Abstract
A stochastic quasigradient algorithm is suggested for solving the quantile optimization problem with a convex loss function. The algorithm is based on stochastic finite-difference approximations of gradients of the quantile function by using the order statistics. The algorithm convergence almost surely is proved.
Similar content being viewed by others
References
Bahadur, R. R. (1966). A note on quantile in large samples. Annals of Mathematical Statistics, 37, 577–580.
Borell, C. (1975). Convex set functions in d-spaces. Periodica Mathematica Hungarica, 6(2), 111–136.
Brascamp, H. J., & Lieb, E. H. (1976). On extensions of the Brunn-Minkowski and Prekopa-Liendler theorems, including inequalities for log-concave functions, and with application to the diffusion equations. Journal of Functional Analysis, 22, 366–389.
Das Gupta, S. (1980). Brunn-Minkowski inequality and its aftermath. Journal of Multivariate Analysis, 10, 296–318.
David, H. A., & Nagaraja, H. N. (2003). Order statistics, 3rd edn. New Jersey: Willey.
Deák, I. (2003). Solving stochastic programming problems by successive regression approximations—numerical results. In K. Marti, Y. Ermoliev, & G. Pug (Eds.), Lecture notes in economics and mathematical systems: Vol. 532. Dynamic stochastic optimization (pp. 209–224). Berlin: Springer.
Dentcheva, D. (2009). Optimization models with probabilistic constraints. In A. Shapiro, D. Dentcheva, & A. Ruszczynski (Eds.), Lectures on stochastic programming: modeling and theory (pp. 87–155). Philadelphia: SIAM.
Ermoliev, Y. M. (1976). Stochastic programming methods. Moscow: Nauka [in Russian].
Gardner, R. J. (2002). The Brunn-Minkowski inequality. Bulletin (New Series) of the American Mathematical Society, 39(3), 355–405.
Kan, Y. S. (1996). A quasi-gradient algorithm for minimizing the quantile function. Journal of Computer & Systems Sciences International, 35(2), 245–251.
Kan, Y. S. (2003). On convergence of a stochastic quasigradient algorithm of quantile optimization. Automation and Remote Control, 64(2), 263–278.
Kan, Y. S., & Kibzun, A. I. (1996). Convex properties of probability and quantile functions in optimization problems. Automation and Remote Control, 57(3), 368–383.
Kibzun, A. I., & Kan, Y. S. (1996). Stochastic programming problems with probability and quantile functions. Chichester: Wiley.
Kibzun, A. I., & Kurbakovskiy, V. Y. (1991). Guaranteeing approach to solving quantile optimization problems. Annals of Operations Research, 30, 81–93.
Kibzun, A. I., & Matveev, E. L. (2010). Sufficient conditions for quasiconcavity of the probability function. Automation and Remote Control, 71(3), 413–430.
Lepp, R. (1983). Stochastic approximation type algorithm for the maximization of the probability functions. Eesti NSV Teaduste Akadeemia Toimetised, Füüsika * Matemaatika, 32(2), 150–156.
Lusternik, L. (1935). The Brunn-Minkowski inequality for Lebesgue-measurable sets. Doklady Akademii Nauk SSSR, 3(2), 55–58 [in Russian].
Malyshev, V. V., & Kibzun, A. I. (1987). Analysis and synthesis of high precision aircraft control. Moscow: Mashinostroenie [in Russian].
Mihalevich, V. S., Gupal, A. M., & Norkin, V. I. (1987). Optimization methods of non-convex functions. Moscow: Nauka [in Russian].
Norkin, V. I., & Roenko, N. V. (1991). α-Concave functions and measures and their applications. Kibernetika I Sistemnyj Analiz, 6, 77–88 [in Russian].
Prékopa, A. (1973). On logarithmic concave measures and functions. Acta Scientiarum Mathematicarum, 34, 335–343.
Prékopa, A. (1995). Stochastic programming. Dordrecht: Kluwer Academic.
Prékopa, A. (2003). Probabilistic programming. In A. Ruszczynsk & A. Shapiro (Eds.), Handbooks oper. res. management sci.: Vol. 10. Stochastic programming (pp. 267–351). New York: Elsevier.
Reiss, R. D. (1989). Approximate distributions of order statistics. New York: Springer.
Rinott, Y. (1976). On convexity of measures. Annals of Probability, 6, 1020–1026.
Shiryaev, A. N. (1980). Probability. Moscow: Nauka [in Russian].
Tretiakov, G. (2002). Stochastic quasi-gradient algorithms for maximization of the probability function. A new formula for the gradient of the probability function. In Lecture notes in economics and mathematical systems: Vol. 513. Stochastic optimization techniques. Numerical methods and technical applications (pp. 117–139).
Zhigljavsky, A., & Zhilinskas, A. (2008). Stochastic global optimization. Berlin: Springer.
Author information
Authors and Affiliations
Corresponding author
Additional information
The work was supported by Russian Foundation of Basic Research (project N0 09-08-00369).
Rights and permissions
About this article
Cite this article
Kibzun, A., Matveev, E. Optimization of the quantile criterion for the convex loss function by a stochastic quasigradient algorithm. Ann Oper Res 200, 183–198 (2012). https://doi.org/10.1007/s10479-011-0987-z
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10479-011-0987-z