Skip to content

Agent Based Modelling in Python (Formative Coursework)


These notes are based on what I found might be useful to me when modelling the wealth distribution in asset exchange models. This was the paper that my group chose for our formative coursework.

For the formative coursework, we established that the best plan of action would be to individually look at the paper, then see if we could all understand it, then implement our own, simpler models.

Following this, we should then have the ability to implement the more complicated model as is outlined in the paper. We chose to implement the model in Python as it would not be running for huge lengths of time, and Python is a fairly well supported, accessible language for all members of the team.

Library Choice

As the paper models wealth in terms of agents, we looked towards a library that would help us. mesa is a Python library that will do this for us (also, side note, the open source graphics driver for Linux, not confusing at all).

There is a good tutorial on a simpler version of a wealth distribution model (covered in this paper). This is done in a Google Colab notebook, so that the code at each step can be explained, and the toolchain is not too difficult to get up and running.