Probability Theory and Related Fields

, Volume 71, Issue 4, pp 517–552 | Cite as

Asymptotic behaviour of a class of stochastic approximation procedures

  • Erich Berger


LetB be a real Banach space. Considering stochastic approximation procedures of Robbins-Monro and Kiefer-Wolfowitz type, one is led to recursion formulas of the form
$$U_{k + 1} = U_k - k^{ - 1} \Gamma _k U_k - k^{ - \rho } V_k - k^{ - \eta } T_k$$
(ϱ and η) positive real numbers,Uk, Vk, TkB-valued random variables, Γ k random linear operatorsBB, where — in addition to some further assumptions —Γ k Γ almost surely andTkT, say, almost surely. Letting (Y k,n )k∈{1,...,n},n∈N be an array ofB-valued random variables such that the partial sums\(\sum\limits_{k = 1}^m {V_k }\) and\(\sum\limits_{k = 1}^m {Y_{k,n} }\) do not differ “too much” (uniformly inm∈{1,...,n}), we shall investigate the distance of (U m )m∈{1,...,n} and (U (n) m )m∈{1,...,n}, where
$$U_{k + 1}^{\left( n \right)} = U_k^{\left( n \right)} - k^{ - 1} \Gamma U_k^{\left( n \right)} - k^{ - \rho } Y_{k,n} - k^{ - \eta } T$$
fork∈{1,...,n−1}. The proofs are based on a deterministic argument that enables us to present a unified approach for proving central limit theorems, weak and almost sure invariance principles and variants of the bounded and the functional law of the iterated logarithm for the sequence (Un).


