Lagrangian matching invariants for fibred four-manifolds. II. (Q929216)

The paper under review is the second part of the author's work on a geometric construction of Lagrangian matching invariants for a singular Lefschetz fibration 4-manifold. See the review on the first part [Geom. Topol. 11, 759--828 (2007; Zbl 1143.53079)] for a background. The paper starts with section 2 in restating theorem B of part I. Section 3 is to develop a general framework for Lagrangian matching invariants (not just for broken fibrations on 4-manifolds). The setup in section 3.1 for almost complex structures of class \(C^{\infty}\) requires a special Banach space technique in order to apply the implicit function theorem and the Baire category theorem (developed by Floer), the paper omits the issue for the symplectic Morse-Bott fibration case to deal with the singular part. The moduli space of the zero set of sections for the symplectic Morse-Bott fibration is formally defined in section 3.2, this is unlike the usual case due to the nontrivial boundary condition. Various analysis issues for this moduli space are not addressed. The Fredholm and transversal properties of the linearization of the \(J\)-holomorphic section are discussed in section 3.3. For the surface with nonempty boundary, the Sobolev space is a kind of trace classes. It would be better to give a explicit proof for the surjectivity in this setup. In section 3.4, the author compactifies the moduli space by considering (i) sections, (ii) bubbles in regular fibres, (iii) bubbles in singular fibres, (iv) boundary bubbles. Dealing with the codimension one components in the moduli space compactification is problematic, and the author shows that the symplectic Morse-Bott fibration is weakly monotone and the codimension one can be avoided. A slight generalization of weak monotonicity to fibred \(k\)-negativity is given in section 3.5. Orientations in section 3.6 is an exposition from some previous book manuscripts. With all these preparations in previous subsections, the definition of the Lagrangian matching invariant is given in section 3.6.1. The Floer theory and Lagrangian matching invariant can be described as an extension to an open-closed topological field theory coupled with singular fibrations. Relative invariants as morphisms are also given in subsection 3.8. The quantum cap product and reduction to homology of the Floer homology are studied in subsection 3.9. Section 4 is to apply section 3 techniques to the broken fibrations on 4-manifolds. Define the admissibility to have the 3-negative for the symmetric product of the regular fibre and the 2-negative for fibres in the Lagrangian matching condition. Thus the Lagrangian matching invariant is given in definition 4.7, and the index computation in theorem D, as well as the Lagrangian matching invariant as an morphism from the Floer homology of admissible pairs in Theorem 5.1. A few examples, trivial bundle over a disc, trivial \(S^2\)-bundle over \(S^2\), broken fibration on \((S^1\times S^3)\# (S^2 \times S^2)\), connected sum with \(\mathbb{C}\mathbb{P}^2\) and broken fibration on \((S^1\times S^3)\# (S^2 \times S^2)\# \overline{\mathbb{C}\mathbb{P}^2}\) are given in section 6, the vanishing matching invariant of connected sum of broken fibrations (Theorem F) is proved in subsection 6.2.1. Other interesting discussions are given in the end of subsection 6.3.1. Again it would be better to have some examples on broken fibrations for \(S^4\) and \(\mathbb{C}\mathbb{P}^2\) which there are nontrivial Lagrangian matching invariants.