Determination of generalized horseshoe maps inducing all link types (Q1877614)

Let \(\varphi\) be an orientation-preserving automorphism of the unit \(2\)-disk \(D\) in \(\mathbb{R}^2\), and \(\Phi=\{\varphi_t\}_{0\leq t\leq 1}\) an isotopy of \(D\) such that \(\varphi_0=\text{id}_D\), \(\varphi_1=\varphi\). Each periodic orbit \(P\) of \(\varphi\) yields a braid \(\beta_{\Phi}P:=\bigcup_{0\leq t \leq 1}(\varphi_t(P)\times\{t\})\) lying in the solid cylinder \(C_{\Phi}:=\bigcup_{0\leq t \leq 1}(\varphi_t(D)\times\{t\})\subset\mathbb{R}^3\) specified by \(\Phi\). The isotopy \(\Phi\) uniquely determines the link \(S_\Phi P\) in \(\mathbb{R}^3\), the closure of \(\beta_\Phi P\), lying in the solid torus obtained by gluing the top and the bottom of \(C_\Phi\) naturally without twists. One says that \(\varphi\) induces all link types if there exists \(k\in\mathbb{Z}\) such that for any link \(L\subset \mathbb{R}^3\), \(L\in\{h^k(S_\Phi P)\;| \;P\;\text{is a periodic orbit of}\;\varphi\}\), where the homeomorphism \(h\) twists the solid torus once along the meridian. Let \(N(\varphi):=\#\{k\in\mathbb{Z}\;| \;k \;\text{satisfies the previous condition}\}\). The authors define the concept of generalized horseshoe map \(G\) and its twist signature \(t(G)\), and construct the template \(\mathcal{U}(G)\) and the braid \(b(G)\). They prove that \(G\) is positive (resp. negative) if and only if \(b(G)\) is positive (resp. negative), and the template \(\mathcal{U}(G)\) is universal if and only if \(G\) is neither positive nor negative. They completely determine the number \(N(G)\) and, as a direct corollary, when such maps \(G\) induce all link types (for the study of the problem concerning which automorphism induces all link types, see also the second author [J. Knot Theory Ramifications 9, 771--795 (2000; Zbl 0995.37028) and ibid. 9, 939--953 (2000; Zbl 0995.37029)]). Computable necessary and sufficient conditions for a pleated vector field to support a universal template are given [see also \textit{R. Ghrist} and \textit{T. Young}, Nonlinearity 11, 1111--1125 (1998; Zbl 0901.57005)]. One should note that the authors' definition of the twist signature is different from that which has been given by Ghrist and Young.