Skip to content

CLAIR-LAB-TECHNION/multi-taxi

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

74 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Multi Taxi Environment

taxi env example map

multi-taxi is a highly configurable multi-agent environment, based on gym's taxi environment, that adheres to the PettingZoo API. Some configurations include:

  1. the number of taxis and passengers in the environment (limited to the size of the map)
  2. the domain map itself
  3. the environment objective
  4. individual taxi configurations:
    1. reward function
    2. action and observation space
    3. passenger and fuel capacity
  5. and so much more!

For a quickstart guide and a deeper dive into the environment and its configuraions, please consult our demonstration notebook, also available in colab and nbviewer.

Installation

The easiest way to install multi-taxi is directly from the git repository using pip. Here is how to install the latest stable version:

pip install "git+https://github.com/CLAIR-LAB-TECHNION/multi-taxi@0.4.0"

You can also download our latest updates by not specifying a tag, like so:

pip install "git+https://github.com/CLAIR-LAB-TECHNION/multi-taxi"

If you wish to install the environment that uses the legacy pettingzoo API, please install version 0.3.0 like so:

pip install "git+https://github.com/CLAIR-LAB-TECHNION/multi-taxi@0.3.0"

If you are seeking the legacy version, which is based on the RLLib API, please install version 0.0.0 like so:

pip install "git+https://github.com/CLAIR-LAB-TECHNION/multi-taxi@0.0.0"

Acknowledgements

This library is based on MultiTaxiLib by Ofir Abu. The original implementation paper can be found here.

Citation

To cite this repository in academic works or any other purpose, please use the following BibTeX citation:

@software{Azran_Multi_Taxi_A_2023,
      author = {Azran, Guy and Keren, Sarah},
      month = {3},
      title = {{Multi Taxi: A Modular Setting for Multi-Agent Systems Experiments}},
      url = {https://github.com/CLAIR-LAB-TECHNION/multi-taxi},
      version = {0.4.0},
      year = {2023}
}

Alternatively, we offer a CITATION.cff file with GitHub and Zotero integration.