Instanton Floer homology for lens spaces (Q1936625): Difference between revisions

To compute the Donaldson invariants of a \(4\)-manifold split along a general \(3\)-manifold, one would need a Floer theory that could address two problems. One problem is that non-trivial reducible flat connections that exist on many three manifolds, including lens spaces induce singularities in the space of ASD conections which keep the naive definition of the boundary map from having square zero. The other problem is that some of the homology classes in the \(4\)-manifold will cut through the \(3\)-manifold requiring one to understand a Floer theory for a knot in a \(3\)-manifold. There is old work studying the Floer theories for lens spaces, [\textit{D. M. Austin}, J. Differ. Geom. 32, No. 2, 383--413 (1990; Zbl 0731.58012)]; \textit{D. M. Austin} and \textit{P. J. Braam} [Topology 35, No. 1, 167--200 (1996; Zbl 0901.57027)]; \textit{P. J. Braam} and \textit{S. K. Donaldson} [The Floer memorial volume. Basel: Birkhäuser. Prog. Math. 133, 257--281 (1995; Zbl 0996.57517)]; \textit{K. Fukaya} [Aspects of low dimensional manifolds. Tokyo: Kinokuniya Company Ltd. Adv. Stud. Pure Math. 20, 1--92 (1992; Zbl 0852.57014)]; \textit{M. Furuta} [Lond. Math. Soc. Lect. Note Ser. 150, 161--174 (1990; Zbl 0837.57015)]; \textit{M. Furuta} and \textit{Y. Hashimoto} [J. Fac. Sci., Univ. Tokyo, Sect. I A 37, No. 3, 585--600 (1990; Zbl 0729.58018)], but no prior result addresses both problems at the same time. In this paper, Sasahira defines a Floer theory for knots in lens spaces that addresses both of these issues. He does so by combining the idea of Furuta using twisted Dirac operators to fix the boundary map [Zbl 0837.57015], with the construction of Fukaya of a Floer theory that sees knots in the \(3\)-manifold [Zbl 0852.57014]. Briefly, Furuta's idea is to only count points in the moduli space for which the index of the associated twisted Dirac operator is odd, and Fukaya's idea is to consider \(2\)-dimensional components of the moduli space in addition to the zero dimensional components, and count the zeros of a section of the determinant line bundle restricted to the knot cross the line. These ideas are explained well in this paper. As an application, Sasahira considers how \(\mathbb{C}P^2\#\mathbb{C}P^2\) may be split along lens spaces. It may only be split along \(L(p,q)\) when \(p=a^2+b^2\). It is possible to split it in many ways along many Lens spaces, but it is not possible to split it along \(L(16N+1,2)\) for prime \(16N+1\) as a consequence of this present work.