/OpenAI-Gym-Taxi-v2

About solving the Taxi v2 from OpenAI using sarsa, sarsamax/q-learning, or expected sarsa

Primary LanguageJupyter Notebook

Taxi-V2

This repository contains an implementation of SARSA, SARSA-MAX, and Expected SARSA to solve the Taxi-v2 environment of OpenAI. The result is published here.

Content of this repository

  • openai-gym-taxi-v2.ipynb: the jupyter notebook that contains the implementation

Requirements

To run the codes, follow the next steps:

  • Create a new environment:
    • Linux or Mac:
     conda create --name taxi_v2 python=3.6
 source activate taxi_v2
    • Windows:
     conda create --name taxi_v2 python=3.6 
 activate taxi_v2
  • Perform a minimal install of OpenAI gym 
     pip install gym
  • Create an IPython kernel for the taxi_v2 environment
	python -m ipykernel install --user --name taxi_v2 --display-name "taxi_v2"
  • Start jupyter notebook from the root of this python codes
jupyter notebook
  • Once started, change the kernel through the menu Kernel>Change kernel>taxi_v2