Supnick matrix

A Supnick matrix is a square Monge array that is symmetric around the main diagonal.

An n-by-n matrix is a Supnick matrix if, for all i, j, k, l such that if

${\displaystyle 1\leq i<k\leq n}$ and ${\displaystyle 1\leq j<l\leq n}$

then

${\displaystyle a_{ij}+a_{kl}\leq a_{il}+a_{kj}\,}$

and also

${\displaystyle a_{ij}=a_{ji}.\,}$

A logically equivalent definition is given by Rudolf & Woeginger who in 1995 proved that

A matrix is a Supnick matrix iff it can be written as the sum of a sum matrix S and a non-negative linear combination of LL-UR block matrices.

The sum matrix is defined in terms of a sequence of n real numbers {α_i}:

${\displaystyle S=[s_{ij}]=[\alpha _{i}+\alpha _{j}];\,}$

and an LL-UR block matrix consists of two symmetrically placed rectangles in the lower-left and upper right corners for which a_ij = 1, with all the rest of the matrix elements equal to zero.

Adding two Supnick matrices together will result in a new Supnick matrix (Deineko and Woeginger 2006).

Multiplying a Supnick matrix by a non-negative real number produces a new Supnick matrix (Deineko and Woeginger 2006).

If the distance matrix in a traveling salesman problem can be written as a Supnick matrix, that particular instance of the problem admits an easy solution (even though the problem is, in general, NP hard).

• Supnick, Fred (July 1957). "Extreme Hamiltonian Lines". Annals of Mathematics. Second Series. 66 (1): 179–201. doi:10.2307/1970124. JSTOR 1970124.
• Woeginger, Gerhard J. (June 2003). "Computational Problems without Computation" (PDF). Nieuwarchief. 5 (4): 140–147.
• Deineko, Vladimir G.; Woeginger, Gerhard J. (October 2006). "Some problems around travelling salesmen, dart boards, and euro-coins" (PDF). Bulletin of the European Association for Theoretical Computer Science. EATCS. 90: 43–52. ISSN 0252-9742.