On multiplicatively dependent vectors of polynomial values
The aim of the paper is to count multiplicatively dependent polynomial and rational function values under certain constraints.\N\NFor a subset \(S\in\mathbb{C}^m\) and an \(n\)-tple of rational functions \(F=(f_1,\ldots,f_n)\in \mathbb{C}(X_1,\ldots,x_m)^n\) denote by \(N_F(S)\) the number of \(n\)-tuples \(({u}_1,\ldots,{u}_n)\in S^n\) such that \((f_1({u}_1),\ldots,f_n({u}_n))\) is multiplicatively dependent and by \(N_F^*(S)\) the number of \({u}\in S\) such that \((f({u}_1),\ldots,f_n({u}))\) is multiplicatively dependent. For \({v}\in (\mathbb{C}^\times)^n\), the multiplicative rank \(s\) of \({v}\) is 0 if \({v}\) has a coordinate which is a root of unity and otherwise the largest integer \(s\) with \(1\le s\le n\) for which any \(s\) coordinates of \({v}\) form a multiplicatively independent vector. For \(0\le s<n\), \(N_{F,s}(S)\subset N_F(S)\) denotes the elements \(({u}_1,\ldots,{u}_n)\) of \(N_F(S)\) such that \((f_1({u}_1),\ldots,f_n({u}_n))\) is multiplicatively dependent of rank \(s\). For \(K\) a global field, \(f\in K[X_1,\ldots,x_m]\) and \(k\ge 1\), \(f\) is \(k\)-irreducible over \(K\) if \(F\) does not have any factors of degree at most \(k\) over \(K\). For an algebraic number \(\alpha\) with minimal polynomial \(a_dX^d+\cdots+a_1X+a_0\) over \(\mathbb{Z}\) and conjugates \(\alpha_i\) \((1\le i\le d)\), the (absolute Weil) height is \(H(\alpha) = (|a_d|\prod_{i=1}^d\max\{1,|\alpha_i|\})^{1/d}\).\N\NThe first finiteness result is reached quickly. Let \(F=(f_1,\ldots,f_m)\in \mathbb{Z}[X_1,\ldots,X_m]^n\) and \(S\subseteq\mathbb{Z}^m\). If pairwise the \(f_i\) have no common zeros in \(\mathbb{C}^m\), \(f_i({u})=\pm1\) has only finitely many solutions \({u}\in S\) for each \(i\) and, for all but at most one \(i\), \(P^+(f_i({u}))\to\infty\) as \(\max_{1\le i\le m}|u_i|\to\infty\) for \({u}=(u_1,\ldots,u_m)\in S\) where \(P^+(x)\) denotes the largest prime factor of \(x\in\mathbb{Z}\), then \(N_F^*(S)\) is finite. In particular, if \(F=(f_1,\ldots,f_n) \in \mathbb{Z}[X,Y]^n\), each \(f_i\) has at least three distinct linear factors and the \(f_i\) are pairwise coprime in \(\mathbb{C}[X,Y]\) and \(S=\{(x,y)\in\mathbb{Z}^2:\gcd(x,y)=1\}\) then all the conditions are satisfied and \(N_F^*(S)\) is finite. The proof of the finiteness result follows from the observation that for a collection \(f_1,\ldots,f_n\in\mathbb{Z}[X_1,\ldots,X_n]\) of coprime polynomials, \(\gcd(f_1({u}),\ldots,f_n({u}))\) takes only finitely many values as \({u}\) ranges over \(\mathbb{Z}^n\). This is a consequence of the Nullstellensatz.\N\NQuantitative results follow from bounds on the number of integral solutions of polynomial solution and on the size of a multiplicative dependence relation between algebraic numbers. Gven \(n\ge 2\) and non-zero multiplicatively dependent algebraic numbers \(\alpha_1,\ldots,\alpha_n\) with degree at most \(D\) and height at most \(H\), there is a positive number \(A\) depending only on \(n\) and \(D\) and rational integers \(k_1,\ldots,k_n\) not all zero such that \(\alpha_1^{k_1}\cdots\alpha_n^{k_n}=1\) and \(\max_{1\le i\le n}|k_i| < A(\log H)^{n-1}\). This leads to a quantitative version of the finiteness result. If \(F=(f_1,\ldots,f_n)\in\overline{\mathbb{Q}}(X_1,\ldots,x_n)^n\) and \(S\subseteq \{x\in\overline{\mathbb{Q}}: H(x)\le H, [\mathbb{Q}(x):\mathbb{Q}]\le D\}\), then \(N_F(S^m)\ll |S|^{mn-1}(\log H)^{n^2-1}\), where the implied constant depends only on \(f\) and \(D\).