Implementation of the Normalized Advantage Functions algorithm (from the paper Continuous Deep Q-Learning with Model-based Acceleration, Gu et al., 2016: https://arxiv.org/abs/1603.00748) in Python TensorFlow.
Motivated by OpenAI request for research: https://openai.com/requests-for-research/#improved-q-learning-with-continuous-actions
Dependencies: gym, tensorflow (1.5.0), numpy, matplotlib
Usage: python experiment.py [-h]
OpenAI environments: https://gym.openai.com/envs/