Heegaard genus formula for Haken manifolds (Q2498423)

Heegaard splittings have long been used in the study of 3-manifolds. The Heegaard genus of a compact 3-manifold has proven to capture the topology of the 3-manifold more accurately than many other invariants. In particular, it provides an upper bound for the rank of the fundamental group of the 3-manifold, and this upper bound need not be sharp, as seen in the examples provided by \textit{M. Boileau} and \textit{H. Zieschang} [Invent. Math. 76, 455--468 (1984; Zbl 0538.57004)]. In this paper the author proves the following: Let \(M\) be a compact orientable 3-manifold and \(Q\subset M\) an orientable boundary incompressible essential surface. Denote the completions of the components of \(M-Q\) with respect to the path metric by \(M^1,\dots, M^k\). Denote the smallest possible genus of a Heegaard splitting of \(M\), or \(M^j\) respectively, for which \(\partial M\), or \(\partial M^j\) respectively, is contained in one compression body, by \(g(M,\partial M)\) or \(g(M^j,\partial M^j)\), respectively. Here \(g(M, \partial M)\) is called the relative genus of \(M\). Denote the maximal number of non parallel essential annuli that can be simultaneously embedded in \(M^j\) by \(n_j\). Then \[ g(M,\partial M)\geq\frac 15\left(\sum_jg(M^j,\partial M^j)-|M-Q| +5-2\chi(\partial-V)+4\chi(Q)-4\sum_jn_j\right). \] A stronger inequality is obtained in the case in which \(M\) and the manifolds \(M^j\) are acylindrical. The author remarks that this formula provides a topological analogue to the algebraic formula \(\text{rank}\,G\geq\frac 13 (\text{rank}\,A+\text{rank}\,B-2 \text{rank}\,C+5)\) provided by \textit{R. Weidmann} [J. Reine. Angew. Math. 540, 77--86 (2001; Zbl 0987.20010)] for the rank of a proper amalgamated product \(G=A*_CB\) with malnormal amalgam \(C\). If \(M\) is a 3-manifold containing a separating incompressible surface \(Q\) and \(Q\) cuts \(M\) into two acylindrical 3-manifolds \(M^1\sqcup M^2\), then Weidmann's formula tells us that the fundamental groups \(\pi_1(M)\), \(\pi_1(M^1)\), \(\pi_1 (M^2)\), \(\pi_1(Q)\) satisfy the following inequality: \(r(\pi_1(M))\geq\frac 13(r(\pi_1(M^1))+r(\pi_1(M^2))-2r(\pi_1(Q))+ 5)\). The correspondence between the two results makes the formula derived here particularly interesting, since it shows that the rank of the fundamental group and the genus of a 3-manifold satisfy a similar linear inequality. The construction and techniques used by the author are a generalization of those in her joint work with \textit{M. Scharlemann} [Math. Ann. 317, No. 4, 783--820 (2000; Zbl 0953.57002)]. The author mentions a similar formula from the book of \textit{K. Johannson} [Topology and Combinatorics of 3-manifolds. Lecture Notes in Mathematics. 1599. Springer-Verlag, Berlin Heidelberg (1995; Zbl 0820.57001), (Proposition 23, 40)]: \(g(M)\geq\frac 15(\sum_jg(M^j)-2+\chi(Q))\), for \(M\) closed and \(M^j\) acylindrical. But the author remarks that while her formula applies in greater generality, the Johannson formula does not apply to the interesting case of a surface bundle over the circle.