Board Games URAP Project 🎲

This repository contains models of board games for an undergraduate research project started in Fall 2015, carried out under the direction and supervision of Falk Lieder, Elizabeth Kon, and Dr. Tom Griffiths of the Computational Cognitive Science Lab at UC Berkeley.

Games List

The 10 games selected for analysis based on ratings pulled from this database. The relevant files for each game are also listed.

1. Peg Solitaire (peg_solitaire.py peg_markov.py)
2. Solitaire/Patience
3. Wolfpack
4. Jasper and Zot (JasperAndZot.py)
5. Legions of Darkness
6. B-29 Superfortress
7. Utopia Engine (UtopiaEngine.py minigame.py currently in progress)
8. Field Commander: Napoleon
9. Where There is Discord: War in the South Atlantic
10. Thunderbolt Apache Leader

A model of a 2x2 game of Tic-Tac-Toe is also included in this repository as a bonus. (TicTacToe.py Markov.py)

Other Dependencies

Besides downloading the relevant files for the game, you will also need NumPy ver1.9 or later. It is recommended that you download it as part of the Anaconda package to ensure that further instructions work for you.

The UC Berkeley python library datascience is also necessary in order to run Features.py. Once you have Anaconda installed, go ahead and run

pip install datascience

Playing a Game

To run any file in interactive mode, use the command python -i followed by the name of the file. Once you are running the main game file (the file that doesn't have "markov" in the title) in interactive mode, type in play() hit Enter and enjoy. Refer to code comments for further instructions on how to run learning algorithms on our models.

Functions According to the MDP Framework

(Functionally equivalent functions are listed on the same line)

States

• create_states
• state_space
• next_states
• create_state_tree
• board2state
• state2board
• state_transition

Actions

• action_space, possible_actions
• legal_actions

Transition Probabilities

• transition_prob
• state_transition, simulate_transition
• transition_prob_matrix

Reward

• reward_function, reward

URAP Apprentices

---

• Fall 2015 - Spring 2017: Priyam Das, David Lu
• Fall 2015 - Fall 2016: Jackie Dong