Topological recursions of Eynard-Orantin type for intersection numbers on moduli spaces of curves (Q369667)

The topological recursion is a framework which associates certain invariants to a spectral curve via a recursive definition. These invariants have very interesting properties and appear in a variety of contexts in mathematics and physics. Many known and well studied invariants can be recovered via certain specializations; the topological recursion gives a broader pictures which connects different areas of mathematics. The importance of the topic is well recognized in the literature, and techniques based on the topological recursion are widely applied, starting from the seminal works by Eynard and Orantin. This paper concerns the study of the intersection numbers on the Deligne-Mumford moduli spaces of curves, and in particular the properties of a generating function of higher Weil-Petersson volumes. Higher Weil-Petersson volumes are a generalization of the ordinary Weil-Petersson volumes, which are the hyperbolic volumes of the moduli spaces of bordered Riemann surfaces with fixed genus \(g\) and \(n\) geodesic boundaries of fixed length. In this ordinary Weil-Petersson case the existence of a recursion formula has been established by \textit{M. Mirzakhani} [J. Am. Math. Soc. 20, No. 1, 1--23 (2007; Zbl 1120.32008)]. This result has been generalized to higher Weil-Petersson volumes by \textit{K. Liu} and \textit{H. Xu} [Int. Math. Res. Not. 2009, No. 5, 835--859 (2009; Zbl 1186.14059)]. These invariants also obey the Virasoro constraints, in the sense that an appropriate generating function is annihilated by operators which obey a Virasoro algebra. The main result of the paper is the proof that such a Virasoro constraint is equivalent to the Eynard-Orantin topological recursion for a certain spectral curve. A different proof using matrix models was provided earlier by \textit{B. Eynard} [Ann. Henri Poincaré 12, No. 8, 1431--1447 (2011; Zbl 1245.14013)]. The approach taken in the paper is more geometric. The paper is very technical, probably not suited for someone not familiar with the subject. On the other hand the proofs of the results are explained carefully and are easy to follow; the author has made an effort to make a rather technical result readable.