A demo of agent-based modelling using Mesa in Python.
To begin:
Run
make install
conda activate abm_demo
Try out the visualisations of a basic model and a more complicated model (these are both based off Mesa's money model example).
Watch as N agents move randomly around a width x height grid.
python abm_demo/examples/basic_model_visualisation.py --N 20 --width 10 --height 10
Watch as agents move randomly around a grid whilst sharing candy with each other or not (based off Mesa's money model example):
python abm_demo/examples/candy_model_visualisation.py --N 20 --width 10 --height 10 --init_candy 1 --init_sharer_prop 0.9 --mutation_rate 0.01
Initial conditions to modify:
init_candy- the number of pieces of candy given to every agent initially.init_sharer_prop- the proportion of agents who are sharers initially.
Plots of the Gini coefficient and the number of sharer agents at a given time step are also given.
Open up the notebook abm_demo/analysis/Trick or Treat.ipynb to experiment with parameter combinations and model outcomes.