The ham sandwich theorem revisited (Q1802729): Difference between revisions

The following main instance \(K(n,k)\) of Knaster's conjecture [\textit{B. Knaster}, Problem 4, Colloq. Math. 1, 30 (1947); \textit{S. A. Bogatyj}, Russ. Math. Surv. 41, No. 6, 43-57 (1986); translation from Usp. Mat. Nauk 41, No. 6(252), 37-48 (1986; Zbl 0698.52005)] is a natural extension of the Borsuk-Ulam theorem, i.e. \(K(n,n-1)\). \(K(n,k)\): If \(f: S^{n- 1}\to\mathbb{R}^ n\) is a continuous mapping and \(A=\{a_ 0,\dots,a_{n- k}\}\subset S^{n-1}\) the set of vertices of a regular \((n-k)\)- dimensional simplex, then there exists an isometry \(\sigma\in 0(n)\) such that \(f(\sigma(a_ i))=f(\sigma(a_ j))\) for all \(i\), \(j\). The authors prove that \(K(n,k)\) and \(B(n,k)\) (that is a \textit{R. Rado} [J. Lond. Math. Soc. 21, 291-300 (1947; Zbl 0061.096)] type statement about probability Borel \((k+1)\) measures in \(\mathbb{R}^ n\)) are both consequences of the same topological conjecture \(C(n,k)\): The Whitney sum \(\bigoplus ^ k \xi\) of \(k\) copies of the canonical vector bundle \(\xi=V_{n,k}\times_{Z_{k+1}}\mathbb{R}^ k\) and \(V^*_{n,n- k}=V_{n,k/Z_{k+1}}\) does not admit a nowhere zero continuous cross- section, where \(V_{n,k}\) denotes the Stiefel manifold and \(Z_{k+1}\) the group of all isometries in \(0(k)\) determined as cyclic permutations of vertices of a fixed regular simplex centered at the origin in \(\mathbb{R}^ n\). \(B(n,n-1)\) turns out to be just the ham sandwich theorem. One should also remark that \(C(n,n-1)\) is true. One shows that \(C(n,n-2)\) is true for \(n=3\) and all \(n\geq 5\), what implies that \(K(n,n-2)\) and \(B(n,n-2)\) are both true.