Adaptive stochastic approximation by the simultaneous perturbation method
- 28 November 2002
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 4 (01912216) , 3872-3879
- https://doi.org/10.1109/cdc.1998.761833
Abstract
Stochastic approximation (SA) has long been applied for problems of minimizing loss functions or root-finding with noisy input information. As with all stochastic search algorithms, there are adjustable algorithm coefficients that must be specified and that can have a profound effect on algorithm performance. It is known that picking these coefficients according to an SA analogue of the deterministic Newton-Raphson algorithm provides an optimal or near-optimal form of the algorithm. This paper presents a general adaptive SA algorithm that is based on an easy method for estimating the Hessian matrix at each iteration while concurrently estimating the primary parameters of interest. The approach applies in both the gradient-free optimization (Kiefer-Wolfowitz) and root-finding/stochastic gradient-based (Robbins-Monro) settings and is based on the "simultaneous perturbation" idea introduced previously.Keywords
This publication has 14 references indexed in Scilit:
- Accelerated second-order stochastic optimization using only function measurementsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- Model-free control of nonlinear stochastic systems with discrete-time measurementsIEEE Transactions on Automatic Control, 1998
- Implementation of the simultaneous perturbation algorithm for stochastic optimizationIEEE Transactions on Aerospace and Electronic Systems, 1998
- Weighted Means in Stochastic Approximation of MinimaSIAM Journal on Control and Optimization, 1997
- Stochastic Approximation Algorithms and ApplicationsPublished by Springer Nature ,1997
- Acceleration of Stochastic Approximation by AveragingSIAM Journal on Control and Optimization, 1992
- Averaging procedures in adaptive filtering: an efficient approachIEEE Transactions on Automatic Control, 1992
- Multivariate stochastic approximation using a simultaneous perturbation gradient approximationIEEE Transactions on Automatic Control, 1992
- Multivariate Adaptive Stochastic ApproximationThe Annals of Statistics, 1987
- An Extension of the Robbins-Monro ProcedureThe Annals of Mathematical Statistics, 1967