Supporting Functions for Stochastic Programs

Supporting functions for stochastic models

The stochastic Benders algorithm (see Using the Stochastic Benders Algorithm) is implemented in AIMMS as a combination of a system module that can be included into your model, and a number of supporting functions in the GMP::Stochastic namespace of the GMP library. The procedures and functions of the GMP::Stochastic namespace are listed in this table.

Table 44 : GMP::Stochastic functions and procedures

BendersFindFeasibilityReference(GMP, stage, scenario):math:toAllGeneratedMathematicalPrograms

BendersFindReference(GMP, stage, scenario):math:toAllGeneratedMathematicalPrograms

CreateBendersRootproblem(GMP[, name]):math:toAllGeneratedMathematicalPrograms

UpdateBendersSubproblem(GMP, solution)

AddBendersFeasibilityCut(GMP, solution, cutNo)

AddBendersOptimalityCut( GMP, solution, cutNo)

MergeSolution(GMP, solution1, solution2[, updObj])

GetRepresentativeScenario(GMP, stage, scenario):math:toAllStochasticScenarios

GetObjectiveBound(GMP, solution)

GetRelativeWeight(GMP, stage, scenario)

Overview of functionality

For a more detailed overview of the functionality offered by the functions in the GMP::Stochastic namespace, we refer to

  • Using the Stochastic Benders Algorithm for an outline of the stochastic Benders algorithm,

  • the system module containing the AIMMS implementation of the stochastic Benders algorithm, and

  • the AIMMS Function Reference for a detailed explanation of the functionality of each function.