Irreducible cycles and points in special position in moduli spaces for tropical curves (Q1953332)

In many respects, the objects we call tropical varieties seem perhaps more analogous to cycles than varieties. For instance, balanced polyhedral complexes with the data of the facet weights are closely related to cycle classes on toric compactifications of the ambient torus. The distinction of cycle versus cycle class is a bit murky, as the fan-based intersection pursued by Gathmann et al. is based on tropical varieties up to refinement, rather than the presumably weaker notion of rational equivalence used in classical intersection theory. Another yet-to-be fully understand part of the story is the notion of a \textit{prime} cycle. Indeed, irreducibility is not well understood in the tropical setting: a standard example is the decomposition of 6 planar line-segments into two tropical Y-graph ``lines'' versus three straight lines. Work of Katz and Huh on proving Rota's conjecture about log concavity of coefficients of the chromatic polynomial of a matroid suggests the need for an appropriate definition of prime cycle in the tropical setting. The present paper offers and studies one such possible definition, not in the framework of Katz-Huh, but rather with the running example of the moduli space of rational tropical curves. This is perhaps not too far from the Katz-Huh setting, since there the relevant ambient space is the toric variety associated to the permutahedron, which is the Losev-Manin space, an example of Hassett's weighted variants of the moduli space of stable rational curves. It would be interesting if the notion of irreducibility introduced in the present paper relates to that needed in the context of matroids. For now, however, the authors are content to simply illustrate that some natural objects (boundary cycles and psi classes) in the tropical setting satisfy their notion of irreducibility, though any theorems or further justification for their definition seems postponed till future work. The paper also discusses a (largely independent) topic of general position for points. More specifically, the authors show that there are two natural ways of defining general position for points in the tropical torus, and the complement of these naturally has the structure of a so-called tropical cycle of codimension one. The paper is thus largely exploratory, simply looking at examples and phenomena related to moduli of tropical curves, perhaps making one small step towards a comprehensive understanding of tropical intersection theory and its role in, e.g., GW theory.