On the number of Tverberg partitions in the prime power case (Q854837): Difference between revisions

\textit{H. Tverberg} showed in [J. Lond. Math. Soc. 41, 123--128 (1966; Zbl 0131.20002)] that any set of \((d+1)(q-1)+1\) points in \(\mathbb R^d\) admits a partition into \(q\) subsets such that the intersection of their convex hulls is nonempty. The result is the best possible; for fewer than this number of points the statemtne does not hold. Placed in a topological setting this becomes: If \(q\geq 2\), \(d\geq 1\), and \(N:=(d+1)(q-1)\), for every affine map \(f: \|\sigma^N\|\to \mathbb R^d\) there are \(q\) disjoint faces \(F_1,\dots,F_q\) of the standard \(N\)-simplex \(\sigma^N\) whose images under \(f\) intersect, \(\bigcap^q_{i=1}f(\|F_i\|)\neq \emptyset\). For \(q\) a prime and \(f\) a continuous map, this was proved by \textit{I. Bárány} et al. [J. Lond. Math. Soc. II. Ser. 23, 158--164 (1981; Zbl 0453.55003)]. The conjecture that there are at least \([(q-1)!]^d\) Tverberg partitions for any set of \((d+1)(q-1)+1\) points in \(\mathbb R^d\) is still unproved. For \(q\) a prime a non-trivial lower bound is known; here this result is extended to the prime power case.