Equivalence of ELSV and Bouchard-Mariño conjectures for \(r\)-spin Hurwitz numbers (Q2339317)

The authors generalize two conjectures (now proved) about Hurwitz numbers, the ELSV formula, which relates them to Hodge integrals, and the Bouchard-Mariño (BM) conjecture, which relates them to the topological recursion of \textit{L. Chekhov} and \textit{B. Eynard} [J. High Energy Phys. 2006, No. 12, 026, 29 p. (2006; Zbl 1226.81138)], \textit{B. Eynard} and \textit{N. Orantin} [Commun. Number Theory Phys. 1, No. 2, 347--452 (2007; Zbl 1161.14026)]. The generalization concerns ``Hurwitz numbers with completed cycles'' that enumerate ramified coverings of the sphere of a certain kind. Okounkov and Pandharipande showed that these numbers express the Gromov-Witten invariants of \(r\)-spin structures over smooth curves. What the authors call the \(r\)-ELSV formula is the conjectural equality of these numbers with Chiodo's \(r\)-spin integrals, an analog of the Hodge integrals. The \(r\)-BM conjecture is that the \(n\)-point functions for completed Hurwitz numbers satisfy the topological recursion with the spectral curve \(x=-y^r+\log y\). The main result is a rigorous proof that the two conjectures are equivalent. The proof uses Frobenius algebra techniques, Givental's \(R\)-matrix, and leaf forms generated by the Bergman kernel, it identifies the Givental-Teleman semi-simple cohomological field theories with the topological recursion. Ample independent evidence for both conjectures is presented. The \(r\)-ELSV formula is known in genus zero, and in genus one with one marked point. It was also successfully tested numerically in other cases. \textit{V. Bouchard} and \textit{M. Mariño} [Proc. Symp. Pure Math. 78, 263--283 (2008; Zbl 1151.14335)] (BM) gave a ``physics proof'' of their conjecture (before it was rigorously derived from the ELSV formula), which carries over to the \(r\)-BM case. Unfortunately, it can not be easily converted into a mathematical proof, although according to the authors ``it will probably convince a mathematical physicist''. One issue seems to be that the potential in the matrix model is neither a rational function, nor independent of a certain parameter, both of which are required for deriving the topological recursion. Another issue is that the infinite contour in the ``physics proof'' has to be replaced by a finite one for the integral to make sense, but then it does not remain invariant under the performed change of variables. Additionally, for the completed Hurwitz numbers the spectral curve has been computed and it coincides with the conjectured one. The quantization of that curve leads to an operator annihilating the principal specialization of the generating function for Hurwitz numbers according to another conjecture \textit{S. Gukov} and \textit{P. Sulkowski} [J. High Energy Phys. 2012, No. 2, Paper No. 070, 57 p. (2012; Zbl 1309.81220)]. Finally, the \(r\)-BM conjecture is in line with the general philosophy that the spectral curve of an enumerative problem should be given by its ``\((0,1)\) geometry'', namely the genus zero \(1\)-pointed generating function in the case of Hurwitz numbers.