The index of a vector field on an orbifold with boundary (Q2390639)

\textit{I. Satake} [Proc. Nat. Acad. Sci. U.S.A. Proc. Natl. Acad. Sci. USA 42, 359--363 (1956; Zbl 0074.18103), J. Math. Soc. Japan 9, 464--492 (1957; Zbl 0080.37403)] proved a Poincaré-Hopf index theorem for a smooth closed orbifold \(Q\) equipped with a vector field \(X\) having isolated zeros, which states that \(\mathfrak{Ind}^{\text{orb}}(X; Q)=\chi_{\text{orb}}(X)\) where the expressions \(\mathfrak{Ind}^{\text{orb}}\) and \(\chi_{\text{orb}}\) are orbifold analogues of the topological index of a vector field and the Euler characteristic respectively. In this paper the authors give a generalization of this theorem for such an orbifold \(Q\) with boundary where the vector field \(X\) is in generic contact with \(\partial Q\). This is indeed also an orbifold generalization of the Poincaré-Hopf index theorem of \textit{C. C. Pugh} [Topology 7, 217--226 (1968; Zbl 0194.24602)] for a vector field on a manifold with boundary. Orbifold analogues of Pugh's submanifolds are constructed in a similar way. Let \(R_-^1\) be the closure of the subset of \(\partial Q\) where \(X\) points out of \(Q\) and \(\Gamma^1\) the subset of \(\partial Q\) where \(X\) is tangent to \(\partial Q\). Then \(R_-^1\) becomes an \((n-1)\)-dimensional orbifold with boundary \(\Gamma^1\) where \(n=\dim Q\). Define \(R_-^2\) to be the closure of the subset of \(\Gamma^1\) where \(X\) points out of \(R_-^1\) and also define \(\Gamma^2\) to be the subset of \(\Gamma^1\) where \(X\) is tangent to \(\Gamma^1\). In the same way one obtains orbifolds \(\Gamma^i\) and \(R_-^i\) for \(n \geq 3\) where \(R_-^i\) are \((n-i)\)-dimensional orbifolds and then \(\Gamma^i\) are their boundaries. Then the main result of this paper is the following formula (Theorem 1.1): \[ \mathfrak{Ind}^{\text{orb}}(X; Q)=\chi_{\text{orb}}(Q, \partial Q) + \sum^n_{i=1}\chi_{\text{orb}}(R_-^i, \Gamma^i). \] The proof proceeds as in the manifold case. The key idea in this approach is to consider the double orbifold \(\hat{Q}\) of \(Q\) in order to apply Satake's formula above. Its essence consists in the construction of a vector field \(\hat{X}\) on \(\hat{Q}\), obtained by perturbing (near the boundary) the vector field which coincides with \(X\) on each copy of \(Q\). The use of this vector field enables one to get \(\mathfrak{Ind}^{\text{orb}}(X; Q) =\chi_{\text{orb}}(Q, \partial Q)+ \mathfrak{Ind}^{\text{orb}}(X; R_{-}^1)\). Since by definition \(\hat{X}\) coincides with \(X\) on \(\Gamma^1\), this formula can be used recursively for all \(R_-^i\). In fact by replacing \(Q\) by \(R_-^i\) one has \(\mathfrak{Ind}^{\text{orb}}(X; R_-^i) =\chi_{\text{orb}}(R_-^i, \Gamma^i) +\mathfrak{Ind}^{\text{orb}}(X; R_{-}^{i+1})\) for each \(i\). Combining these formulas yields \(\mathfrak{Ind}^{\text{orb}}(X; R_{-}^1)= \sum_{i=1}^n\chi_{\text{orb}}(R_{-}^i, \Gamma^i)\) and hence substituting this formula into the first formula one obtains the desired result. This paper concludes with an application of the result now obtained to the vector field induced on the inertia orbifold of \(Q\) (Corollary 5.2).