Quantum indices and refined enumeration of real plane curves (Q1699388)

This is an outstanding work which suggest a series of real enumerative invariants of new type and provides a new, deep and unexpected explanation of the geometry behind the refined tropical enumerative invariants. The enumerative problem stated and solved in the paper is as follows: count real rational curves in the plane (or another real toric surface) having a given degree and passing through a fixed configuration of points on the coordinate axes such that the fixed points have either real or purely imaginary coordinates. The key observation is that, in such a case, the signed area of the complex amoeba (the image under the map \((z,w)\mapsto(\log|z|,\log|w|)\)) of a chosen component of the set of non-real points of the curve equals \(\frac{\pi^2}{2}\) times an integer, called the quantum index of the curve with a chosen component of the set of non-real points (or, equivalently, with a given orientation of the real point set). The first main result of the paper is that the number of the real plane curves of a given degree and of a given quantum index, and obeying point constraints as above, does not depend on the choice of the constraint when counting curves with the Welschinger sign (see [\textit{J.-Y. Welschinger}, Invent. Math. 162, No. 1, 195--234 (2005; Zbl 1082.14052)]). The second main theorem states that the generating Laurent polynomial for the above invariants equals the refined Block-Göttsche invariant (multiplied with a standard polynomial) for the corresponding tropical enumerative problem. Namely, the tropical enumerative problem reads as the count of plane rational tropical curves of a given (tropical) degree that pass through a fixed configuration of points on the boundary of the tropical toric surface and are counted with the Block-Göttsche weights. The tropical curves in count appear to be trivalent trees, and the Block-Göttsche weight then equals the product of the expressions \(\frac{y^{\mu/2}-y^{-\mu/2}}{y^{1/2}-y^{-1/2}}\) over all the trivalent vertices, where \(\mu\) is Mikhalkin's weight of the vertex (see [\textit{F. Block} and \textit{L. Göttsche}, Compos. Math. 152, No. 1, 115--151 (2016; Zbl 1348.14125)]). These two results uncover a nice geometry hidden in the formally defined Block-Göttsche invariants. One of the exciting consequences is that the count of real curves in the considered problem determines the count of all complex curves in the complexified problem. This, in particular, explains the phenomenon of the logarithmic asymptotic equivalence of the real and complex count.