CPLEXCP.jl

CPLEXCP.jl is a wrapper for the IBM® ILOG® CPLEX® CP Optimizer.

You cannot use CPLEXCP.jl without having purchased and installed a copy of CPLEX Optimization Studio from IBM. However, CPLEX is available for free to academics and students.

CPLEXCP.jl has two components:

• a thin wrapper around the complete Java API
• an interface to MathOptInterface and ConstraintProgrammingExtensions.jl

The Java API can be accessed via CPLEXCP.cpo_java_xxx functions, where the names and arguments are built from the Java API. See the CPLEX documentation for details.

Note: This wrapper is maintained by the community and is not officially supported by IBM. If you are a commercial customer interested in official support for CPLEX CP Optimizer in Julia, let them know!

Installation

Minimum version requirement: CPLEXCP.jl requires CPLEX version 20.1. Other versions may work, but are not tested. Versions 12.9 and 12.10 are automatically detected, though.

First, obtain a license of CPLEX and install CPLEX CP optimizer, following the instructions on IBM's website. Then, set the CPLEX_STUDIO_DIR environment variable as appropriate and run Pkg.add("CPLEXCP"), then Pkg.build("CPLEXCP"). For example:

# On Windows, this might be
ENV["CPLEX_STUDIO_DIR"] = "C:\\Program Files\\CPLEX_Studio1210\\"
import Pkg
Pkg.add("CPLEXCP")
Pkg.build("CPLEXCP")

# On macOS, this might be
ENV["CPLEX_STUDIO_DIR"] = "/Applications/CPLEX_Studio1210/"
import Pkg
Pkg.add("CPLEXCP")
Pkg.build("CPLEXCP")

# On Unix, this might be
ENV["CPLEX_STUDIO_DIR"] = "/opt/CPLEX_Studio1210/"
import Pkg
Pkg.add("CPLEXCP")
Pkg.build("CPLEXCP")

Note: your path may differ. Check which folder you installed CPLEX in, and update the path accordingly.

Example with MOI

Here is an example of use of CPLEXCP.jl through MOI and CP:

using MathOptInterface
using ConstraintProgrammingExtensions
using CPLEXCP

const MOI = MathOptInterface
const CP = ConstraintProgrammingExtensions

model = CPLEXCP.Optimizer()

# Create the variables: six countriers; the value is the colour number for each country
belgium, _ = MOI.add_constrained_variable(model, MOI.Integer())
denmark, _ = MOI.add_constrained_variable(model, MOI.Integer())
france, _ = MOI.add_constrained_variable(model, MOI.Integer())
germany, _ = MOI.add_constrained_variable(model, MOI.Integer())
luxembourg, _ = MOI.add_constrained_variable(model, MOI.Integer())
netherlands, _ = MOI.add_constrained_variable(model, MOI.Integer())

# Constrain the colours to be in {0, 1, 2, 3}
MOI.add_constraint(model, belgium, MOI.Interval(0, 3))
MOI.add_constraint(model, denmark, MOI.Interval(0, 3))
MOI.add_constraint(model, france, MOI.Interval(0, 3))
MOI.add_constraint(model, germany, MOI.Interval(0, 3))
MOI.add_constraint(model, luxembourg, MOI.Interval(0, 3))
MOI.add_constraint(model, netherlands, MOI.Interval(0, 3))

# Two adjacent countries must have different colours.
countries(c1, c2) = MOI.ScalarAffineFunction(MOI.ScalarAffineTerm.([1, -1], [c1, c2]), 0)
MOI.add_constraint(model, countries(belgium, france), CP.DifferentFrom(0))
MOI.add_constraint(model, countries(belgium, germany), CP.DifferentFrom(0))
MOI.add_constraint(model, countries(belgium, netherlands), CP.DifferentFrom(0))
MOI.add_constraint(model, countries(belgium, luxembourg), CP.DifferentFrom(0))
MOI.add_constraint(model, countries(denmark, germany), CP.DifferentFrom(0))
MOI.add_constraint(model, countries(france, germany), CP.DifferentFrom(0))
MOI.add_constraint(model, countries(france, luxembourg), CP.DifferentFrom(0))
MOI.add_constraint(model, countries(germany, luxembourg), CP.DifferentFrom(0))
MOI.add_constraint(model, countries(germany, netherlands), CP.DifferentFrom(0))

# Solve the model.
MOI.optimize!(model)

# Check if the solution is optimum.
@assert MOI.get(model, MOI.TerminationStatus()) == MOI.OPTIMAL

# Get the solution
@show MOI.get(model, MOI.VariablePrimal(), belgium)
@show MOI.get(model, MOI.VariablePrimal(), denmark)
@show MOI.get(model, MOI.VariablePrimal(), france)
@show MOI.get(model, MOI.VariablePrimal(), germany)
@show MOI.get(model, MOI.VariablePrimal(), luxembourg)
@show MOI.get(model, MOI.VariablePrimal(), netherlands)

Use with JuMP

We highly recommend that you use the CPLEXCP.jl package with higher level packages such as JuMP.jl. However, for now, JuMP hasn't caught up with MOI 0.10; you will not be able to use JuMP with the latest version of CPLEXCP.jl.

This can be done using a CPLEXCP.Optimizer object. Here is how to create a JuMP model that uses Chuffed as solver.

using JuMP, CPLEXCP

model = Model(CPLEXCP.Optimizer)