Real log curves in toric varieties, tropical curves, and log Welschinger invariants (Q2675333)

The tropical enumerative geometry provides an efficient way to count complex and real curves. The authors develop a log deformation theory approach to establish a correspondence between enumeration of tropical curves on one side and complex and real curves on the other side with a focus on the real case. The resulting technical tool can be regarded as a real version of the Nishinou-Siebert correspondence for rational complex curves [\textit{T. Nishinou} and \textit{B. Siebert}, Duke Math. J. 135, No. 1, 1--51 (2006; Zbl 1105.14073)] and of the Nishinou correspondence for non-superabundant curves of positive genera [\textit{T. Nishinou}, ``Correspondence theorems for tropical curves. I'', Preprint, \url{arXiv:0912.5090}]. The main authors' theorem compares enumeration of tropical curves in the Euclidean space \({\mathbb R}^n\) of arbitrary dimension \(n\ge2\) and enumeration of real curves in an appropriate real toric variety. Namely, the tropical data consists of the genus \(g\ge0\), the tropical degree \(\Delta\), and the tuple \({\mathbb A}\) of affine linear subspaces (tropical constraints) such that the expected dimension of the space of tropical curves in \({\mathbb R}^n\) of genus \(g\) and degree \(\Delta\), matching the given tropical constraints is zero. Moreover, the extra non-superabundance condition requires that the latter space of tropical curves \({\mathcal T}_{g,\Delta}({\mathbb A})\) is finite (the non-superabundance condition always holds for either \(g=0\), or \(n=2\)). The algebraic data consists of a real \(n\)-dimensional toric variety \(X\) determined by \(\Delta\), the same genus \(g\), and the tuple of families over the punctured disk of constraints being orbits of the action of subtori determined by the tuple \({\mathbf A}\). The statement is that the number of real curves in \(X\) of genus \(g\), intersecting the toric divisors in a way prescribed by \(\Delta\), and matching the given constraints for the sufficiently small real value of the parameter (so-called ``constraints close to the tropical limit'') equals the number of tropical curves in \({\mathcal T}_{g,\Delta}({\mathbf A})\) counted with appropriate weights. This weight splits into three factors, one of which resembles the ``complex'' weight introduced by Nishinou and Siebert and the two others reflect the real structure. Another authors' result is the new proof of Mikhalkin's theorem [\textit{G. Mikhalkin}, J. Am. Math. Soc. 18, No. 2, 313--377 (2005; Zbl 1092.14068)] which expresses the number of real curves of a given degree and genus in a given toric surface that are counted with Welschinger signs via enumeration of appropriate plane tropical curves of the same ganus and corresponding tropical degree that are counted with tropical Welschinger signs. Some open problems and further directions of study are discussed.