\N\NMore intricate arguments yield improved bounds under certain conditions on \(F\) which still allow relatively generic examples. Let \(K\) be a number field of degree \(D\) over \(\mathbb{Q}\) with ring of integers \(\mathcal{O}_K\). Let \(\mathcal{B}_K(H)\) denote the algebraic integers in \(\mathcal{O}_K\) of height at most \(H\). Call a polynomial \(f\in K[X_1,\ldots,X_m]\) cylindrical over a curve if there exist a polynomial \(g\in K[Y_1,Y_2]\) and linear forms \(\ell_1({X}),\ell_2({X})\) over \(K\) such that \(f({X})=g(\ell_1({X}),\ell_2({x}))\). In charcteristic zero, this means that the vector subspace of \(K[X_1,\ldots,X_m]\) spanned by \(\{\partial f/\partial X_i\}_{i=1}^m\) has dimension at least 3. Let \(F=(f_1,\ldots,f_n)\in\mathcal{O}[X_1,\ldots,X_m]^n\). If \(m>1\), assume each \(f_i-\alpha\) is irreducibe and not cylindrical over a curve for all \(\alpha\in\mathcal{O}_K\), the homogeneous part of \(f_i\) of degree \(\deg f_i\) is 1-irreducible (and geometrically irreducible if \(m=2\)) and \(d=\min_i\deg f_i\ge 3\). Then \(N_F(\mathcal{B}(H)^m)\ll H^{D(mn-v(m,d))}(\log H)^{rm(n-1)+\kappa+3}\), where \(\kappa\) depends only on \(m\), \(v(m,d)\) is an explicit number between 1 and 2 and the implied constant depends only on \(F\) and \(K\). There are more precise statements for larger \(m\) and \(d\). There is a similar result for \(N_{F,s}(\mathcal{B}_K(H)^m)\) showing that most of the contribution to the bound comes from the rank 0 and rank 1 terms.
- Coprime values of polynomials in several variables
- Effective results for hyper- and superelliptic equations over number fields
- scientific article; zbMATH DE number 1467843 (Why is no real title available?)
- scientific article; zbMATH DE number 1402564 (Why is no real title available?)
- scientific article; zbMATH DE number 3226335 (Why is no real title available?)
- Improvements on dimension growth results and effective Hilbert's irreducibility theorem
- Integral points of fixed degree and bounded height
- Multiplicative Dependence Among Iterated Values of Rational Functions Modulo Finitely Generated Groups
- Multiplicative dependence of rational values modulo approximate finitely generated groups
- Multiplicative relations in number fields
- On abelian multiplicatively dependent points on a curve in a torus
- On abelian points of varieties intersecting subgroups in a torus
- On multiplicative dependence of values of rational functions and a generalization of the Northcott theorem
- On multiplicatively dependent vectors of algebraic numbers
- On Some Inequalities for Polynomials in Several Variables
- On the greatest prime factors of decomposable forms at integer points
- Sharp bounds for the number of rational points on algebraic curves and dimension growth, over all global fields
- Uniform bounds for the number of rational points on varieties over global fields
This page was built for publication: On multiplicatively dependent vectors of polynomial values
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6896442)
