Geodesic
Parallel algorithms and probability of large deviation for stochastic convex optimization problems
Sibirskij žurnal vyčislitelʹnoj matematiki, Tome 21 (2018) no. 1, pp. 47-53.

Voir la notice de l'article provenant de la source Math-Net.Ru

In this paper, convex stochastic optimization problems under different assumptions on the properties of the available stochastic subgradients are considered. It is known that if a value of the objective function is available, one can obtain, in parallel, several independent approximate solutions in terms of the objective residual expectation. Then, choosing a solution with the minimum function value, one can control the probability of large deviations of the objective residual. On the contrary, in this short paper we address the situation when the objective function value is unavailable or is too expensive to calculate. Under the “light-tail” assumption for stochastic subgradients and in the general case with a moderate probability of large deviations, it is shown that parallelization combined with averaging gives bounds for the probability of large deviations similar to those of a serial method. Thus, in these cases one can benefit from parallel computations and reduce the computational time without any loss in the solution quality. 
@article{SJVM_2018_21_1_a2,
     author = {P. Dvurechensky and A. Gasnikov and A. Lagunovskaya},
     title = {Parallel algorithms and probability of large deviation for stochastic convex optimization problems},
     journal = {Sibirskij \v{z}urnal vy\v{c}islitelʹnoj matematiki},
     pages = {47--53},
     publisher = {mathdoc},
     volume = {21},
     number = {1},
     year = {2018},
     language = {ru},
     url = {http://geodesic.mathdoc.fr/item/SJVM_2018_21_1_a2/}
}
TY  - JOUR
AU  - P. Dvurechensky
AU  - A. Gasnikov
AU  - A. Lagunovskaya
TI  - Parallel algorithms and probability of large deviation for stochastic convex optimization problems
JO  - Sibirskij žurnal vyčislitelʹnoj matematiki
PY  - 2018
SP  - 47
EP  - 53
VL  - 21
IS  - 1
PB  - mathdoc
UR  - http://geodesic.mathdoc.fr/item/SJVM_2018_21_1_a2/
LA  - ru
ID  - SJVM_2018_21_1_a2
ER  - 
%0 Journal Article
%A P. Dvurechensky
%A A. Gasnikov
%A A. Lagunovskaya
%T Parallel algorithms and probability of large deviation for stochastic convex optimization problems
%J Sibirskij žurnal vyčislitelʹnoj matematiki
%D 2018
%P 47-53
%V 21
%N 1
%I mathdoc
%U http://geodesic.mathdoc.fr/item/SJVM_2018_21_1_a2/
%G ru
%F SJVM_2018_21_1_a2
P. Dvurechensky; A. Gasnikov; A. Lagunovskaya. Parallel algorithms and probability of large deviation for stochastic convex optimization problems. Sibirskij žurnal vyčislitelʹnoj matematiki, Tome 21 (2018) no. 1, pp. 47-53. http://geodesic.mathdoc.fr/item/SJVM_2018_21_1_a2/

[1] Shapiro A., Dentcheva D., Ruszczynski A., Lectures on Stochastic Programming: Modeling and Theory, MPS–SIAM Series on Optimization, 2014 | MR

[2] Nemirovski A., Juditsky A., Lan G., Shapiro A., “Robust stochastic approximation approach to stochastic programming”, SIAM J. Optim., 19 (2009), 1574–1609 | DOI | MR

[3] Ben-Tal A., Nemirovski A., Lectures on Modern Convex Optimization, , 2013 http://www2.isye.gatech.edu/~nemirovs/Lect_ModConvOpt.pdf

[4] Durrett R., Probability: Theory and Examples, Cambridge University Press, 2010 | MR

[5] Guigues V., Juditsky A., Nemirovski A., “Non-asymptotic confidence bounds for the optimal value of a stochastic program”, Optimization Methods and Software, 32:5 (2017), 1033–1058 | DOI | MR

[6] Gasnikov A. V., Effektivnye chislennye metody poiska ravnovesii v bolshikh transportnykh setyakh, Dis. $\dots$ doktora fiz.-mat.nauk: 05.13.18, Moskva, 2016

[7] Nesterov Yu., Vial J.-Ph., “Confidence level solutions for stochastic programming”, Automatica, 44:6 (2008), 1559–1568 | DOI | MR