Tropical intersection products on smooth varieties (Q653290)

This paper extends recent work of the author's mathematical sibling [\textit{L. Allermann} and \textit{J. Rau}, Math. Z. 264, No. 3, 633--670 (2010; Zbl 1193.14074); \url{arxiv:0709.3705}] and of their adviser and his collaborators [\textit{A. Gathmann, M. Kerber} and \textit{H. Markwig}, Compos. Math. 145, No. 1, 173--195 (2009; Zbl 1169.51021); \url{arxiv:0708.2268}] to offer a possible route towards intersection theory in the rapidly emerging field of tropical geometry. The author seems to have omitted an introduction, applications, and motivation from this paper (perhaps because this is essentially embedded in the two above-cited works), so allow me to attempt to fill this gap and offer the reader a proper context for the present work and an interpretation of its position in the mathematical landscape. Based on ideas and a suggestion of Kontsevich, Mikhalkin demonstrated that certain enumerative invariants (e.g., the number of rational curves in the plane passing through a fixed number of points) coincide with their tropical counterparts. The former invariants are special instances of a well-developed theory called Gromov-Witten (GW) theory, in which the enumerative invariants arise as intersection numbers on a compactified moduli space of maps from an abstract curve to the target space. The latter, on the other hand, were for the most part defined only combinatorially without proper foundations as the number of tropical curves (metric graphs) passing through the specified points. As such, various basic properties such as the independence of the location of these points, assuming sufficient generality, had to be addressed by very cumbersome and ad hoc methods, whereas on the algebro-geometric side this invariance follows immediately from the intersection-theoretic interpretation of the enumerative invariants. Thus, in an effort to remedy the situation and place the tropical side on firm foundations, \textit{G. Mikhalkin} sketched in his 2006 Madrid ICM talk [in: Proceedings of the international congress of mathematicians (ICM), Madrid, Spain, August 22--30, 2006. Volume II: Invited lectures. Zürich: EMS. 827--852 (2006; Zbl 1103.14034); \url{arxiv:math/0601041}) a possible framework for tropical intersection theory. Unfortunately he only provided definitions and no proofs of any statements, so the subsequent series of papers, including the present one, are part of the elaboration, fleshing out, and in some cases correcting, of Mikhalkin's sketch. The launching point for these papers on tropical intersection theory is that tropical varieties are polyhedral complexes with weights (or multiplicities) associated to their top-dimensional polyhedra. Moreover, the gluing of the constituent polyhedra is given locally by the data of a fan structure -- and when incorporating the weights, it is something called a balanced fan. The papers in this series do not address the issue of tropicalization -- that is, the bridge from the classical world -- they simply take the polyhedral objects as the intrinsic definition of a tropical variety and proceed to develop a purely combinatorial intersection theory (with inspiration from the classical setting) for it. It should be pointed out that this is not necessarily a universally accepted point of view of tropical geometry. Nonetheless, it is sufficient for the original intent of the authors: the tropical Kontsevich moduli space (of maps from a tropical curve to a tropical target, e.g., the tropical plane) can be built directly from fan data -- rather than something like trying to tropicalize the original Kontsevich space as an actual variety/stack -- and the enumerative invariants appearing in Mikhalkin's correspondence theorem are indeed compatible with this fan-based tropical intersection theory. Thus Gathmann-Kerber-Markwig accomplish their original task of inheriting independence of the position of the points from an incipient tropical intersection that, though not fully developed, does apply to the tropical Kontsevich moduli space. Based on the success of this application, and on the lack of a more general and fully-formed tropical intersection theory, the Allermann-Rau paper abandons specific applications as a goal and proceeds to round out this fan-based intersection theory. Their theory works best at producing a tropical cycle when intersecting so-called Cartier divisors with a tropical cycle, as opposed to intersecting tropical cycles with other tropical cycles. Nonetheless, and despite the lack of definitive applications, the properties they verify do closely enough mirror the corresponding algebro-geometric properties that one is tempted to believe this is not merely a coincidence and there is indeed lurking in the background a deeper explanation for why this fan-based tropical intersection theory seems to work. Rather than developing tropical intersection theory as a parallel yet independent branch of mathematics to algebro-geometric intersection, it would be nice to see in future work a tighter integration of the two theories -- for instance, perhaps an explanation of why classical intersection multitiplicites/cycles tropicalize to these fan-based ones could help put Mikhalkin's correspondence theorem in a much more general framework. In the mean time, however, Allermann has decided in this paper to continue developing this fan-based tropical intersection as an intrinsic piece of mathematics. Deciding, without any motivation/explanation, that a ``smooth'' tropical variety is one built locally out of tropicalized linear spaces, he shows in this paper how to extend ideas from his previous one with Rau to yield an intersection pairing of tropical cycles (i.e., how to intersect to cycles and produce a third cycle) in the case of one of these so-called smooth tropical varieties. This is accomplished by a careful study of the case of tropicalized linear spaces which then extends almost immediately to the global setting. The paper once again demonstrates various desirable properties of this intersection pairing which suggest that it is somehow the ``correct'' one, but the paper leaves the reader a bit desirous of a follow-up as there is almost nothing in the way of applications, motivation for what this tropical intersection pairing could be used for, or explaining how it relates to the classical world. Nonetheless, it is a nice step towards these future goals and also a nice illustration that tropical geometry really does seem to boil everything down to essentially pure combinatorics.