YasolSolver.jl

YasolSolver.jl is an interface between MathOptInterface.jl and Yasol solver.

Please consult the providing website for further information about the solver and how to build models.

First, download the Yasol solver from here.

Second, install Yasol interface using Pkg.add.

import Pkg
Pkg.add("YasolSolver")

Models can be build using JuMP.jl package and will be solved using Yasol interface and Yasol solver.

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

using JuMP, YasolSolver

cd("C:/Yasol") # change path to Yasol .exe directory

model = Model(() -> YasolSolver.Optimizer()) # use the path to Yasol solver .exe

set_optimizer_attribute(model, "solver path", "C:/Yasol/Yasol_CLP")
set_optimizer_attribute(model, "time limit", 60)
set_optimizer_attribute(model, "output info", 1)
set_optimizer_attribute(model, "problem file name", "Example.qlp")

The solver supports 3 attributes that can be used with JuMP:

• solver path -> defines the path to the Yasol executable
• time limit -> defines the time limit in seconds
• output info -> defines output level (1 is recommended)
• problem file name -> defines filename of model; solution file will have the same name

Further, the solver specific parameter saved in Yasol.ini can be set and retrieved the following:

# change Yasol initial parameter
# format: solver directory, parameter name, value
YasolSolver.setInitialParameter("C:/Yasol", "writeOutputFile", 1)
# get initial parameter
# format: solver directory
@show YasolSolver.getInitialParameter("C:/Yasol")

Note: Do not change the default parameter without knowing their purpose!

Do the following to build and solve a JuMP model using Yasol solver:

@variable(model, x1, binary=true, lower_bound=0, upper_bound=1, YasolVariable, quantifier="exists", block=1)

@variable(model, x2, binary=true, lower_bound=0, upper_bound=1, YasolVariable, quantifier="exists", block=2)

@variable(model, x3, binary=true, lower_bound=0, upper_bound=1, YasolVariable, quantifier="all", block=3)

@variable(model, x4, binary=true, lower_bound=0, upper_bound=1, YasolVariable, quantifier="exists", block=4)

@constraint(model, con1, 1*x1 -2*x2 +1*x3 -1*x4 <= 1, YasolConstraint, quantifier="exists")

@constraint(model, con2, 1*x1 + 1*x2 +1*x3 -1*x4 <= 2, YasolConstraint, quantifier="exists")

@constraint(model, con3, 1*x1 + 1*x2 +1*x3 <= 2, YasolConstraint, quantifier="all")

@objective(model, Min, -1*x1 -2*x2 +2*x3 +1x4)

optimize!(model)

The package provides two JuMP extensions that are used in the example above:

To use Yasol variables, the keyword YasolVariable needs to be used followed by the parameter quantifier, that can have the values 'exists' or 'all' and the parameter block that needs to be an integer >= 1. Every variable can either be binary or an interger variable.

To use Yasol constraints, the keyword YasolConstraint needs to be used followed by the parameter quantifier, that can have the values 'exists' or 'all'. Constraints can also be used without the constraint extension.

After calling the optimize function, the solution will be available in the selected project directory. Additionally, the solution can be accessed using JuMP the following way:

@show termination_status(model)
@show value(x1)
@show objective_value(model)
@show solve_time(model)