Adaptive initial step size selection for simultaneous perturbation stochastic approximation
- Author
- Keiichi Ito (UGent) and Tom Dhaene (UGent)
- Organization
- Abstract
- A difficulty in using Simultaneous Perturbation Stochastics Approximation (SPSA) is its performance sensitivity to the step sizes chosen at the initial stage of the iteration. If the step size is too large, the solution estimate may fail to converge. The proposed adaptive stepping method automatically reduces the initial step size of the SPSA so that reduction of the objective function value occurs more reliably. Ten mathematical functions each with three different noise levels were used to empirically show the effectiveness of the proposed idea. A parameter estimation example of a nonlinear dynamical system is also included.
- Keywords
- OPTIMIZATION, GRADIENT APPROXIMATION, ALGORITHM, SPSA, Stochastic approximation, Optimization, Direct method, Noisy function, Parameter estimation, IBCN
Downloads
-
6555.pdf
- full text
- |
- open access
- |
- |
- 1.91 MB
Citation
Please use this url to cite or link to this publication: http://hdl.handle.net/1854/LU-7240462
- MLA
- Ito, Keiichi, and Tom Dhaene. “Adaptive Initial Step Size Selection for Simultaneous Perturbation Stochastic Approximation.” SPRINGERPLUS, vol. 5, 2016, doi:10.1186/s40064-016-1823-3.
- APA
- Ito, K., & Dhaene, T. (2016). Adaptive initial step size selection for simultaneous perturbation stochastic approximation. SPRINGERPLUS, 5. https://doi.org/10.1186/s40064-016-1823-3
- Chicago author-date
- Ito, Keiichi, and Tom Dhaene. 2016. “Adaptive Initial Step Size Selection for Simultaneous Perturbation Stochastic Approximation.” SPRINGERPLUS 5. https://doi.org/10.1186/s40064-016-1823-3.
- Chicago author-date (all authors)
- Ito, Keiichi, and Tom Dhaene. 2016. “Adaptive Initial Step Size Selection for Simultaneous Perturbation Stochastic Approximation.” SPRINGERPLUS 5. doi:10.1186/s40064-016-1823-3.
- Vancouver
- 1.Ito K, Dhaene T. Adaptive initial step size selection for simultaneous perturbation stochastic approximation. SPRINGERPLUS. 2016;5.
- IEEE
- [1]K. Ito and T. Dhaene, “Adaptive initial step size selection for simultaneous perturbation stochastic approximation,” SPRINGERPLUS, vol. 5, 2016.
@article{7240462,
abstract = {{A difficulty in using Simultaneous Perturbation Stochastics Approximation (SPSA) is its performance sensitivity to the step sizes chosen at the initial stage of the iteration. If the step size is too large, the solution estimate may fail to converge. The proposed adaptive stepping method automatically reduces the initial step size of the SPSA so that reduction of the objective function value occurs more reliably. Ten mathematical functions each with three different noise levels were used to empirically show the effectiveness of the proposed idea. A parameter estimation example of a nonlinear dynamical system is also included.}},
articleno = {{200}},
author = {{Ito, Keiichi and Dhaene, Tom}},
issn = {{2193-1801}},
journal = {{SPRINGERPLUS}},
keywords = {{OPTIMIZATION,GRADIENT APPROXIMATION,ALGORITHM,SPSA,Stochastic approximation,Optimization,Direct method,Noisy function,Parameter estimation,IBCN}},
language = {{eng}},
title = {{Adaptive initial step size selection for simultaneous perturbation stochastic approximation}},
url = {{http://doi.org/10.1186/s40064-016-1823-3}},
volume = {{5}},
year = {{2016}},
}
- Altmetric
- View in Altmetric
- Web of Science
- Times cited: