Some mapping theorems for continuous functions defined on the sphere (Q412635)

Let \(f: S^{k-1}\to \mathbb R^m\) be a continuous mapping from the sphere \(S^{k-1}\) to the Euclidean space \(\mathbb R^m\), and let \(u_1, \dots, u_n\) be \(n\) distinct points in \(S^{k-1}\) such that \(u_{i+\alpha}\cdot u_{j+\alpha} = u_{i}\cdot u_{j}\) for all possible \(i,j\), where \(\alpha| n\). The main result in this paper are some sufficient conditions, related to the integers \(k, m, \alpha\), and the rank of the \(n\)-vectors, to guarantee the existence of a rotation \(r\) in \(SO(k)\) such that \(f(r u_i) = f(r u_{i+q\alpha})\) for all possible \(i\) and \(q\). The topic concerned here originates from a conjecture posed by Knaster: Given a continuous mapping \(f: S^{m+n-2}\to \mathbb R^{m}\) and \(n\) distinct points \(u_1, \dots, u_n\) in \(S^{m+n-2}\), does there exist a rotation \(r\in SO(m+n-1)\) such that \(f(ru_1) = \cdots =f(ru_n)\)? This is a kind of generalization of the classical Borsuk-Ulam theorem, where \(n=2\) and \(u_1\cdot u_2=-1\).