Comparative study of stochastic algorithms for system optimization based on gradient approximations
Author:
- 1 April 1997
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics)
- Vol. 27 (2) , 244-249
Abstract
Stochastic approximation (SA) algorithms can be used in system optimization problems for which only noisy measurements of the system are available and the gradient of the loss function is not. This type of problem can be found in adaptive control, neural network training, experimental design, stochastic optimization, and many other areas. This paper studies three types of SA algorithms in a multivariate Kiefer-Wolfowitz setting, which uses only noisy measurements of the loss function (i.e., no loss function gradient measurements). The algorithms considered are: the standard finite-difference SA (FDSA) and two accelerated algorithms, the random directions SA (RDSA) and the simultaneous-perturbation SA (SPSA). RDSA and SPSA use randomized gradient approximations based on (generally) far fewer function measurements than FDSA in each Iteration. This paper describes the asymptotic error distribution for a class of RDSA algorithms, and compares the RDSA, SPSA, and FDSA algorithms theoretically (using mean-square errors computed from asymptotic distributions) and numerically. Based on the theoretical and numerical results, SPSA is the preferable algorithm to use.Keywords
This publication has 16 references indexed in Scilit:
- Multivariate stochastic approximation using a simultaneous perturbation gradient approximationIEEE Transactions on Automatic Control, 1992
- Perturbation Analysis of Discrete Event Dynamic Systems1991
- Sensitivity analysis via likelihood ratios1986
- Stochastic approximation method with gradient averaging for unconstrained problemsIEEE Transactions on Automatic Control, 1983
- Stochastic Approximation Methods for Constrained and Unconstrained Systems1978
- STOCHASTIC APPROXIMATION1971
- On Asymptotic Normality in Stochastic ApproximationThe Annals of Mathematical Statistics, 1968
- Asymptotic Distribution of Stochastic Approximation ProceduresThe Annals of Mathematical Statistics, 1958
- On a Class of Stochastic Approximation ProcessesThe Annals of Mathematical Statistics, 1956
- A Stochastic Approximation MethodThe Annals of Mathematical Statistics, 1951