Overview
This is an example of Deep Q learning algorithm for solving reaching target tasks on the 2D robot arm. The environment is available through this link https://github.com/ekorudiawan/gym-robot-arm
Example of agent performance is shown in this animation
Result
Here is an example result of DQN agent performance that was trained after 2000 episodes.
Dependencies
- OpenAI Gym
- PyGame
- PyTorch 1.4.0