Generating and handling scenarios in stochastic programming<br/><br/>

Monday, October 18, 2010 - 3:00pm - 4:00pm
Keller 3-180
Werner Römisch (Humboldt-Universität)
First, three approaches to scenario generation besides
Monte Carlo methods are considered: (i) Optimal quantization
of probability distributions, (ii) Quasi-Monte Carlo
methods and (iii) Quadrature rules based on sparse
grids. The available theory is discussed and related
to applying them in stochastic programming. Second,
the problem of optimal scenario reduction and the
generation of scenario trees for multistage models
are addressed.
MSC Code: