- Date Finished: 1/2/2026
- Project was part of the AMSI Summer school 2026 for the subject Mathematical Optimisation Theory and Applications.
The project is about finding 50 postcode locations across Australia such that the distance between these postcodes are maximised under specified constraints. The fictional reason for this is to install the high-capacity communication hubs across the country and have it organised so that if there's a regional power out, it minimises other hubs being affected too. The algorithm is using real postcode location data collected by Matthew Proctor. It can found on his website
Note the algorithm used to solve this problem is the Outer Approximation Technique with cutting planes, inspired from this paper by Sandy Spiers, Hoa T. Bui, and Ryan Loxton. This is simply my own implementation of the algorithm for this particular problem.
More details about the project outline in the 'project 1.pdf' file and my mathematical formulation in 'Lfiles-report/report.pdf'. Accessing the Python implementation can be found in the 'implementation.ipynb' notebook.
This can be done in either the Anaconda environment or in a virtual environment on your computer. Choice is yours.
Ensure that Python 3.11 is installed as packages like Gurobi have not been installed to later versions by the time this project has finished.
In order to use this package properly, ensure that you have a Gurobi license since this algorithm utilises around 18,000 constraint variables for it to work. Standard gurobi package is limited to 1,000 constraint variables.
Once you have the python environment setup, you would be required to install the appropriate packages from the requirements.txt file. This can be done using the following command.
pip install -r requirements.txtPlease ensure that you have your virtual environment active (or Conda environment active) so the packages are installed locally instead of globally.