abhishek-iyer1 / cartpole-deep-reinforced-learning

We implement a DQN Agent using a Boltzmann Q Policy to tame the Cart Pole problem in the discrete action space. The method is tested against the baseline of making random movements and is implemented using the OpenAI Gym.

cartpole-deep-reinforced-learning

We implement a DQN Agent using a Boltzmann Q Policy to tame the Cart Pole problem in the discrete action space. The method is tested against the baseline of making random movements and is implemented using the OpenAI Gym.

The notebook is split into sections and is pretty self explanatory. You can get the following outputs by following the instructions in the CartPole-DRL.ipynb notebook.

This is how the baseline method which takes random actions performs.

train.mp4

Here is how our trained DQN Agent performs.

test.mp4