Abstract:
|
Building on the proliferation of data, on the advances in statistical methods, and on the growth of computing power, the opportunities for optimization under uncertainty are tremendous. We consider a general approach in which one can move from data to stochastic models of uncertainty based on statistical methods and then to stochastic optimization and control based on these stochastic models of uncertainty. Various examples will be used to describe instances of our general approach, where each is intended to illustrate different types and sources of data and different types of stochastic models and optimization formulations, as well as the significant benefits of combining these various elements.
|