Skip to content

Chapter 5: Cellular Automata

Part II — Simulation Paradigms

Implemented by the dissmodel-ca package.

Live demo: try the models in this chapter directly in the browser — dissmodel-ca-demo on Hugging Face Spaces.

Learning objectives

  • Install and use the dissmodel-ca extension
  • Know the classic and research models included
  • Choose between vector and raster substrate for a CA model

5.1 Overview

dissmodel-ca provides a collection of cellular automata models implemented on top of the dissmodel engine, in both vector (GeoDataFrame) and raster (NumPy) versions.

5.2 Included models

Model Substrate Description
GameOfLife Vector / Raster Classic Conway's simulation
FireModel Vector / Raster Forest fire spread with probabilistic regrowth
Snow Vector Snowfall accumulation and gravity dynamics
Growth Vector Stochastic radial growth
Anneal Vector Binary system relaxation via majority-vote rule
Excitable Vector Excitable medium waves (spiral/ring patterns)
Parasit Vector Host-parasite spatial dynamics
Interspecific Vector Grass species competition model

5.3 Installation and quick start

pip install .
python examples/cli/ca_game_of_life.py
streamlit run examples/streamlit/ca_all.py
jupyter lab examples/notebooks/ca_game_of_life.ipynb

5.4 Repository structure

  • src/dissmodel_ca/models/ — core implementations (the "Science" layer)
  • examples/notebooks/ — 15+ didactic notebooks (in Portuguese)
  • examples/cli/ — self-contained scripts for quick testing
  • examples/streamlit/ — reactive UI components

Exercises

  1. CellularAutomaton.rule(idx) is an abstract method — every model in 5.2 must implement it. Open src/dissmodel_ca/models/fire_model.py and identify how FireModel.rule(idx) decides a cell's next FireState (FOREST/BURNING/BURNED) based on its Rook neighbors. Why does the model use Rook instead of the default Queen neighborhood used elsewhere in the package?
  2. CellularAutomaton.execute() applies rule to every cell via self.gdf.index.map(self.rule), which cannot be vectorized because rule is arbitrary Python. Compare the vector FireModel to fire_model_raster.py and describe, in your own words, what vectorization technique the raster version uses instead of per-cell rule() calls.
  3. Every CellularAutomaton subclass overrides initialize() to set the starting state instead of doing it in __init__. Explain why — what does initialize() have access to that __init__ doesn't, given the Model.__init__ → setup(**kwargs) lifecycle from Chapter 2?
  4. Pick one model from 5.2 you haven't read yet (e.g. Anneal, Excitable, or Interspecific) and, from its docstring and rule() body alone, write a one-paragraph description of the phenomenon it models and which neighborhood strategy it relies on.

Summary

Every model in dissmodel-ca is a CellularAutomaton — a SpatialModel that trades a free-form execute() for a stricter, cell-by-cell rule(idx) contract: initialize() sets the starting state, and execute() applies rule to every cell each tick via gdf.index.map(self.rule). That contract is what makes the package's Streamlit explorer (Chapter 3) possible — auto-discovering every concrete CellularAutomaton subclass works precisely because they all share the same shape. Classic models (GameOfLife, FireModel) and research models (Snow, Growth, Anneal, Excitable, Parasit, Interspecific) plug into this same base class, differing only in their state space and neighborhood strategy (Queen by default, Rook where fire/diffusion logic calls for 4-directional spread). Vector and raster substrates implement the same rule with different mechanics — per-cell rule() lookups on GeoDataFrame versus vectorized NumPy operations on RasterBackend — which is why GameOfLife and FireModel each ship both a vector and a raster class.