Transformations that preserve uniform distribution

From MaRDI portal





A map \(T\) of the unit interval \(I\) to itself is called uniform distribution preserving (u.d.p.) if \(\{T(x_n)\}^{\infty}_{n=1}\) is uniformly distributed in \(I\) whenever \(\{x_n\}^{\infty}_{n=1}\) is a uniformly distributed sequence in \(I\). In the first part of the paper the authors give a collection of simple characterisations of u.d.p. transformations. For example (Theorem 2b): \(T\) is u.d.p. if and only if \(\int^1_0 T^n(x) \,dx=1/(n+1)\) for every \(n=1,2,\ldots\), or (Theorem 4): \(T\) is u.d.p. if and only if \(T\) is measurable in the Jordan sense and \(| T^{-1}(I_1) | =| I_1|\) for every interval \(I_1\subset I\). (\(| E |\) denotes the Jordan or Lebesgue measure of \(E)\). In the second part some properties of u.d.p. transformations are given. In the third part more extensively piecewise linear u.d.p. transformations are studied and explicitly characterized. Moreover all these u.d.p. piecewise linear transformations are determined for which there is any \(x\in I\) such that the orbit \(\{T^n(x)\}^{\infty}_{n=1}\) of \(x\) is uniformly distributed in \(I\). In the last part it is shown that the system of the u.d.p. transformations is nowhere dense in the space \(M\) of all maps of \(I\) to itself, and that it is a perfect set in \(M\).











This page was built for publication: Transformations that preserve uniform distribution

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3803214)