Real Number Stochastic Process Asymptotic Behaviour Probability Theory Limit Theorem 
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.
    Abdelhamid, S.N.: Transformation of observations in stochastic approximation. Ann. Statist.1, 1158–1174 (1973)Google Scholar
  2. 2.
    Berger, E.: A note on the invariance principle for stochastic approximation procedures in a Hilbert space. Unpublished manuscript (1979)Google Scholar
  3. 3.
    Billingsley, P.: Convergence of Probability Measures. New York: Wiley 1968Google Scholar
  4. 4.
    Birnbaum, Z.W., Marshall, A.W.: Some multivariate Chebyshev inequalities with extensions to continuous parameter processes. Ann. Math. Statist.32, 687–704 (1961)Google Scholar
  5. 5.
    Blum, J.R.: Approximation methods which converge with probability one. Ann. Math. Statist.25, 382–386 (1954)Google Scholar
  6. 6.
    Borodin, A.N.: A stochastic approximation procedure in the case of weakly dependent observations. Theory Probab. Appl.24 34–52 (1979)Google Scholar
  7. 7.
    Chung, K.L.: On a stochastic approximation method. Ann. Math. Statist.25, 463–483 (1954)Google Scholar
  8. 8.
    Daleckiî, Ju.L., Kreîn, M.G.: Stability of Solutions of Differential Equations in Banach Space. Translations of Mathematical Monographs43. Providence, Rhode Island: American Math. Society 1974Google Scholar
  9. 9.
    Dehling, H.: Limit theorems for sums of weakly dependent Banach space valued random variables. Z. Wahrscheinlichkeitstheor. Verw. Geb.63, 393–432 (1983)Google Scholar
  10. 10.
    Dudley, R.M.: Distances of probability measures and random variables. Ann. Math. Statist.39, 1563–1572 (1968)Google Scholar
  11. 11.
    Dudley, R.M., Philipp, W.: Invariance principles for sums of Banach space valued random elements and empirical processes. Z. Wahrscheinlichkeitstheor. Verw. Geb.62, 509–552 (1983)Google Scholar
  12. 12.
    Fabian, V.: On asymptotic normality in stochastic approximation. Ann. Math. Statist.39, 1327–1332 (1968)Google Scholar
  13. 13.
    Fabian, V.: Asymptotically efficient stochastic approximation; the RM case. Ann. Statist.1, 485–495 (1973)Google Scholar
  14. 14.
    Fabian, V.: On asymptotically efficient recursive estimation. Ann. Statist.6, 854–866 (1978)Google Scholar
  15. 15.
    Fabian, V.: Local asymptotic minimax properties of some recursive estimates.RM — 409. Department of Statistics and Probability. Michigan State University (1980)Google Scholar
  16. 16.
    Fernique, X.: Intégrabilité des vecteurs gaussiens. C.R. Acad. Sci. Paris Sér.A 270, A 1698–1699 (1970)Google Scholar
  17. 17.
    Gänssler, P., Stute, W.: Wahrscheinlichkeitstheorie. Berlin-Heidelberg-New York: Springer 1977Google Scholar
  18. 18.
    Gaposhkin, V.F., Krasulina, T.P.: On the law of the iterated logarithm in stochastic approximation processes. Theory Probab. Appl.19, 844–850 (1974)Google Scholar
  19. 19.
    Hall, P.G., Heyde, C.C.: On a unified approach to the law of the iterated logarithm for martingales. Bull. Aust. Math. Soc.14, 435–447 (1976)Google Scholar
  20. 20.
    Hall, P., Heyde, C.C.: Martingale Limit Theory and Its Applications. New York-San Francisco-London: Academic Press 1980Google Scholar
  21. 21.
    Heyde, C.C.: On martingale limit theory and strong convergence results for stochastic approximation procedures. Stochastic Processes Appl.2, 371–389 (1974)Google Scholar
  22. 22.
    Jakubowski, A.: On limit theorems for sums of dependent Hilbert space valued random variables. In: Mathematical Statistics and Probability Theory. Proc. 6th Int. Conf., Wisla/Pol. 1978. Lect. Notes Stat.2, 178–187 (1980)Google Scholar
  23. 23.
    Kersting, G.: Almost sure approximation of the Robbins-Monro process by sums of independent random variables. Ann. Probab.5, 954–965 (1977)Google Scholar
  24. 24.
    Komlós, J., Major, P., Tusnády, G.: An approximation of partial sums of independent RV's and the sample DF. I. Z. Wahrscheinlichkeitstheor. Verw. Geb.32, 111–131 (1975)Google Scholar
  25. 25.
    Komlós, J., Major, P., Tusnády, G.: An approximation of partial sums of independent RV's and the sample DF. II. Z. Wahrscheinlichkeitstheor. Verw. Geb.34, 33–58 (1976)Google Scholar
  26. 26.
    Kuelbs, J., LePage, R.: The law of the iterated logarithm for Brownian motion in a Banach space. Trans. Am. Math. Soc.185, 253–264 (1973)Google Scholar
  27. 27.
    Kuelbs, J., Philipp, W.: Almost sure invariance principles for partial sums of mixing B-valued random variables. Ann. Probab.8, 1003–1036 (1980)Google Scholar
  28. 28.
    Lai, T.L., Robbins, H.: Limit theorems for weighted sums and stochastic approximation processes. Proc. Nat. Acad. Sci. U.S.A.75, 1068–1070 (1978)Google Scholar
  29. 29.
    Lai, T.L., Robbins, H.: Adaptive design and stochastic approximation. Ann. Statist.7, 1196–1221 (1979)Google Scholar
  30. 30.
