Leafcutter Ants Fungi Mutualism Agent-Based Model

An agent-based model to study the dynamics of the mutualism between leafcutter ants and their fungus garden. It is implemented using the Mesa framework.

Demo

Installation

To install the dependencies using pip, use the following command

$ pip install -r requirements.txt

This will install all dependencies to require the model and notebooks (used for analyzing and visualizing data)

How to run

To run the model interactively in the browser, run the following command in the root directory

$ mesa runserver

This will start a webserver at http://127.0.0.1:8521 which you can visit using your browser.

Alternatively, one can run the model directly without live visualization by using leafcutter_ants_fungi_mutualism/run_model.py. This is significantly faster and persists the collected data. The usage is as follows

usage: run_model.py [-h] [-r REPETITIONS] [-t TIME_STEPS] [-n N_CORES]
                    [-c COLLECT_TIMESERIES]
                    output_file

Leafcutter Ants Fungy Mutualism model runner

positional arguments:
  output_file           location of output file

optional arguments:
  -h, --help            show this help message and exit
  -r REPETITIONS, --repetitions REPETITIONS
                        number of repeated model runs
  -t TIME_STEPS, --time-steps TIME_STEPS
                        number of time steps to execute
  -n N_CORES, --n-cores N_CORES
                        number of processes to use in pool
  -c COLLECT_TIMESERIES, --collect-timeseries COLLECT_TIMESERIES
                        collect timeseries data

For example, the following command runs 100 repetitions of the model using 32 cores for 5000 time steps while collecting timeseries data:

$ python3 run_model.py data/N100_t5000.npz --repetitions=100 --n-cores=32 --collect-timeseries=True --time-steps=5000

Visualization

Data visualization notebooks can be found in the leafcutter_ants_fungi_mutualism folder:

• Experiments.ipynb is used for creating the experimental result figures
• Gif visualization.ipynb is used for creating the gif displayed above
• OFAT_visualization.ipynb is used for visualizing OFAT results
• Sobol_visualization.ipynb is used for visualizing Sobol results
• Time steps visualization.ipynb is used for general timeseries visualization

Mesa Changes

In order to use the parallelized version of Mesa's BatchRunner, we apply one of the fixes proposed in mesa/#1107. The fixed BatchRunnerMP is available in leafcutter_ants_fungi_mutualism/batchrunner.py.