Stochastic Optimization Problems and Dependent Data

It is well-known that empirical estimates are usually employed when it is necessary to solve a stochastic decision problem depending on a completely unknown probability measure. These estimates are based on replacing the theoretical measure by an empirical one. In the stochastic programming theory a great attention has been already paid to them. However, these results are mostly known for underlying independent identically distributed random samples. The aim of this paper is to recall and summarize some rather new results achieved for dependent data that correspond rather often to economic activities. In particular we focus on one-stage and two-stage problems in the case of weakly dependent random sequences and on the multistage problems in the special type of the Markov dependence.