Lokapal

Computational Systemics

Computational Systemics is a series of generative agent-based models that operationalize concepts from the Conciliatorics framework.

An agent-based model (ABM) is a computational simulation approach used to study the behavior and interactions of autonomous entities—known as "agents"—and observe how these micro-level actions create complex global patterns or macro-level phenomena.

Unlike traditional top-down models that look at a system as a whole using aggregate equations, ABMs use a bottom-up approach. They simulate individual participants (such as people, cars, animals, or companies) following a set of specific rules to see how the system behaves organically over time.

The models are written in Mesa so they remain easy to inspect, modify, and extend. Each major version is preserved and offered separately, so a simpler, more familiar case stays available alongside later elaborations. All entries includes a short description, a showcase recording, and a link to its GitHub repository. To interact with a model, clone that repository and run it locally.

v0.1 — Schelling Operational Status

This first model recovers the classic Schelling segregation result while expressing the agents' decision rule through Operational Status.

  • Fronterization holds the agent's current attribute (binary in this version: two types).
  • Permeability is the minimum share of similar neighbors the agent requires. A higher value is stricter: the agent accepts less local difference.
  • Receptivity detects when the proportion of similar neighbors falls below that share.
  • Proceptivity moves the agent to a random empty cell when it is dissatisfied.

With a moderate permeability, around 0.3 to 0.5, the familiar pattern appears: agents start relatively mixed and gradually form larger same-type clusters. The recording shows the grid, the rising segregation index, and the count of dissatisfied agents.

The mapping stays close to the original Schelling micromotives so the Operational Status vocabulary can be checked against a well-known benchmark. Later versions will introduce multiple simultaneous criteria and dynamic changes to Fronterization.

Repository lokapal-xyz/computational-systemics-schelling-v0.1 Opens in a new tab.

How to run

The script needs Python 3.12 or newer. It prints a local address. Open that address in a browser. Stop the dashboard with Ctrl+C.

git clone https://github.com/lokapal-xyz/computational-systemics-schelling-v0.1
cd computational-systemics-schelling-v0.1
chmod +x setup.sh   # only needed once
./setup.sh