Functional CLT for nonstationary strongly mixing processes (Q2288740)

The authors prove a weak invariance principle for triangular schemes of weakly dependent random variables. They do not assume stationarity, so they study a version of the partial sum process with a transformed time to achieve convergence to a standard Wiener process. The theorem does only require mild assumptions, mainly the Lindeberg-condition and a mixing condition which is weaker than the classical strong mixing condition. Examples include Markov chains under some conditions and linear processes with strongly mixing observations. The proof is based on martingale approximations.