论文信息 - Stochastic approximation method with gradient averaging for unconstrained problems

Stochastic approximation method with gradient averaging for unconstrained problems

A stochastic approximation algorithm is studied, in which random gradient estimates are averaged by an auxiliary, filter. Convergence of the algorithm is proved under a rather general noise condition. A practical version of the method is described and numerical results are given.

A. Ruszczynski | W. Syski

[1] Lennart Ljung,et al. Analysis of a general recursive prediction error identification algorithm , 1981, Autom..

[2] H. Kushner,et al. Stochastic approximation type methods for constrained systems: Algorithms and numerical results , 1974 .

[3] Andrzej Ruszczynski,et al. Feasible direction methods for stochastic programming problems , 1980, Math. Program..

[4] H. Kushner,et al. Numerical studies of stochastic approximation procedures for constrained problems , 1977 .

[5] E. Fogel. A fundamental approach to the convergence analysis of least squares algorithms , 1980 .

[6] Y. Tsypkin,et al. On convergence of the recursive identification algorithms , 1981 .

[7] A. M. Gupal,et al. Stochastic analog of the conjugant-gradient method , 1972 .

[8] F. Clarke. Generalized gradients and applications , 1975 .

[9] H. Kushner,et al. Asymptotic Properties of Stochastic Approximations with Constant Coefficients. , 1981 .

[10] Lennart Ljung,et al. Analysis of recursive stochastic algorithms , 1977 .

[11] V. Fabian. STOCHASTIC APPROXIMATION METHODS , 1960 .

[12] H. Kesten. Accelerated Stochastic Approximation , 1958 .