Title :
Some guidelines for using iterate averaging in stochastic approximation
Author_Institution :
Appl. Phys. Lab., Johns Hopkins Univ., Laurel, MD, USA
Abstract :
Averaging of the output (iterates) from a stochastic approximation (SA) recursion has been shown to be a useful technique for the gradient-based Robbins-Monro form of SA. For the gradient-free form, iterate averaging can produce an improvement in the stability of the algorithm and competitive mean-square errors relative to the standard (unaveraged) recursion. We discuss guidelines on how and when to use averaging in this context
Keywords :
approximation theory; iterative methods; competitive mean-square errors; gradient-based Robbins-Monro form; gradient-free form; iterate averaging; stochastic approximation recursion; Approximation algorithms; Finite difference methods; Guidelines; Laboratories; Loss measurement; Mean square error methods; Performance loss; Physics; Stability; Stochastic processes;
Conference_Titel :
Decision and Control, 1997., Proceedings of the 36th IEEE Conference on
Conference_Location :
San Diego, CA
Print_ISBN :
0-7803-4187-2
DOI :
10.1109/CDC.1997.657115
