DSPopt.Methods
— TypeMethods
List of methods available.
ExtensiveForm
: deterministic equivalent formulation solveBenders
: Benders decompositionDW
: Dantzig-Wolfe decompositionDual
: Dual decomposition
DSPopt.getBlockIds
— FunctiongetBlockIds
Get the vector of block IDs that are assigned to the current MPI rank
Arguments
nblocks
: number of blocks (by default,DSPProblem.nblocks
)master_has_subblocks
: indicate whether the master process solves (scenario) blocks or not (should always befalse
). This should not be modified.
DSPopt.getNumBlockCols
— MethodgetNumBlockCols
Get the number of columns for each block
DSPopt.loadStochasticProblem!
— MethodloadStochasticProblem!
Load stochastic programming problem from StructJuMP
Arguments
model
: StructJuMP model
DSPopt.loadStructuredProblem!
— MethodloadStructuredProblem!
Load generic block-structured problem from StructJuMP
Arguments
model
: StructJuMP model
DSPopt.load_problem!
— Methodload_problem!
Load problem from StructJuMP
Arguments
m
: StructJuMP model
DSPopt.myrank
— Methodmyrank
Returns the MPI rank (i.e., process id), starting from 0
DSPopt.mysize
— Methodmysize
Returns the number of MPI processes
DSPopt.parallelize
— Methodparallelize
Initialize DSP environment related to MPI, if MPI is initialized.
DSPopt.set
— Methodset
Set the parameters for Wasserstein ambiguity set
Arguments
dsp
: DSPProblem structp
: norm of the Wasserstein distanceϵ
: the size of the Wasserstein ball
DSPopt.setBlocks
— MethodsetBlocks
The function sets the number of blocks (e.g., scenarios for stochastic program) and their Ids.
DSPopt.setoptions!
— Methodsetoptions!
This takes options
arguments for setting DSP options.
Arguments
options
: possible keys are:param
(path to parameter file),:is_stochastic
(true
if stochastic model; otherwise,false
), and:solve_type
(algorithm type; seeMethods
).
DSPopt.writeMps!
— MethodwriteMps!
Write the deterministic equivalent formulation in MPS format.