This is an implementation of a multi-rover control problem in a continuous domain using deep reinforcement learning.
The training script is stored in runners
as run_contworld.py
. To visualize certain results, use the script vis_contworld.py
in visualization
.
This library requires Python>=3.5 and the packages listed in requirements.txt.
-
rllab is a framework for developing and evaluating reinforcement learning algorithms, developed by Yan Duan, Xi Chen, Rein Houthooft, John Schulman and Pieter Abbeel at Berkeley and OpenAI.
-
rltools is a library that contains helper tools for working with reinforcement learning algorithms, developed by the Stanford Intelligent Systems Laboratory.