A method to generate upper bounds for the sums of squares formulae problem
The smallest integer \(t\) for which an identity \[ (x^2_1+ \cdots+ x^2_r) (y^2_1+ \cdots +y^2_s)^2 =z^2_1+ \cdots+ z^2_t \tag{1} \] exists is denoted by \(r*_\mathbb{Z} s\) \((z_i\) bilinear in the \(x_j,y_k\) with integer coefficients). It is known that (1) exists iff an \(r\times s\) consistently signed intercalate matrix of type \([r,s,t]\) exists. Best upper bounds for \(r*_\mathbb{Z} s\) \((1\leq r,s \leq 64)\) have been obtained in the past by juxtaposing two or more matrices of smaller size. The present paper gives irreducible configurations that juxtapose \(k\) matrices for \(k\leq q\). Further, the author gives an algorithm that yields the lowest bound on \(r*_\mathbb{Z} s\) from upper bounds on \(a*_\mathbb{Z} b\) \((a\leq r,b\leq s,a+ b<r+s)\). More new upper bounds on \(r*_\mathbb{Z} s\) \((1\leq r,s \leq 64)\) are obtained by using the above machinery.
- A step towards Yuzvinsky's conjecture
- New constructions for integral sums of squares formulae
- Intercalate coloring of matrices and the Yuzvinsky conjecture
- On lower bounds on the size of sums-of-squares formulas
- Some upper bounds for composition numbers
- scientific article; zbMATH DE number 784889 (Why is no real title available?)
- scientific article; zbMATH DE number 796979 (Why is no real title available?)
- New solutions to the Hurwitz problem on square identities
- On the uniqueness of balanced complex orthogonal design
This page was built for publication: A method to generate upper bounds for the sums of squares formulae problem
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2565259)