Improving computational efficiency in risk averse two-stage stochastic programs with fixed recourse for location problems

R. Gázquez, P. Díaz-Cachinero, R. Mínguez

A large class of practical problems can be formulated as two-stage stochastic programs. When uncertainty is represented by a finite set of scenarios—a common case—accounting for risk aversion, such as conditional value-at-risk, becomes crucial. This is particularly challenging when the number of scenarios is large, as is often the case in location problems, where decisions must be robust against demand or cost fluctuations. To improve computational efficiency, we propose a (clustering and) constraint generation algorithm. We demonstrate the effectiveness of our approach in a large-scale location and transportation problem, highlighting its potential for solving complex stochastic optimization models.

Keywords: Stochastic programming, location problems, linear programming, risk measures, cvar

Location (GELOCA2)
June 12, 2025  3:30 PM
MR 3

M. Merino, U. Aldasoro, I. Gago-Carro, D. Lee

M. Landete, M. Leal Palazón, L. Nácher, M. Labbé

L. Nácher, M. Landete, M. Leal Palazón, H. Yaman

H. I. Calvete, C. Galé, A. Hernández, J. A. Iranzo