The StructJuMP package provides a parallel algebraic modeling framework for block structured optimization models in Julia. StructJuMP, originally known as StochJuMP, is tailored to two-stage stochastic optimization problems and uses MPI to enable a parallel, distributed memory instantiation of the problem. StructJuMP.jl is an extension of the JuMP.jl package, which is as fast as AMPL and faster than any other modeling tools such as GAMS and Pyomo (see this).
The most recently tagged version of this package works with most recently tagged version of JuMP, i.e. JuMP v0.18.x, to try this package with JuMP v0.18, simply do
] add StructJuMP
The development of this package for JuMP v0.18.x is done in the
master branch of this package works with the JuMP development version. To
try this package with JuMP master, do:
] add JuMP#master ] add StructJuMP#master
using StructJuMP numScen = 2 m = StructuredModel(num_scenarios=numScen) @variable(m, x[1:2]) @NLconstraint(m, x + x == 100) @NLobjective(m, Min, x^2 + x^2 + x*x) for i in 1:numScen bl = StructuredModel(parent=m, id=i) @variable(bl, y[1:2]) @NLconstraint(bl, x + y + y ≥ 0) @NLconstraint(bl, x + y + y ≤ 50) @NLobjective(bl, Min, y^2 + y^2 + y*y) end
The above example builds a two level structured model
m with 2 scenarios.
Problems modeled in StructJuMP models can be solved in parallel using the PIPS-NLP parallel optimization solver. In addition, StructJuMP models can be solved (in serial only) using Ipopt. The SturctJuMP models interface with the solvers via StructJuMPSolverInterface.jl.
StructDualDynProg can run the SDDP algorithm on multi-stage models from StructJuMP.
StructJuMP has been developed under the financial support of Department of Energy (DOE), Office of Advanced Scientific Computing Research, Office of Electricity Delivery and Energy Reliability, and Grid Modernization Laboratory Consortium (GMLC) (PIs: Cosmin G. Petra, Lawrence Livermore National Laboratory and Mihai Anitescu, Argonne National Laboratory).
about 2 months ago