Selectively balancing unit vectors (Q1747993)

A set \(U\) of unit vectors is called selectively balancing if one can select two disjoint subsets \(U^{+}\) and \(U^{-}\), not both empty, such that Euclidean distance between the sum of \(U^{+}\) and the sum of \(U^{-}\) is less than 1. In this paper, the authors try to balance unit vectors in the \(n\)-dimensional Euclidean space. Let \(\sigma(n)\) denote the minimum number such that any \(\sigma(n)\) unit vectors in \(\mathbb{R}^n\) are selectively balancing. It is proved that \(\sigma(n)\) is asymptotically \(\frac12 n\log n\). The results can be applied to the dot product representation of cube graphs.