A branch-and-bound method for multistage stochastic integer programs with risk objectives

Abstract

We identify multistage stochastic integer programs with risk objectives where the related wait-and-see problems enjoy similiar separability as in the risk neutral case. For models belonging to this class, we present a solution method combining branch-and-bound with relaxation of non-anticipativity and constraint branching along non-anticipativity subspaces.

Keywords:

• stochastic integer programming
• multistage models
• mean-risk models
• mixed-integer optimization

AMS Subject Classification:

• 90C15
• 90C11
• 90C06

Acknowledgements

We would like to thank an anonymous referee for suggestions that have improved the final version of this paper. This research has been supported by the DFG-Schwerpunktprogramm 1126 ‘Algorithmik großer und komplexer Netzwerke’.