@MISC{Delyon94adeterministic, author = {Bernard Delyon and Bernard Delyon}, title = {A Deterministic Approach To Stochastic Approximation}, year = {1994} }
Share
OpenURL
Abstract
: A deterministic approach is proposed for proving the convergence of stochastic algorithms of the most general form, under necessary conditions on the input noise, and reasonable and verifiable conditions on the (non-necessarily continuous) mean field. A condition is also given for obtaining convergence when the mean field has a non-discrete set of stationary points. Incidentally, a Lyapunov theorem is given for non-necessarily stable matrices. Key-words: Stochastic approximation, Lyapunov theorem (R'esum'e : tsvp) delyon@irisa.fr Centre National de la Recherche Scientifique Institut National de Recherche en Informatique (URA 227) Universit e de Rennes 1 -- Insa de Rennes et en Automatique -- unit e de recherche de Rennes Une approche d'eterministe pour l'approximation stochastique R'esum'e : On propose ici une approche d'eterministe au probl`eme de la convergence des algorithmes stochastiques g'en'eraux. Les conditions demand'ees sur le bruit d'excitation sont n'ecessaires et s...