A Monte Carlo Tree Search Agent used to control agents in a Pacman competition.
Python TeX Ruby
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
doc
docs
layouts
teams
.gitignore
README.md
capture.py
captureAgents.py
commands.txt
config.py
contest.html
distanceCalculator.py
expectimax.py
game.py
ghostAgents.py
graphicsDisplay.py
graphicsUtils.py
ideas.txt
keyboardAgents.py
layout.py
mazeGenerator.py
pacclient.py
pacman.py
pacmanAgents.py
projects.css
screen.png
search.py
test_maker.rb
textDisplay.py
unpack.py
util.py

README.md

MCTS for Pacman Capture the Flag Contest

screenshot

This project implements an agent to participate in the Pacman Capture the Flag contest.

More information on the contest can be found at: http://ai.berkeley.edu/contest.html. The contest rules can also be found on this link.

The team implemented in this project is named MonteCarloTeam. It uses Monte Carlo Tree Search (MCTS) to implement an offensive and a defensive agent. Its source code and its documentation can be found here.