01:30 PM - 03:10 PM
Robust Optimization in Energy and Environmental Problems
Uncertainty in environmental modeling and assessment is present everywhere and is recognized to be a key factor to be taken into consideration. Unfortunately, the implementation of classical methods such as stochastic programming or chance-constrained programming are quickly numerically intractable due to the large size of the models and to the difficulty of representing the uncertainty in reliable probabilistic terms. Robust optimization proposes an alternative approach that avoids some of these limitations, mainly because it is based on a non-probabilistic description of the uncertainty and because its primarily concern is to lead to optimization models that are numerically tractable by modern convex optimization tools. In this tutorial we shall introduce the methodology on a simple energy planning problem with a control on pollutant emission. We shall show how the important feature of adaptability is efficiently treated via decision rules. We shall also evoke an application to the large bottom-up global energy-environment model TIAM, with focus on the security of European energy supply.