Advertisement

Science in China Series F: Information Sciences

, Volume 51, Issue 9, pp 1269–1280 | Cite as

Performance analysis of stochastic gradient algorithms under weak conditions

  • Feng Ding
  • HuiZhong Yang
  • Fei Liu
Article

Abstract

By using the stochastic martingale theory, convergence properties of stochastic gradient (SG) identification algorithms are studied under weak conditions. The analysis indicates that the parameter estimates by the SG algorithms consistently converge to the true parameters, as long as the information vector is persistently exciting (i.e., the data product moment matrix has a bounded condition number) and that the process noises are zero mean and uncorrelated. These results remove the strict assumptions, made in existing references, that the noise variances and high-order moments exist, and the processes are stationary and ergodic and the strong persistent excitation condition holds. This contribution greatly relaxes the convergence conditions of stochastic gradient algorithms. The simulation results with bounded and unbounded noise variances confirm the convergence conclusions proposed.

Keywords

recursive identification parameter estimation least squares stochastic gradient multivariable systems convergence properties martingale convergence theorem 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Ljung L. Consistency of the least-squares identification method. IEEE Trans Automatic Contr, 1976, 21(5): 779–781MATHCrossRefMathSciNetGoogle Scholar
  2. 2.
    Solo V. The convergence of AML. IEEE Trans Automatic Contr, 1979, 24(6): 958–962MATHCrossRefGoogle Scholar
  3. 3.
    Lai T L, Wei C Z. Least squares estimates in stochastic regression models with applications to identification and control of dynamic systems. Ann Stat, 1982, 10(1): 154–166CrossRefMathSciNetGoogle Scholar
  4. 4.
    Lai T L, Wei C Z. Extended least squares and their applications to adaptive control and prediction in linear systems. IEEE Trans Automatic Contr, 1986, 31(10): 898–906MATHCrossRefMathSciNetGoogle Scholar
  5. 5.
    Wei C Z. Adaptive prediction by least squares prediction in stochastic regression models. Ann Stat, 1987, 15(4): 1667–1682MATHCrossRefGoogle Scholar
  6. 6.
    Lai T L, Ying Z L. Recursive identification and adaptive prediction in linear stochastic systems. SIAM J Contr Optim, 1991, 29(5): 1061–1090MATHCrossRefMathSciNetGoogle Scholar
  7. 7.
    Toussi K, Ren W. On the convergence least squares estimates in white noise. IEEE Trans Automatic Contr, 1994, 39(2): 364–368MATHCrossRefGoogle Scholar
  8. 8.
    Ren W, Kumar P R. Stochastic adaptive prediction and model reference control. IEEE Trans Automatic Contr, 1994, 39(10): 2047–2060MATHCrossRefMathSciNetGoogle Scholar
  9. 9.
    Chen H F, Guo L. Identification and Stochastic Adaptive Control. Boston, MA: Birkhäuser, 1991MATHCrossRefGoogle Scholar
  10. 10.
    Ding F, Yang J B, Ding T. Performance analysis of least mean square algorithm for time-varying systems. Contr Theory Appl, 2001, 18(3): 433–437MATHMathSciNetGoogle Scholar
  11. 11.
    Ding F, Ding T, Xiao D Y, et al. Bounded convergence of finite data window least squares identification for time-varying systems. Acta Automat Sin, 2002, 28(5): 754–761MathSciNetGoogle Scholar
  12. 12.
    Ding F, Ding T, Yang J B, et al. Convergence of forgetting gradient estimation algorithm for time-varying parameters. Acta Automat Sin, 2002, 28(6): 962–968MathSciNetGoogle Scholar
  13. 13.
    Ding F, Chen T. Least squares based self-tuning control of dual-rate systems. Int J Adapt Contr Sig Proc, 2004, 18(8): 697–714MATHCrossRefMathSciNetGoogle Scholar
  14. 14.
    Ding F, Chen T. Combined parameter and output estimation of dual-rate systems using an auxiliary model. Automatica, 2004, 40(10): 1739–1748MATHCrossRefMathSciNetGoogle Scholar
  15. 15.
    Ding F, Chen T. Hierarchical least squares identification methods for multivariable systems. IEEE Trans Automat Contr, 2005, 50(3): 397–402CrossRefMathSciNetGoogle Scholar
  16. 16.
    Ding F, Chen T. Hierarchical gradient-based identification of multivariable discrete-time systems. Automatica, 2005, 41(2): 315–325MATHCrossRefMathSciNetGoogle Scholar
  17. 17.
    Ding F, Chen T. Performance bounds of forgetting factor least squares algorithm for time-varying systems with finite measurement data. IEEE Trans Circ Syst-I: Regular papers, 2005, 52(3): 555–566CrossRefMathSciNetGoogle Scholar
  18. 18.
    Ding F, Chen T. Identification of Hammerstein nonlinear ARMAX systems. Automatica, 2005, 41(9): 1479–1489MATHCrossRefMathSciNetGoogle Scholar
  19. 19.
    Ding F, Chen T. Modeling and identification for multirate systems. Acta Automat Sin, 2005, 31(1): 105–122MathSciNetGoogle Scholar
  20. 20.
    Ding F, Chen T. Performance analysis of multi-innovation gradient type identification methods. Automatica, 2007, 43(1): 1–14MATHCrossRefMathSciNetGoogle Scholar
  21. 21.
    Xie X M, Ding F. Adaptive Control Systems. Beijing: Tsinghua University Press, 2002Google Scholar
  22. 22.
    Goodwin G C, Sin K S. Adaptive Filtering, Prediction and Control. Englewood Cliffs, NJ: Prentice-Hall, 1984MATHGoogle Scholar
  23. 23.
    Ding F, Chen T. Parameter estimation of dual-rate stochastic systems by using an output error method. IEEE Trans Automat Contr, 2005, 50(9): 1436–1441CrossRefMathSciNetGoogle Scholar
  24. 24.
    Ding F, Chen T. Adaptive digital control of Hammerstein nonlinear systems with limited output sampling. SIAM J Contr Optim, 2007, 45(6): 2257–2276MATHCrossRefMathSciNetGoogle Scholar
  25. 25.
    Ding F, Shi Y, Chen T. Auxiliary model based least-squares identification methods for Hammerstein output-error systems. Syst Contr Lett, 2007, 5(56): 373–380CrossRefMathSciNetGoogle Scholar

Copyright information

© Science in China Press and Springer-Verlag GmbH 2008

Authors and Affiliations

  1. 1.Control Science and Engineering Research CenterJiangnan UniversityWuxiChina

Personalised recommendations