On uniform distribution of subsequences (Q1063630): Difference between revisions

A multi-sequence \(\omega =(x_ n)\), \(n\in {\mathbb{N}}^ r\) with elements \(x_ n\) in a compact topological AA2-space X is said to be P-uniformly distributed \((P=(p_ 1(n_ 1),...,p_ r(n_ r))\), \(n=(n_ 1,...,n_ r)\) a positive weighted mean) with respect to the regular normed Borel measure \(\mu\) on X (for short: P(\(\mu)\)-u.d.) if and only if \[ (*)\quad \lim_{N_ 1,...,N_ r\to \infty}(1/P(N))\sum_{n\leq N}p(n) f(x_ n)=\int_{X}f(x) d\mu \] for all continuous real-valued functions f on X (the coordinates of \(N=(N_ 1,...,N_ r)\) are tending independently to \(\infty)\); \(p(n)=\prod p_ j(n_ j)\), \(P_ j(n_ j)=\sum_{k\leq n_ j}p_ j^{(k)}\), \(P(n)=\prod P_ j(n_ j)\). \(\omega\) is called P- weakly uniformly distributed with respect to \(\mu\) ((P,\(\mu)\)-w.u.d.) iff in (*) the coordinates \(N_ 1,...,N_ r\) are tending simultaneously to \(\infty\) \((N_ 1=...=N_ r=N\to \infty)\). In a natural way a probability measure is introduced on the space of all subsequences of a given multi-sequence \(\omega\). Theorem II: Supposing certain regularity conditions on P, \(\omega\) is (P,\(\mu)\)-w.u.d. if the set of all (P,\(\mu)\)-w.u.d. subsequences of \(\omega\) has positive measure. - Theorem I: Supposing certain regularity conditions on P, almost all subsequences of \(\omega\) are (P,\(\mu)\)-w.u.d. if \(\omega\) is (P,\(\mu)\)-w.u.d. The statement of Theorem II remains true for (P,\(\mu)\)-u.d. multi- sequences but not the statement of Theorem I; a counter-example is given in Theorem III \((r=2\), \(X=\{-1,0,1\}\), \(\mu (0)=0\), \(\mu (-1)=\mu (1)=1/2\), \(p_ 1=p_ 2=1)\). In Theorems IV, V it is shown that Theorems I, II do not hold for arbitrary means (without any regularity conditions); counter-examples are given in the case \(r=1\). Different constructions were given by the first author [Lect. Notes Math. 1114, 84- 97 (1985; Zbl 0558.10044)]. The proofs of these results make use of the central limit theorem and the law of large numbers. In the case \(r=1\) Theorems I, II are due to \textit{F. R. Keogh} and \textit{G. M. Petersen} [Q. J. Math., Oxf. II. Ser. 12, 33-44 (1961; Zbl 0097.04701)]; in the case \(r>1\) weaker versions of Theorems I, II are due to the second author [Sitzungsber., Abt. II, Österr. Akad. Wiss., Math.-Naturwiss. Kl. 188, 317-327 (1979; Zbl 0446.10043)]. The proofs in the cited papers are based on summation properties of Rademacher functions.