PipsSolver.jl is a Julia interface to the PIPS-NLP nonlinear optimization solver.
Running the solver requires a working PIPS-NLP installation following the instructions.
The PipsSolver.jl package works with the graph-based algebraic modeling package Plasmo.jl.
PipsSolver.jl can be installed using the following Julia Pkg command. Note that currently, PipsSolver.jl only works with Plasmo.jl v0.3.0.
using Pkg
Pkg.add(Pkg.PackageSpec(name="Plasmo", version="0.3.0"))
Pkg.add(PackageSpec(url="https://github.com/zavalab/PipsSolver.jl.git"))using MPIClusterManagers # to import MPIManager
using Distributed # need to also import Distributed to use addprocs()
#Setup worker environments
#This will load the environment specified in this script's directory onto each worker
@everywhere using Pkg
@everywhere Pkg.activate((@__DIR__))
#Load Plasmo and PipsSolver on every worker
@everywhere using Plasmo
@everywhere using PipsSolver
graph = OptiGraph()
#Add nodes to a GraphModel
@optinode(graph,n1)
@optinode(graph,n2)
@variable(n1,0 <= x <= 2)
@variable(n1,0 <= y <= 3)
@variable(n1, z >= 0)
@constraint(n1,x+y+z >= 4)
@objective(n1,Min,y)
@variable(n2,x)
@NLnodeconstraint(n2,ref,exp(x) >= 2)
@variable(n2,z >= 0)
@constraint(n2,z + x >= 4)
@objective(n2,Min,x)
@linkconstraint(graph,n1[:x] == n2[:x])
#Setup MPI manager
manager=MPIManager(np=2) # specify, number of mpi workers, launch cmd, etc.
addprocs(manager) # start mpi workers and add them as julia workers too.
#Map julia workers to MPI ranks
julia_workers = sort(collect(values(manager.mpi2j)))
#Distribute the graph to workers. This creates the variable `pipsgraph` on each worker with an allocation of optinodes.
remote_references = PipsSolver.distribute(graph,julia_workers,remote_name = :pipsgraph)
# The remote optigraphs can be queried if they are fetched from the other workers.
r1 = fetch(remote_references[1])
r2 = fetch(remote_references[2])
#Solve with PIPS-NLP
@mpi_do manager begin
using MPI
PipsSolver.pipsnlp_solve(pipsgraph)
end
#Retrieve Solution
rank_zero = manager.mpi2j[0] #julia process representing rank 0
solution = fetch(@spawnat(rank_zero, pipsgraph))