Code for the paper 'Monte Carlo Tree Search for Asymmetric Trees'
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


Code for the paper Monte Carlo Tree Search for Asymmetric Trees by Thomas M. Moerland, Joost Broekens, Aske Plaat and Catholijn M. Jonker.


  1. Install recent versions of:
  • Python 3
  • Tensorflow
  • Numpy
  • Matplotlib
  1. Clone this repository:
git clone


You can run a new experiment from the function. Hyperparameters can be parsed through the --hp option. Default hyperparameters are listed in mcts-t+/ For example, to start a default experiment on CartPole-v0:

cd mcts-t+
python3 --hp game=CartPole-v0

Reproducing Paper Results

The results of the paper can be reproduced by:

cd mcts-t+
bash jobs/

This automatically loop over the necessary hyperparameters. Running it will take quite long on a regular laptop though. You can submitted the runs to a SLURM cluster via

bash jobs/

Visualization of Results

Subsequently, you can visualize the output with

cd mcts-t+
python3 --home --plot_type mean --game your_game 

for some your_game of your choice.


	author = "Moerland, Thomas M and Broekens, Joost and Plaat, Aske and Jonker, Catholijn M",
	journal = "arXiv preprint arXiv:1805.09218",
	title = "{Monte Carlo Tree Search for Asymmetric Trees}",
	year = "2018"