Generalization of Two Results of the Theory of Uniform Distribution

From MaRDI portal





Let \(x_ 1,...,x_ N\) be numbers in \([0,1]\) and let \(p_ 1,...,p_ N\) be nonnegative weights with sum equal to~1. Define \[ g(t)=\sum^{N}_{k=1}p_ K c_ t(x_ k)-t\text{ for } 0\leq t\leq 1, \] where \(c_ t\) is the characteristic function of the interval \([0,t)\). Then \(D_ N=\sup_{0\leq t\leq 1}| g(t)|\) is called the discrepancy of \(x_ 1,...,x_ N\) with respect to the weights \(p_ 1,...,p_ N\). Let \(\phi\) be an increasing function on \([0,1]\) with \(\phi (0)=0\) and let \(F(x)=\int^{x}_{0}\phi (t) \,dt\). It is shown that \[ F(D_ N)\leq \int^{1}_{0}\phi (| g(t)|) \,dt\leq \phi (D_ N). \] In the special case where \(\phi (x)=x^ p\), \(0<p<\infty\), and \(p_ k=1/N\) for \(1\leq k\leq N\), one obtains a result of the reviewer [H. Niederreiter, Répartition modulo 1, Actes Colloq. Marseille-Luminy 1974, Lect. Notes Math. 475, 132--154 (1975; Zbl 0306.10034)]. The generalization is proved by the same method as the special case. The paper also contains criteria for the uniform distribution of triangular arrays of finite sequences.











This page was built for publication: Generalization of Two Results of the Theory of Uniform Distribution

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3731710)