Rate of convergence analysis of simultaneous perturbation stochastic approximation algorithm for time-varying loss function

Author

Qi Wang ; Ming Ye

Author_Institution

Dept. of Appl. Math. & Stat., Johns Hopkins Univ., Baltimore, MD, USA

fYear

2014

fDate

4-6 June 2014

Firstpage

5192

Lastpage

5197

Abstract

A popular method for continuous stochastic optimization problem is simultaneous perturbation stochastic approximation (SPSA). Spall (1992) introduces SPSA and discusses the rate of convergence of SPSA for fixed loss functions. In this paper, we use different criteria to discuss the rate of convergence of SPSA for time-varying loss functions. The rate of convergence result shows that SPSA is an effective algorithm for time-varying problems, such as the model-free adaptive control of nonlinear stochastic systems with unknown dynamics.

Keywords

adaptive control; convergence; nonlinear control systems; perturbation techniques; stochastic processes; stochastic systems; SPSA; continuous stochastic optimization problem; convergence analysis; convergence rate; fixed loss function; model-free adaptive control; nonlinear stochastic systems; simultaneous perturbation stochastic approximation algorithm; time-varying loss function; time-varying problem; unknown dynamics; Optimization; Stochastic systems; Time-varying systems;

fLanguage

English

Publisher

ieee

Conference_Titel

American Control Conference (ACC), 2014

Conference_Location

Portland, OR

ISSN

0743-1619

Print_ISBN

978-1-4799-3272-6

Type

conf

DOI

10.1109/ACC.2014.6859379

Filename

6859379