Double Hurwitz numbers: polynomiality, topological recursion and intersection theory (Q6047232)

To go straight to the point, the paper under review is concerned, as the title announces, with double Hurwitz numbers associated to branched covers of the projective line. What are they about? Here is a short synopsis before than a more detailed description. Hurwitz numbers count, by definition, the cardinality of the equivalence classes of topological coverings of the projective line with a prescribed ramification profile over some distinguished point, conventionally choosen to be \(\infty\), and simply ramified at the other branch points, specified by the celebrated Riemann-Hurwitz formula. In the last few decades, Hurwitz numbers have attracted lot of interest especially with respect to their beautiful and, at the beginning, very surprising relationship with the topological recursion holding for intersection numbers in the moduli space of stable pointed curves, which are in turn concerned with the celebrated Witten's conjecture, soon turned into a theorem by Kontsevich. A breakthrough in the subject was the impactful ELSV paper \textit{T. Ekedahl} et al. [Invent. Math. 146, No. 2, 297--327 (2001; Zbl 1073.14041)], where the authors compute the (finite) \textit{Hurwitz number} \(h_{\mu_1,\ldots,\mu_n,g}\) of all the topological equivalent classes of ramified coverings \(C\to \mathbb{P}^1\), with simple ramification at all branch points but over \(\infty\), with profile prescribed by \((\mu_1,\ldots, \mu_n)\). The point is now that the paper under review, which is a sequel of a previous one by \textit{N. Do} and \textit{M. Karev} [Proc. Symp. Pure Math. 100, 151--178 (2018; Zbl 1452.14051)], walks a further important step forward to the subject. Instead of sticking to simple ones, it is concerned with double Hurwitz numbers, which enumerate branched covers of the complex projective line with prescribed ramification profile over two points (say \(0\) and \(\infty\)), instead of just one. As in the classical case, all the other ramification points are assumed to be simple. The authors point out that the underlying geometry of the double Hurwitz numbers is not yet very well understood, and further investigations are needed. As a matter of fact the authors don not retain from formulating a bunch of conjectures relating the double Hurwitz numbers to the topological recursion. More precisely, the double Hurwitz numbers \(DH_{g,n}({\mu_1,\ldots, \mu_n})\), according to the author's definition, is the weighted enumeration of connected genus \(g\) branched covers \(f : (\Sigma, p_1,\ldots,p_n)\to (\mathbb{CP}^1\) such that i) \(f^{-1}([\infty])= \mu_1[p_1] +\ldots+ \mu_n[p_n]\); ii) the order of any ramification point above \(0\in\mathbb{CP}^1\) is at most \(d\); iii) all other branch points are simple and occur at prescribed points of \(\mathbb{CP}^1\). Hurwitz numbers can be studied also from the point of view of monodromy problems, as the authors point out. If the purpose of a review were that, as it is, of giving the reader the most precise idea possible of a paper's content, the reviewer was rather lucky, this time. Indeed, the main theorems are already listed in the third section of the very well written and exciting introduction. The first main result states that, like in the classical ELSV case, the numbers \(DH_{g,n}\) have polynomial behavior with respect to the profile parameters \(\mu_i\); the second main result establishes a sort of topological recursion producing some correlation differentials, within the best of the traditions; finally, Theorem 1.7 establishes an ELSV-like formula for the Hurwitz numbers counting, proven in the last section before the reference list. It is not difficult to guess that such a dense sixty six pages long article had to end with a rich reference list, especially because of the huge amount of prerequisites demanded by a subject which evolves at an impressive speed. Before than that, however, between the introduction and the bibliography, the authors insert four sections and three appendices. The sections are devoted respectively to (not to list the titles of the almost twenty subsections) i) preliminaries and topological recursion; ii) Double Hurwitz numbers via the semi-infinite wedge; iii) polynomiality results and, finally, iv) the proof of the ELSV-like formula announced in the introduction. So, ending a paper that, besides giving lot of perspectives and raising lots of interesting questions, is literally breathtaking.