davidanthoff / KNITRO.jl

Julia interface to the Knitro solver

KNITRO.jl

The KNITRO.jl package provides an interface for using the KNITRO solver from the Julia language. You cannot use KNITRO.jl without having purchased and installed a copy of KNITRO from Ziena Optimization. This package is available free of charge and in no way replaces or alters any functionality of Ziena's KNITRO solver.

KNITRO functionality is extensive, so coverage is incomplete, but the basic functionality for solving linear, nonlinear, and mixed-integer programs is provided. Documentation is available on ReadTheDocs.

Setting up KNITRO on Linux and OS X

1. First, you must obtain a copy of the KNITRO software and a license; trial versions and academic licenses are available here.

2. Once KNITRO is installed on your machine, point the LD_LIBRARY_PATH (Linux) or DYLD_LIBRARY_PATH (OS X) variable to the KNITRO library by adding, e.g.,

export LD_LIBRARY_PATH="$HOME/knitro-9.0.1-z/lib:$LD_LIBRARY_PATH"

export DYLD_LIBRARY_PATH="$HOME/knitro-9.0.1-z/lib:$DYLD_LIBRARY_PATH"

to your start-up file (e.g. .bash_profile).

3. To activate KNITRO for your computer you will need a valid Ziena license file (which looks like ziena_lic_*.txt). The simplest procedure is to copy each license into your HOME directory.

4. At the Julia prompt, run

julia> Pkg.add("KNITRO")

5. Test that KNITRO works by runnning

julia> Pkg.test("KNITRO")

Setting up KNITRO on Windows

Note that currently only 64-bit Windows is supported. That is, you must use 64-bit Julia and install the Win64 version of KNITRO.

1. First, you must obtain a copy of the KNITRO software and a license; trial versions and academic licenses are available here.

2. Once KNITRO is installed on your machine, add the directory containing knitro.dll to the PATH environment variable, as described in the KNITRO documentation.

3. To activate KNITRO for your computer you will need a valid Ziena license file (which looks like ziena_lic_*.txt). The simplest procedure is to copy each license into your HOME directory.

4. At the Julia prompt, run

julia> Pkg.add("KNITRO")

5. Test that KNITRO works by runnning

julia> Pkg.test("KNITRO")

MathProgBase Interface

KNITRO implements the solver-independent MathProgBase interface, and so can be used within modeling software like JuMP. The solver object is called KnitroSolver. All options listed in the KNITRO documentation may be passed directly. For example, you can run all algorithms by saying KnitroSolver(KTR_PARAM_ALG=KTR_ALG_MULTI).