238.