Date of Award


Document Type

Thesis campus only


Computer Science

First Advisor

Britton Horn

Second Advisor

Sheng Tan


Generating decks in Collectible Card Games (CCG’s) has been a hot spot for artificial intelligence in recent years. Artificial intelligence methods have usually attacked this problem without considering the specific cards accessible to a player. In this paper, we examine the capability to generate decks in Hearthstone, a CCG, for individual players based on their cards using Evolutionary Algorithms (EA) with a neural network fitness function. In these experiments, the approach shows a promising result that EA algorithms can generate good decks for a players collection.