Annals of Operations Research

, Volume 39, Issue 1, pp 251–267 | Cite as

A stochastic quasigradient algorithm with variable metric

  • S. P. Uryas'ev


This paper deals with a new variable metric algorithm for stochastic optimization problems. The essence of this is as follows: there exist two stochastic quasigradient algorithms working simultaneously — the first in the main space, the second with respect to the matrices that modify the space variables. Almost sure convergence of the algorithm is proved for the case of the convex (possiblynonsmooth) objective function.


Objective Function Space Variable Stochastic Optimization Stochastic Optimization Problem Main Space 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    B. Betro and L. De Biase, A Newton-like method for stochastic optimization, in:Towards Global Optimization 2 (North-Holland, 1978), pp. 269–289.Google Scholar
  2. [2]
    J.N. Dennis and J.J. Moré, Quasi-Newton methods, motivation and theory, SIAM Rev. 19(1977)46–89.Google Scholar
  3. [3]
    Yu.M. Ermoliev, Stochastic quasi-gradient methods and their application to systems optimization, Stochastics 4(1983)1–37.Google Scholar
  4. [4]
    A. Gaivoronski, Implementation of stochastic quasigradient methods, in:Numerical Techniques for Stochastic Optimization, ed. Yu.M. Ermoliev and R.J.-B. Wets (Springer, 1988), pp. 313–353.Google Scholar
  5. [5]
    L. Gerencser, Strong consistency theorems related to stochastic quasi-Newton methods, in:Stochastic Optimization, Lecture Notes in Control and Information Science 81 (Springer, 1984).Google Scholar
  6. [6]
    Yu.M. Kaniovski, P.S. Knopov and Z.V. Nekrylova,Limit Theorems for Stochastic Programming Processes (Naukova Dumka, Kiev, 1980), in Russian.Google Scholar
  7. [7]
    H.J. Kushner and G. Jin, Stochastic approximation algorithms for parallel and distributed processing, Stochastics 22(1987)219–250.Google Scholar
  8. [8]
    L. Ljung and T. Söderström,Theory and Practice of Recursive Identification (MIT Press, 1983).Google Scholar
  9. [9]
    K. Marti, Descent stochastic quasigradient methods, in:Numerical Techniques for Stochastic Optimization, ed. Yu.M. Ermoliev and R.J.-B. Wets (Springer, 1988), pp. 393–401.Google Scholar
  10. [10]
    N.M. Novikova, Some stochastic programming methods in Hilbert Space, Zh. Vichislitelnoj Mat. i. Mat. Phis. 25(1985)1795–1813, in Russian [USSR Comput. Math. and Math. Phys. (in English)].Google Scholar
  11. [11]
    E.A. Nurminski,Numerical Methods for Solving Deterministic and Stochastic Minimax Problems (Naukova Dumka, Kiev, 1979), in Russian.Google Scholar
  12. [12]
    G. Pflug, Stepsize rules, stopping times and their implementation in stochastic quasigradient algorithms, in:Numerical Techniques for Stochastic Optimization, ed. Yu.M. Ermoliev and R.J.-B. Wets (Springer, 1988), pp. 353–373.Google Scholar
  13. [13]
    R.T. Rockafellar and R.J.-B. Wets, On the interchange of subdifferentiation and conditional expectation for convex functionals, Stochastics 7(1982)173–182.Google Scholar
  14. [14]
    A. Ruszczynski and W. Syski, A method of aggregate stochastic sub-gradients with on-line stepsize rules for convex stochastic programming problems, Math. Progr. Study 28(1985)113–131.Google Scholar
  15. [15]
    G.S. Saridis, Learning applied to successive approximation algorithms, IEEE Trans. Syst. Sci. Cybern. SSC-6(1970)97–103.Google Scholar
  16. [16]
    D.J. Sarkison, The use of stochastic approximation to solve the system identification problem, IEEE Trans. Auto. Control AC-12(1967)563–567.Google Scholar
  17. [17]
    S.P. Uryas'ev, Stochastic quasi-gradient algorithms with adaptively controlled parameters, Working Paper WP-86-32, IIASA (1986).Google Scholar
  18. [18]
    S.P. Uryas'ev,Adaptive Algorithms of Stochastic Optimization and Game Theory (Nauka, Moskow, 1990), in Russian.Google Scholar
  19. [19]
    S.P. Uryas'ev, Adaptive variable metric algorithms for nondifferentiable optimization problems, in:Analysis and Optimization of Systems, ed. A. Bensoussan and J.L. Lions, Proc. 9th Int. Conf., Antibes (Springer, 1990), pp. 432–441.Google Scholar
  20. [20]
    R.J.-B. Wets, Modeling and solution strategies for constrained stochastic optimization problems, Ann. Oper. Res. 1(1984)3–22.Google Scholar

Copyright information

© J.C. Baltzer AG, Scientific Publishing Company 1992

Authors and Affiliations

  • S. P. Uryas'ev
    • 1
  1. 1.International Institute for Applied Systems AnalysisLaxenburgAustria

Personalised recommendations