Stochastic programming with random processes (Q1178433)

The author considers the following chance-constrained program \[ \min c_ T(x;\omega), \text{s.t. }P\{g_ i(x)\geq b_ i(t;\omega), t=1,\dots,T\}\geq\alpha_ i\quad (i=1,\dots,m), x\in X\tag{1} \] in which one requires particular constraints to be fulfilled with given probabilities for more time periods without interruption (here for \(T\) time periods). The random variables \(b_ i(1),\dots,b_ i(T)\) can be dependent, forming thus a general random process. The author uses the asymptotic approach provided by the extremal theory of random processes for solving the stochastic program (1). Two further possible approaches to stochastic programming problems defined in time are also described: a) an exponential penalty model approach related to scenario analysis; b) a modification of the entropic penalty approach.