On the formulation of stochastic linear programs using algebraic modelling languages

This paper considers extensions to algebraic modelling languages to support formulation, instantiation and solver integration for stochastic linear programs (SLPs). The authors present a taxonomy of SLP problem types and analyze formulation requirements including distribution handling by class of problem. They demonstrate suggested formulations for most problem classes, show solver input in the S-MPS standard, and propose consistency checks for constraints involving stochastic data items. Some unresolved difficulties are identified.