A transference principle for simultaneous rational approximation

From MaRDI portal



Abstract: We establish a general transference principle for the irrationality measure of points with mathbbQ-linearly independent coordinates in mathbbRn+1, for any given integer ngeq1. On this basis, we recover an important inequality of Marnat and Moshchevitin which describes the spectrum of the pairs of ordinary and uniform exponents of rational approximation to those points. For points whose pair of exponents are close to the boundary in the sense that they almost realize the equality, we provide additional information about the corresponding sequence of best rational approximations. We conclude with an application.


Len \(n\) be a positive integer and let \(\boldsymbol{\zeta} =(\zeta_0,\dots ,\zeta_n)\) be a point of \(\mathbb R^{n+1}\) whose coordinates are linearly independent over \(\mathbb Q\). For any integer point \(\mathbf x=(x_0,\dots ,x_n)\in\mathbb Z^{n+1}\) we set \(L_{\boldsymbol{\zeta}}(\mathbf x)=\max_{1\leq k\leq n}\mid \zeta_0x_k-\zeta_kx_0\mid\) and for each \(X\geq 1\) we define \(\mathcal L_{\boldsymbol{\zeta}}(X)=\min\{ L_{\boldsymbol{\zeta}} (\mathbf x);\mathbf x\in\mathbb Z^{n+1}\setminus \{ 0\}, \Vert\mathbf x\Vert\leq X\}\), where \(\Vert . \Vert\) denotes the usual Euclidean norm in \(\mathbb R^{n+1}\). For \(S\subset \mathbb Z^{n+1}\) set \(\mathcal L_{\boldsymbol{\zeta}} (X;S)=\min\{ L_{\boldsymbol{\zeta}} (\mathbf x); x\in S, 0\leq \Vert\mathbf x\Vert\leq X\}\) such that \(\min \emptyset =\infty\). Assume that \(\lim_{X\to\infty}\mathcal L_{\boldsymbol{\zeta}} (X;S)=0\). Let \((\mathbf x_i)_{i\geq 0}\) be a sequence of minimal points for \(\boldsymbol{\zeta}\) with respect to \(S\). For each \(i\geq 0\) set \(X_i=\Vert\mathbf x_i\Vert\) and \(L_i=\mathcal L_{\boldsymbol{\zeta}} (X;S)=L_{\boldsymbol{\zeta}} (\mathbf x_i)\). Fix also an index \(i_0\geq 0\). Then the authors prove that for each \(t=1,\dots , n-1\) there exists a largest integer \(i_t\) with \(i_t\geq i_0\) such that \(\dim(\mathbf x_{i_0},\dots ,\mathbf x_{i_t})_{\mathbb R}=t+1\). For these indices \(i_0,\dots , i_{n-1}\) we have \(X_{i_1}\cdots X_{i_{n-1}}\leq c L_{i_0} X_{i_0+1}\cdots L_{i_{n-1}}X_{i_{n-1}+1}\) with a constant \(c>0\) depending only on \(\boldsymbol{ \zeta}\) and not on \(i_0\).











This page was built for publication: A transference principle for simultaneous rational approximation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2211022)