I would be very grateful if someone guide me in constructing an ellipsoidal uncertainty set for an uncertain right-hand side vector of demands, in the Robust Optimization(RO) paradigm.
Assuming that we have several instances (samples) of the vector as historical data.
I know there are papers on doing it via statistical methods and Risk measures, but Im looking for the most practical and straightforward one at the moment, Just to learn RO by doing it.
Do you think, for a rhs vector, box or ellipsoidal sets, yield different results?!