Sim - Sugarscape

[bordumb]

Sugarscape Model Implementation

Description

This PR introduces a complete implementation of the classic Sugarscape model, designed to serve as a robust testbed for evaluating and comparing different agent cognitive strategies in a resource-scarce environment.

The primary goal of this simulation is to quantitatively measure the survival advantage of learning-based agents over simpler heuristic and random strategies. The experiment has been designed as a rigorous A/B/C/D test with statistical analysis to validate the effectiveness of our Q-learning architecture.

This work is extensively documented in the new blog post, which covers the experimental design, statistical analysis, and a full interpretation of the results.

Key Changes

• New Simulation: Adds the complete simulations/sugarscape_sim module, including all necessary components, systems, actions, and providers.
• New Experiment: Includes a new experiment file (validate_sim.yml) that defines the four-group study comparing heuristic, random, and two Q-learning agent architectures.
• Statistical Analysis: Adds a new analysis script (analyze_sugarscape.py) that uses ANOVA and Tukey's HSD to rigorously evaluate the experimental results.
• Documentation: Adds a new blog post, "Survival of the Smartest: A Statistical Showdown in the Sugarscape," detailing the experiment.

How to Test & Run

1. Run the full experiment:

make run FILE=simulations/sugarscape_sim/experiments/validate_sim.yml WORKERS=8

2. Analyze the results:

docker compose exec app poetry run python simulations/sugarscape_sim/analysis/analyze_sugarscape.py