    Major, P., Révész, P.: A limit theorem for the Robbins-Monro approximation. Z. Wahrscheinlichkeitstheor. Verw. Geb.27, 79–86 (1973)Google Scholar
  31. 31.
    Major, P.: The approximation of partial sums of independent RV's Z. Wahrscheinlichkeitstheor. Verw. Geb.35, 213–220 (1976)Google Scholar
  32. 32.
    Mark, G.: Log-log-Invarianzprinzipien für Prozesse der stochastischen Approximation. Mitteilungen Math. Sem. Giessen, No. 153, 87 pp. (1982)Google Scholar
  33. 33.
    McLeish, D.L.: Invariance principles for dependent random variables. Z. Wahrscheinlichkeitstheor. Verw. Geb.32, 165–178 (1975)Google Scholar
  34. 34.
    McLeish, D.L.: Functional and random central limit theorems for the Robbins-Monro process. J. Appl. Probab.13, 148–154 (1976)Google Scholar
  35. 35.
    McLeish, D.L.: On the invariance principle for nonstationary mixingales. Ann. Probab.5, 616–621 (1977)Google Scholar
  36. 36.
    Métivier, M.: Semimartingales: a Course on Stochastic Processes. Berlin-New York: de Gruyter 1982Google Scholar
  37. 37.
    Morrow, G., Philipp, W.: An almost sure invariance principle for Hilbert space valued martingales. Trans. Am. Math. Soc.273, 231–251 (1982)Google Scholar
  38. 38.
    Nevel'son, M.B., Has'minskiî, R.Z.: Stochastic Approximation and Recursive Estimation. Translations of Mathematical Monographs47. Providence, Rhode Island: American Math. Society 1976Google Scholar
  39. 39.
    Nixdorf, R.: Stochastische Approximation in Hilberträumen durch endlichdimensionale Verfahren. Mitteilungen Math. Sem. Giessen, No. 154, 78 pp. (1982)Google Scholar
  40. 40.
    Nixdorf, R.: An invariance principle for a finite dimensional stochastic approximation method in a Hilbert space. J. Multivariate Analysis15, 252–260 (1984)Google Scholar
  41. 41.
    O'Reilly, N.E.: On weak convergence of empirical processes in sup-norm metrics. Ann. Probab.2, 642–651 (1974)Google Scholar
  42. 42.
    Pantel, M.: Adaptive Verfahren der stochastischen Approximation. Dissertation. Universität Essen — Gesamthochschule (1979)Google Scholar
  43. 43.
    Philipp, W., Stout, W.: Almost sure invariance principles for partial sums of weakly dependent random variables. Mem. Amer. Math. Soc. No. 161. Providence, Rhode Island: American Math. Society 1975Google Scholar
  44. 44.
    Philipp, W.: Almost sure invariance principles for sums ofB-valued random variables. In: Probability in Banach Spaces II. Proc. Conf. Oberwolfach 1978. Lect. Notes in Math.709, pp. 171–193, Berlin-Heidelberg-New York: Springer 1979Google Scholar
  45. 45.
    Pyke, R.: Applications of almost surely convergent constructions of weakly convergent processes. In: Probability and Information Theory. Proc. Int. Symp. McMaster Univ. 1968. Lect. Notes in Math.89, 187–200. Berlin-Heidelberg-New York: Springer 1969Google Scholar
  46. 46.
    Ruppert, D.: Almost sure approximations to the Robbins-Monro and Kiefer-Wolfowitz processes with dependent noise. Ann. Probab.10, 178–187 (1982)Google Scholar
  47. 47.
    Sacks, J.: Asymptotic distribution of stochastic approximation procedures. Ann. Math. Statist.29, 373–405 (1958)Google Scholar
  48. 48.
    Skorohod, A.V.: Limit theorems for stochastic processes. Theory Probab. Appl.1, 261–290 (1956)Google Scholar
  49. 49.
    Strassen, V.: An invariance principle for the law of the iterated logarithm. Z. Wahrscheinlichkeitstheor. Verw. Geb.3, 211–226 (1964)Google Scholar
  50. 50.
    Strassen, V.: Almost sure behavior of sums of independent random variables and martingales. Proc. Fifth Berkeley Sympos. Math. Statist. Probab.2, 315–344 (1965)Google Scholar
  51. 51.
    Venter, J.H.: An extension of the Robbins-Monro procedure. Ann. Math. Statist.38, 181–190 (1967)Google Scholar
  52. 52.
    Walk, H.: An invariance principle for the Robbins-Monro process in a Hilbert space. Z. Wahrscheinlichkeitstheor. Verw. Geb.39, 135–150 (1977)Google Scholar
  53. 53.
    Walk, H.: Sequential estimation of the solution of an integral equation in filtering theory. In: Stochastic Control Theory and Stochastic Differential Systems (eds.: M. Kohlmann, W. Vogel). Lect. Notes in Control and Information Sciences16, 598–605 (1979)Google Scholar
  54. 54.
    Walk, H.: Martingales and the Robbins-Monro procedure inD[0, 1].J. Multivariate Anal. 8, 430–452 (1978)Google Scholar
  55. 55.
    Walk, H.: A functional central limit theorem for martingales inC(K) and its application to sequential estimates. J. Reine Angew. Math.314, 117–135 (1980)Google Scholar
  56. 56.
    Watanabe, M.: A stochastic approximation from dependent observations. Z. Wahrscheinlichkeitstheor. Verw. Geb.62, 279–292 (1983)Google Scholar

Copyright information

© Springer-Verlag 1986

Authors and Affiliations

  • Erich Berger
    • 1
  1. 1.Institut für Mathematische StochastikUniversität GöttingenGöttingenGermany

Personalised recommendations