Efficiency of the stochastic approximation method

DOI10.2298/YJOR101124003JMaRDI QIDQ2853274zbMATH OpenOpenAlexFDO

Authors M. Japundžić

Publication date 18 October 2013

Published in YUJOR. Yugoslav Journal of Operations Research (Search for Journal in Brave)

Full work available at URL https://doi.org/10.2298/yjor101124003j

zbMATH Keywords

efficiency of the stochastic methodsnoisestep length

Mathematics Subject Classification ID

Optimality conditions for problems involving randomness (49K45) White noise theory (60H40) Stochastic approximation (62L20)

Review text

The author presents a set of sufficient conditions for almost sure convergence of stochastic approximation method, with emphasis on the condition that refers to the step length. Next, he analyses of choice of the coefficients that generate step length, in order to achieve better performance of the algorithm. Numerical results in Matlab are also presented, which justify the proposed choice of coefficients.

Recommendations

• Sufficient and necessary condition for the convergence of stochastic approximation algorithms
• Asymptotically efficient stochastic approximation
• Necessary and sufficient conditions for a stochastic approximation method
• scientific article; zbMATH DE number 952465
• Adaptive stochastic approximation algorithm

Describes a project that uses

Uses Software

• Matlab

This page was built for publication: Efficiency of the stochastic approximation method

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2853274)