Some geometric properties of successive difference substitutions (Q415953)

A polynomial in \(n\) variables \(F\in \mathbb{R}\left[ x_{1},\dots,x_{n}\right] \) is said to be positive semi-definite when \(F\left( x_{1},\dots,x_{n}\right) \geq 0\) for all \(x\in \mathbb{R}_{+}^{n}.\) The complexity of deterministic methods for deciding the positivity of polynomials increases rapidly as variable numbers and degrees of the polynomials increase, so that heuristic methods have been proposed. One of these heuristic methods, called successive difference substitution, was proposed by \textit{L. Yang} [in: Proceedings of the 10th Asian Technology Conference in Mathematics. Singapore: National Institute of Education, 37--46 (2005)], and it is based on the computation of a (possibly infinite) sequence of sets associated with \(F\) and the permutations of the columns of the unit matrix. This paper provides a sufficient condition for the mentioned sequence corresponding to a positive semi-definite polynomial \(F\) to be infinite. So, the paper identifies a class of polynomials which cannot be proved to be positive semi-definite by the successive difference substitution method.