Optimization under Exogenous and Endogenous Uncertainty

Customary stochastic programs aim at the best feasible decision made before the realization of the random element is observed. The common assumption is that the probability distribution does not depend on decisions - the case of the exogenous uncertainty. This paper focuses on stochastic programming models for which through decisions, a decision-dependent, endogenous randomness is put into effect. Problem structure then becomes important. Examples point out at tractable cases and solution techniques.