# Tropical geometry and polynomial systems

Abstract : Real polynomial systems are ubiquitous in many areas of pure and applied mathematics. A. Khovanskii provided a fewnomial upper bound on the number of non-degenerate positive solutions of a real polynomial system of n equations in n variables that depends only on the number of monomials appearing in the equations. The latter bound was recently improved by F. Bihan and F. Sottile, but the resulting bound still has room for improvement, even in some simple cases.The aim of this work is to tackle three main problems in Fewnomial theory. Consider a family of real polynomial systems with a given structure (for instance, supports or number of monomials). One problem is to find good upper bounds for their numbers of real (or positive) solutions. Another problem is to construct systems whose numbers of real (or positive) solutions are close to the best known upper bound. When a sharp upper bound is known, what can be said about reaching it?In this thesis, we refine a result by M. Avendaño by proving that the number of real intersection points of a real line with a real plane curve defined by a polynomial with at most t monomials is either infinite or does not exceed 6t -7. Furthermore, we prove that our bound is sharp for t=3 using Grothendieck's real dessins d'enfant. This shows that the maximal number of real intersection points of a real line with a real plane trinomial curve is eleven.We then consider the problem of estimating the maximal number of transversal positive intersection points of a trinomial plane curve and a t-nomial plane curve. T-Y Li, J.-M. Rojas and X. Wang showed that this number is bounded by 2^t-2, and recently P. Koiran, N. Portier and S. Tavenas proved the upper bound 2t^3/3 +5t. We provide the upper bound 3*2 {t-2} - 1 that is sharp for t=3 and is the tightest for t=4,...,9. This is achieved using the notion of real dessins d'enfant. Moreover, we study closely the case t=3 and give a restriction on the supports of systems reaching the sharp bound five.A circuit is a set of n+2 points in mathbb{R} n that is minimally affinely dependent. It is known that a system supported on a circuit has at most n+1 non-degenerate positive solutions, and that this bound is sharp. We use real dessins d'enfant and Viro's combinatorial patchworking to give a full characterization of circuits supporting polynomial systems with the maximal number of non-degenerate positive solutions.We consider polynomial systems of two equations in two variables with a total of five distinct monomials. This is one of the simplest cases where the sharp upper bound on the number of non-degenerate positive solutions is not known. F. Bihan and F. Sottile proved that this sharp bound is not greater than fifteen. On the other hand, the best examples had only five non-degenerate positive solutions. We consider polynomial systems as before, but defined over the field of real generalized locally convergent Puiseux series. The images by the valuation map of the solutions of such a system are intersection points of two plane tropical curves. Using non-transversal intersections of plane tropical curves, we obtain a construction of a real polynomial system as above having seven non-degenerate positive solutions.
Keywords :
Document type :
Theses

https://tel.archives-ouvertes.fr/tel-01679316
Contributor : Abes Star :  Contact
Submitted on : Tuesday, January 9, 2018 - 6:07:11 PM
Last modification on : Wednesday, October 14, 2020 - 4:10:48 AM
Long-term archiving on: : Friday, May 4, 2018 - 12:37:06 AM

### File

EL_HILANY_2016_archivage.pdf
Version validated by the jury (STAR)

### Identifiers

• HAL Id : tel-01679316, version 1

### Citation

Boulos El Hilany. Tropical geometry and polynomial systems. General Mathematics [math.GM]. Université Grenoble Alpes, 2016. English. ⟨NNT : 2016GREAM037⟩. ⟨tel-01679316⟩

Record views