UDC 519.21
1 V.M. Glushkov Institute of Cybernetics, National Academy of Sciences of Ukraine, Kyiv, Ukraine
knopov1@yahoo.com
|
2 V.M. Glushkov Institute of Cybernetics, National Academy of Sciences of Ukraine, Kyiv, Ukraine
e.kasitskaya@gmail.com
|
ON LARGE DEVIATIONS OF EMPIRICAL ESTIMATES IN A STOCHASTIC PROGRAMMING
PROBLEM WITH NONSTATIONARY OBSERVATIONS AND CONTINUOUS TIME
Abstract. The paper considers a stochastic programming problem with the empirical function constructed on nonstationary observations and continuous time. A stationary in a strict sense random process satisfying the strong mixing condition is investigated in the problem. The conditions under which the empirical estimate is consistent are given and large deviations of the estimate are considered.
Keywords: stochastic programming problem, stationary ergodic random process, strong mixing condition, large deviations.
FULL TEXT
REFERENCES
- Knopov PS, Kasitskaya EI On large deviations of empirical estimates in the problem of stochastic programming for non-stationary observations. Kibernetika i sistemnyj analiz. 2010. Т. 46, № 5. С. 46–50.
- Deuscel J.-D., Stroock D.W. Large deviations. Boston, etc.: Academ. Press, Inc., 1989. 310 p.
- Dunford N., Schwartz J. Linear operators, Part I: General theory. New York: Interscience, 1957. 896 p.
- Knopov P., Kasitskaya E. Large deviations for the method of empirical means in stochastic optimization problems with continuous time observations. Optimization Methods and Applications. In Honor of the 80-th Birthday of Ivan V. Sergienko. Butenko S., Pardalos P.M., and Shylo V. (Eds.). New York: Springer, 2017. P. 263–275.