Sparse polynomial equations and other enumerative problems whose Galois groups are wreath products (Q2065446)

This remarkable paper deals with the monodromy of solutions of generic systems of \(n\) algebraic equations in \(n\) non-zero complex variables. The Bernstein-Koushnirenko theorem states that the number of solutions of such a system does not depend on the (generic) coefficients and equals the mixed lattice volume of the Newton polytopes of the given equations. Fixing the set of monomials for each equation, one obtains the monodromy group or the Galois group of the system induced by the fundamental group of the complement to the discriminant on the space of all systems spanned by fixed monomials. Recently, the first author [Compos. Math. 155, No. 2, 229--245 (2019; Zbl 1451.14152)] showed that the Galois group of a reduced, irreducible system of generic equations is isomorphic to the symmetric group on the number of solutions. In the case of one univariate equation, one obtains the full symmetric group whenever the gcd of the exponents of the monomials equals \(1\). However, if the gcd is non-trivial, the Galois group appears to be the wreath product of the reduced symmetric group and a cyclic group. The current paper is mainly devoted to the study of the latter phenomenon in the multivariate case. The expected multivariate answer is as follows. Let \(A_1,\dots,A_n\subset{\mathbb Z}^n\) be the sets of monomials for the considered system of algebraic equations, \(\widetilde d\) the mixed volume of the Newton polytopes \(P(A_1),\dots,P(A_n)\), and let \(\Lambda\subset{\mathbb Z}^n\) be a full rank sublattice of index \(k\) generated by the differences of the points of \(A_1,\dots,A_n\). Then the expected Galois group is the wreath product of the symmetric group \(S_d\), \(d=\widetilde d/k\), and the cyclic group \({\mathbb Z}_k\). The authors show that the actual answer is not always the expected one, and it is widely open though, in principle, depends only on the combinatorial data. The main theorem, proved in the paper, provides almost coinciding necessary and sufficient conditions for the Galois group to be the expected one. A very particular instance of it sounds like that: If the Newton polytopes \(P(A_1),\dots,P(A_n)\) share the same dual fan, then the Galois group is expected if the tuple \((A_1,\dots,A_n)\) is combinatorially ample, and the Galois group is strictly smaller than the expected one, otherwise. To prove the main theorem and to provide applications to enumerative geometry and to the problem of connectivity of covering spaces, the authors develop new combinatorial and topological techniques. In particular, the connectivity of covering spaces in certain important cases can be reduced to the study of the first homology and its behavior under covering, instead of consideration of fundamental groups. An application to enumerative problems consists in computaion of Galois (or monodromy) groups whenever the counted objects appear in a family, and the enumeration is maid for a generic choice of the parameters. A typical example is the Galois group of the enumeration of roots of generic systems of algebraic equations. Another important example is the enumeration of lines in a projective space that meet and appropriate collection of projective subspaces or points. The novelty of the authors' approach is that they separate a class of wreath enumerative problems, for which they provide criteria to have an expected Galois group. Numerous illustrating examples and a very accurate exposition help much to follow the argument.