https://github.com/jolares/deep-reinforcement-learning-navigation

https://github.com/jolares/deep-reinforcement-learning-navigation

deep-reinforcement-learning dqn pytorch

Last synced: 24 days ago
JSON representation

• Host: GitHub
• URL: https://github.com/jolares/deep-reinforcement-learning-navigation
• Owner: jolares
• License: mit
• Created: 2020-06-09T21:13:53.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2022-11-21T22:24:31.000Z (almost 2 years ago)
• Last Synced: 2023-03-07T06:26:55.928Z (over 1 year ago)
• Topics: deep-reinforcement-learning, dqn, pytorch
• Language: ASP
• Homepage:
• Size: 26.2 MB
• Stars: 0
• Watchers: 3
• Forks: 0
• Open Issues: 2
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Dueling Double DQN (D3QN) Learning: 
 Collecting the Ripe Bananas

TODO: GIF of trained Agent next to Plot of Rewards per episode & avg reward per episode

This project's goal is to implement an agent capable of learning navigation in a large grid-world environment


as it also learns how to identify ripe bananas from reading raw-pixel data real-time and maximize the number it

can collect in a given time frame.

For the time being, this is a living document listing some of the project's specifications.

## The problem

TODO

## The Environment

TODO

## Markov Decision Process (MDP)

* TODO: state and action spaces

* TODO:Consideration of environment-solved

## Implementation Details

The report (Report.ipynb) describes:

* TODO: The D3QN learning algorithm and chosen hyperparameters

* TODO: The model neural networks architecture

* TODO: Future ideas for improving the agent's performance.

## Running the Project

TODO:

### Software Dependencies & Installation

* Python 3

* PyTorch

* Numpy

## Resources

[Deep Reinforcement Learning with Double Q-learning](https://arxiv.org/abs/1509.06461)

[Dueling Network Architectures for Deep Reinforcement Learning](https://arxiv.org/abs/1511.06581)

[Prioritized Experience Replay](https://arxiv.org/pdf/1511.05952v4.pdf)

[Reinforcement Learning (Sutton & Barto, 2020)](http://incompleteideas.net/book/RLbook2020.pdf)