zbMATH — the first resource for mathematics

The convergence rate in the invariance principle for differently distributed variables with exponential moments. (Russian) Zbl 0541.60024
Predel’nye Teoremy dlya Summ Sluchajnykh Velichin, Tr. Inst. Mat. 3, 4-49 (1984).
Let \((\eta_ j)\) be a sequence of normally distributed r.v.’s with \(E\eta_ j=0\) and \(E\eta^ 2_ j<\infty\). Let \(F_ j(x)\) be distribution functions such that \(\int xF_ j(dx)=0\), \(\int x^ 2F_ j(dx)<\infty\) and for some \(\lambda>0 \lambda \int | x|^ 3\exp(\lambda | x|)F_ j(dx)\leq \int x^ 2F_ j(dx)\quad \forall j.\) The sequence of independent r.v.’s \((\xi_ j)\) with distribution functions \((F_ j(x))\) and such that \(E \exp(e\lambda \Delta_ n)\leq 1+\lambda B_ n,\) where \(c>0\) is some absolute constant, \(B^ 2_ n=\sum_{j\leq n}D\xi_ j, \Delta_ n=\max_{m\leq n}| \sum_{j\leq m}\xi_ j-\sum_{j\leq m}\mu_ j|,\) is constructed. This result generalizes the earlier one of J. Komlos, P. Major, and G. Tusnady, Z. Wahrscheinlichkeitstheor. Verw. Geb. 34, 33-58 (1976; Zbl 0307.60045).
Reviewer: K.Kubilius

60F15 Strong limit theorems
60F17 Functional limit theorems; invariance principles