Agent based modeling in Python
summary
Describes a concrete ABM stack: Mesa in Python with examples (Schelling, Wolf‑Sheep), a Tornado‑backed server, a batch runner for parameter sweeps and a data collector that outputs pandas frames. Demonstrates that activation order matters — sequential vs simultaneous activation produces qualitatively different outcomes — and that browser visualizations run live but heavy geospatial features push into separate libraries. Project is pre‑1.0, started with two contributors and grows to ~26, and the demos mostly work though some chart bugs remain.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 3:39Complexity in agent-based modeling
- 6:08Agent-based models and uses
- 10:48Model types and examples
- 13:00Live demo to Mesa
- 17:59Agent-based modeling tools
- 21:37Mesa structure and models
- 28:31Mesa server and visualization
- 31:25Scheduling in Mesa
- 34:36Mesa's future plans
- 37:05Growth and contributions
related talks
25mDirecting a Swarm of Agents for Fun and Profit
21mScalable agent-based models for optimized policy design: applications to the ec...
30mDistributing AI with Python in the Browser: Edge Inference and Flexibility...
41mBuilding LLM Agents – Challenges and Learnings
18mMulti-Agent Systems for Traceable Bayesian Workflow
48mAI Coding Agents and how to code them
53mFrom Mixture of Experts to Mixture of Agents with Super Fast Inference
47mThe State of the Art for Probabilistic Programming