On dispersion and Markov constants (Q5903295): Difference between revisions

Let \(x_ n=n\delta mod 1\) for an irrational number \(\delta\). Then define \(d_ N=\sup_{0\leq x\leq 1}\min_{1\leq n\leq N}| x-x_ n|\) and \(D(\delta)=\limsup_{N}N d_ N.\) On the other hand the Lagrange-Markov constant M(\(\delta)\) is defined by \(M(\delta)^{-1}=\liminf_{n}n \| n\delta \|.\) The following relation between the dispersion constant D(\(\delta)\) and the Lagrange-Markov constant M(\(\delta)\) is shown: \[ M(\delta)+2\leq 4D(\delta)\leq M(\delta)+2+M(\delta)^{-1}. \] It is shown that in some sense these inequalities are best possible.