Combinatorial cobordism maps in hat Heegaard Floer theory (Q955153)

Hat Heegaard Floer theory is the simplest Heegaard Floer theory. To a closed, connected, oriented 3-manifold \(Y\) it associates a vector space \(\widehat{HF}(Y)\) coming from a Heegaard diagram of \(Y\), and to a cobordism \(W\) between 3-manifolds \(Y_1\) and \(Y_2\) it associates a map \(\widehat{F}_W\) between \(\widehat{HF}(Y_1)\) and \(\widehat{HF}(Y_2)\). In this way, hat Heegaard Floer theory is a sort of ``decorated \((3+1)\) TQFT''. The map \(\widehat{F}_W\) can be used to detect exotic smooth structures on 4-manifolds with boundary, and its \(HF^-\) and \(HF^+\) versions conjecturally recover the gauge-theoretic Seiberg-Witten invariant. Both \(\widehat{HF}(Y)\) and \(\widehat{F}_W\) can be filtered by \(\text{Spin}^c\) structures. Our coefficient ring is \(\mathbb{Z}/2\mathbb{Z}\). Calculating Heegaard Floer theory in general involves counting pseudoholomorphic disks in symmetric products of Riemann surfaces, which is a difficult analytic problem. However, \textit{S. Sarkar} and \textit{J. Wang} [An algorithm for computing some Heegaard Floer homologies, preprint, arXiv:math/0607777v3, to appear in Annals of Math.] showed that \(\widehat{HF}(Y)\) may be calculated combinatorially, using the fact that \(Y\) admits a nice Heegaard diagram in which all elementary domains are either rectangles or bigons. This paper shows, under a technical assumption (that \(H_1(Y_1;\mathbb{Z})\) and \(H_1(Y_2;\mathbb{Z})\) surject onto \(H_1(W;\mathbb{Z})\), all modulo torsion), how to also calculate the rank of \(\hat{F}_W\) combinatorially. First, in Section 2, it is proved that in the context of cobordisms, a Heegaard diagram can be chosen in which all fundamental regions are either rectangles, triangles or bigons. The problem then reduces to counting holomorphic triangles combinatorially (see [\textit{S. Sarkar}, Maslov index of holomorphic triangles, preprint, arXiv:math/0609673v2] for a different solution to this problem). In Section 3 this is achieved first for a single two-handle addition, and then for multiple two-handle additions. The paper concludes with the example of \(+1\)-surgery on the right-hand trefoil. An interesting problem highlighted in the introduction is to combinatorially describe the isomorphism between \(\widehat{HF}(Y)\) for different nice Heegaard diagrams. This would enable us to calculate \(\hat{F}_W\) combinatorially (not merely its rank) and to remove the assumption on homologies. Editorial remark: An erratum corrects some minor labeling-errors in the printed version that have been corrected in the online version of the article.