The topological Tverberg theorem and winding numbers (Q2568970): Difference between revisions

Let \(\Delta_N^{\leq k}\) be the \(k\)-skeleton of the \(N\)-dimensional simplex \(\Delta_N\). For \(d\geq 0, q\geq 1,\) and \(k\geq 0\), a set \(S\) of \(q\) disjoint faces \(\sigma\) of \(\Delta_{(d+1)(q-1)}^{\leq k}\) is called a Tverberg partition for a map \(f:\Delta_{(d+1)(q-1)}^{\leq k} \to \mathbb R^d\) if \(\bigcap_{\sigma \in S} f(\sigma)\neq \emptyset\). The Tverberg theorem [\textit{H. Tverberg}, J. Lond. Math. Soc. 41, 123--128 (1966; Zbl 0131.20002)] asserts that every linear map \(f:\Delta_{(d+1)(q-1)} \to \mathbb R^d\) has a Tverberg partition. Its topological version is that when linear is replaced by continuous. The topological version has been proven for the case when \(q\) is a prime power. The main result of the paper under review is to show that proving the topological Tverberg conjecture can be reduced to proving the so-called winding number conjecture, which asserts that for \(k=d-1\), there exist \(q\) disjoint faces \(\sigma_1,\dots,\sigma_q\) of \(\Delta_{(d+1)(q-1)}^{\leq d}\) and a point \(p\in \mathbb R^d\) such that for each \(i\), either (i) \(\dim (\sigma_i)\leq d-1\) and \(p\in f(\sigma_i)\) or (ii) \(\dim (\sigma_i)=d\), and either (a) \(p\in f(\partial \sigma_i)\) or (b) \(p\notin f(\partial \sigma_i)\) and \(W(f| _{\partial \sigma_i},p)\neq 0\) where \(W(g,x)\) denotes the winding number of a map \(g\) with respect to a point \(x\).