Publication Date

Spring 2020

Degree Type

Master's Project

Degree Name

Master of Science (MS)

Department

Computer Science

First Advisor

Chris Pollett

Second Advisor

Mark Stamp

Third Advisor

Fabio Di Troia

Keywords

Dou Di Zhu, deep Q-learning, rule-based

Abstract

We describe our implementation of AIs for the Chinese game Dou Di Zhu. Dou Di Zhu is a three-player game played with a standard 52 card deck together with two jokers. One player acts as a landlord and has the advantage of receiving three extra cards, the other two players play as peasants. We designed and implemented a Deep Q-learning Neural Network (DQN) agent to play the Dou Di Zhu. At the same time, we also designed and made a pure Q-learning based agent as well as a Zhou rule-based agent to compare with our main agent. We show the DQN model has a 10% higher win rate than the Q-learning model and Zhou rule-based model when playing as the landlord, and a 5% higher win rate than the other models when playing as a peasant.

Download

Share

COinS