A Python implementation of the World2 model

• Quick install
• Hello World2
• How to go further
• How to cite the projet with Bibtex
• References & acknowledgment

PyWorld2 is a Python implementation of the World2 model, designed by Jay W. Forrester to predict trajectories of growth-focused Humankind in a world bounded by its physical limits. The model structure is thouroughly described in the book World Dynamics. Although World3 later became the reference, World2 is much less complex than its successor: it helps to get a good understanding of how the world could be modeled by a System Dynamics approach. World2 is a nonlinear 5-th order differential equation model with the following state variables:

• human population P,
• nonreplaceable natural resources NR,
• pollution POL,
• capital investement CI,
• and the fraction of capital devoted to agriculture CIAF.

The easiest way to install pyworld2 is via pip:

pip install pyworld2

It will automatically install the requirements numpy, scipy and matplotlib. Otherwise you can clone the repository and do:

python setup.py install

Run the provided example:

import pyworld2
pyworld2.hello_world2()

It simulates the 2 scenarios described in the Forrester's book:

• the standard run, or Business as Usual
• an hypothetical scenario where the usage rate of Natural Resources is reduced by 75% in 1970.

Below is a comparaison with the original figures. They plot the population P, pollution rate POLR, capital investment CI, quality of life QL and natural resources NR.

Standard run - Business as usual	Reduced usage of Natural Resources

If you want to run your own simulations with different policies, you need to:

• copy-paste the configuration file ./pyworld2/functions_switch_default.json into your current directory
• modify it at your convenience. Each policy is defined by its due date and its due variable (NAME switches to NAME1 at the year trigger.value).
• Finally, run the following lines of code:

from pyworld2 import World2

w2 = World2()
w2.set_state_variables()
w2.set_initial_state()
w2.set_table_functions()
w2.set_switch_functions("./functions_switch_modified.json")
w2.run()

pyworld2.utils.plot_world_state(w2)

The project is under the MIT Licence & open-source, see the licence terms for more details.

To cite PyWorld2 via BibTex:

@software{vanwynsberghe2020pyworld2,
  author       = {Charles Vanwynsberghe},
  title        = {{PyWorld2 - a Python implementation of the World2 
                   model.}},
  month        = nov,
  year         = 2020,
  publisher    = {Zenodo},
  version      = {v1.1},
  doi          = {10.5281/zenodo.4274586},
  url          = {https://doi.org/10.5281/zenodo.4274586}
}

PyWorld2 is forked from the Rworld2 Software held by Arnaud Mignan.

• World Dynamics, Jay W. Forrester (1973 second edition)
• RWorld2: reproducing doomsday models in R, Arnaud Mignan (2020).