On discrepancies of irrational rotations: an approach via rational rotation (Q682126)

From MaRDI portal
scientific article
Language Label Description Also known as
English
On discrepancies of irrational rotations: an approach via rational rotation
scientific article

    Statements

    On discrepancies of irrational rotations: an approach via rational rotation (English)
    0 references
    0 references
    0 references
    13 February 2018
    0 references
    For continued fractions, let \(a_n\) and \(p_n/q_n\) denote the \(n\)th partial quotient and the \(n\)th convergent, respectively. Let \(0<\alpha<1\) be an irrational number with a single isolated large quotient \(a_{\eta}\) in its continued fraction expansion. For example \(\alpha=\log_{10}7\) has \(a_{\eta}=4813\) for \(\eta=7\) and \(\alpha=\pi-3\) has \(a_{\eta}=292\) for \(\eta=4\). The authors give some estimates of discrepancies of irrational rotations based on such irrational numbers. For the initial segment of the graph of the discrepancy it is shown that if \(\eta \geq 3\) and \(n=\nu q_{\eta-1}\) \((0<\nu<a_{\eta})\), then \(D_n^*(\{n\alpha\})=\frac 1{q_{\eta -1}}(1-\Delta-\frac{\nu-1}{L})\), where \(\Delta=| q_{\eta-1}\alpha-p_{\eta-1}| \) and \(L=1/(q_{\eta-1}\Delta)\). For the next segment they also prove that for \(\eta\geq 3\) the following estimate holds \(nD_n^*(\{n\alpha\})=\nu (1-\Delta-\frac{\nu-1}{L})+\theta\) for \(n=q_{\eta+1}+\nu q_{\eta-1}\) if \(a_{\eta+1}=1\), for \(n=q_{\eta}+\nu q_{\eta-1}\) if \(a_{\eta+1}\geq 2\) and for \(0<\nu<a_{\eta}\), where \(0\leq \theta\leq 2\). Moreover it is derived that if \(\eta\geq 3\) and \(\lambda>\eta\), then \(nD_n^*(\{n\alpha\})=\nu (1-\Delta -\frac{\nu-1}{L})+\theta\) for \(n=q_{\lambda}+\nu q_{\eta-1}\), \(0<\nu<a_{\eta}\), where \(0\leq \theta\leq 2\).
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    rational rotations
    0 references
    irrational rotations
    0 references
    isolated large partial quotients
    0 references
    continued fractions
    0 references
    discrepancy
    0 references
    0 references