Gurobi

Gurobi#

The Gurobi mathematical optimisation solver is available to use on HPC. For details see the Gurobi Documentation.

Gurobi module#

To access the Gurobi solver you need to load the gurobi module. You can do this with the command:

$ module add gurobi

This sets a number of environment variables which ensure any Gurobi APIs know how to communicate with the solver, these include:

GUROBI_HOME
GRB_LICENSE_FILE
GUROBI_R_PKG

Python setup#

To use Gurobi in a python script you will need to install the Gurobipy package. You can do this using either pip or conda with the following commands:

# installing with pip
$ pip install --user -i https://pypi.gurobi.com gurobipy

# installing with conda
$ conda install -c gurobi gurobi

You can then use the Gurobi Python API by adding the line import gurobi to your python script and calling functions from the module.

R setup#

Note

The Gurobi R package is currently only tested as working with R version 4.0.3

To use Gurobi with the R language you need to install the R gurobi package and the R package slam. This is not on CRAN so to install this on ARC3/ARC4 you must install the package using one of the following methods:

# install Gurobi and slam within the R console
> install.packages("slam"))
> install.packages(c(Sys.getenv("GUROBI_R_PKG"))

# install Gurobi and slam from the shell
$ R -e "install.packages('slam')"
$ R -e "install.packages(c('$GUROBI_R_PKG'))"

Once you have installed the package you can then load it within the R environment using the line library(gurobi). Once the package is loaded you can use Gurobi R API functions to use the solver.

Installing R 4.0.3 via Anaconda module#

You can install R 4.0.3 on the HPC by using the Anaconda module via the conda package manager. In its simplest form you can do this with the following commands:

# add the anaconda module
$ module add anaconda

# create a conda environment in which we install R 4.0.3 from
# the conda forge channel
$ conda create -n R403 -c conda-forge r-base=4.0.3

# to enter the environment and use R 4.0.3
$ source activate R403

Matlab setup#

To configure MATLAB to use the Gurobi module you will need to have loaded the Gurobi module and within the MATLAB command window run the following commands:

% change directory into the Gurobi matlab directory
cd(getenv("GUROBI_HOME")+"/matlab")

% run the gurobi_setup.m file to configure MATLAB
gurobi_setup

This adds the Gurobi MATLAB directory to your MATLAB path in your current MATLAB session. To permanently save this directory to your MATLAB path you can use the following commands after performing the above:

% save your current (updated) MATLAB path in your home directory
savepath(getenv("HOME")+"/pathdef.m")

Once you have run gurobi_setup you should now have access to the Gurobi MATLAB API and can run Gurobi specific functions.

Contents