Formal Comparison of Simultaneous Perturbation Stochastic Approximation and Random Direction Stochastic Approximation

被引：2

作者：

Peng, Ducheng ^{[1
]}

Chen, Yiwen ^{[2
]}

Spall, James C. ^{[1
,3
]}

机构：

[1] Johns Hopkins Univ, Appl Math & Stat Dept, Whitehead Hall,3400 North Charles St, Baltimore, MD 21218 USA

[2] Johns Hopkins Univ, Appl Math & Stat Dept, Baltimore, MD USA

[3] JHU, Appl Phys Lab & Res, Baltimore, MD USA

来源：

2023 AMERICAN CONTROL CONFERENCE, ACC | 2023年

关键词：

OPTIMIZATION; ALGORITHMS; CONTROLLER; SYSTEMS;

D O I：

10.23919/ACC55779.2023.10156400

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

Stochastic approximation (SA) algorithms can be used in system optimization problems when only noisy measurements of a system are available. This paper formally compares the performance of two popular SA algorithms in a multivariate Kiefer-Wolfowitz setting of simultaneous-perturbation SA (SPSA) and the random-directions SA (RDSA). This paper provides sufficient conditions to demonstrate which algorithm has the smaller asymptotic mean squared error (MSE) and numerically presents comparison of SPSA and RDSA in a test function and a model-free control system. The theory and supporting numerics indicate that SPSA has better efficiency (lower MSE) across a broad range of problem settings.

引用

页码：744 / 749

页数：6

共 50 条

[21] OPTIMIZATION OF DYNAMIC RAMP METERING CONTROL WITH SIMULTANEOUS PERTURBATION STOCHASTIC APPROXIMATION [J].

Chien, S. I. ;

Luo, J. .

CONTROL AND INTELLIGENT SYSTEMS, 2008, 36 (01)

[22] Rate of Convergence Analysis of Discrete Simultaneous Perturbation Stochastic Approximation Algorithm [J].

Wang, Qi ;

Spall, James C. .

2013 AMERICAN CONTROL CONFERENCE (ACC), 2013, :4771-4776

[23] A robust recurrent simultaneous perturbation stochastic approximation training algorithm for recurrent neural networks [J].

Xu, Zhao ;

Song, Qing ;

Wang, Danwei .

NEURAL COMPUTING & APPLICATIONS, 2014, 24 (7-8) :1851-1866

[24] Random Directions Stochastic Approximation With Deterministic Perturbations [J].

Prashanth, L. A. ;

Bhatnagar, Shalabh ;

Bhavsar, Nirav ;

Fu, Michael ;

Marcus, Steven, I .

IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 2020, 65 (06) :2450-2465

[25] A method combining genetic algorithm with simultaneous perturbation stochastic approximation for linearly constrained stochastic optimization problems [J].

Zhang Huajun ;

Jin, Zhao ;

Hui, Luo .

JOURNAL OF COMBINATORIAL OPTIMIZATION, 2016, 31 (03) :979-995

[26] Tuning Missile Guidance and Control Algorithms Using Simultaneous Perturbation Stochastic Approximation [J].

Reardon, Brian E. ;

Lloyd, Justin M. ;

Perel, Ron Y. .

JOHNS HOPKINS APL TECHNICAL DIGEST, 2010, 29 (01) :85-100

[27] Tuning of Elasticsearch Configuration: Parameter Optimization Through Simultaneous Perturbation Stochastic Approximation [J].

Haugerud, Harek ;

Sobhie, Mohamad ;

Yazidi, Anis .

FRONTIERS IN BIG DATA, 2022, 5

[28] Constrained source seeking for mobile robots via simultaneous perturbation stochastic approximation [J].

Ramirez-Lianos, Eduardo ;

Martinez, Sonia .

2016 IEEE 55TH CONFERENCE ON DECISION AND CONTROL (CDC), 2016, :6851-6856

[29] Sparse Gaussian Mixture Model Clustering via Simultaneous Perturbation Stochastic Approximation [J].

Boiarov, Andrei ;

Granichin, Oleg .

IFAC PAPERSONLINE, 2020, 53 (02) :995-1000

[30] Controller Design by Using Simultaneous Perturbation Stochastic Approximation with Changeable Sliding Window [J].

Lu, Qing ;

Zhou, Jun .

INTELLIGENT ROBOTICS AND APPLICATIONS, ICIRA 2019, PT III, 2019, 11742 :665-676

← 1 2 3 4 5 →