EthicalGardeners documentation

EthicalGardeners is a PettingZoo multi-agent environment for simulating gardeners tending to a grid-world garden, including ethical considerations.

The goal is to make agents learn an ethically-aligned behaviour that includes and respects these considerations.

Installation

Install the package with pip:

pip install ethical-gardeners

For visualization support (pygame, opencv):

pip install "ethical-gardeners[viz]"

For metrics tracking (wandb):

pip install "ethical-gardeners[metrics]"

Quick Start

To run the simulation with the default configuration. After cloning the project, at the project root, use the following command:

python ethicalgardeners/main.py --config-name config

After installing the package using pip, you can run instead:

python -m ethicalgardeners.main --config-name config

For customization options, see the Launching the Ethical Gardeners Simulation tutorial.

Tutorials:

• Launching the Ethical Gardeners Simulation
  • Hydra Configuration System
  • Launching the Simulation
  • Understanding Metrics
  • Visualization Results
  • Example: Complete Analysis Configuration
• Configuration Guide for Ethical Gardeners Simulation
  • Hydra Configuration System Overview
  • Main Configuration Parameters
  • Grid Configuration
  • Observation Configuration
  • Metrics Configuration
  • Renderer Configuration
  • Advanced Configuration Examples
  • How to modify the Configuration
• How to Extend the Ethical Gardeners Environment
  • 1. Adding or Modifying Reward Functions
  • 2. Adding an Observation Type
  • 3. Adding or Modifying Metrics
  • 4. Adding Actions and Handling Them
  • 5. Adding a Cell Type
  • 6. Adding a New Renderer Type
• Using Ethical Gardeners with Custom Algorithms
  • Compatibility with Custom Algorithms
  • Requirements for Custom Algorithms
  • Using the Core Utility Functions