scientific article; zbMATH DE number 176293

From MaRDI portal
Publication:4035402





[For the entire collection see Zbl 0752.00047.] The Robbins-Monro stochastic approximation procedure in the case of the presence of delayed observations is investigated, more precisely the approach proposed by \textit{M. A. Mahmoud} [Asymptotic Statistics, 4th Prague Symp., Prague/Czech. 1988, 399-403 (1989)] consisting in the rule that the next experiment is made in the same point as the preceding one if no result of any observation has become known during the interval between. For this procedure, a convergence theorem in the sense almost sure, an invariance principle and the asymptotic normality are proved under rather restrictive boundedness assumptions in comparison with the ``classical Robbins-Monro procedures.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4035402)