Co-Author(s)

Chris Lokken, Matthew Carter

Research Mentor(s)

Deneke, Wesley

Description

VikingBot is an automated AI that plays StarCraft by using a combination of machine learning and artificial intelligence. High level strategies are planned using the Brown-UMBC Reinforcement Learning and Planning (BURLAP), library which implements planning algorithms and provides interfaces for defining a domain and models of that domain for planning. For the planning, we used the BURLAP implementation of the sparse sampling algorithm because the time complexity is independent of the size of the state space, and we have to plan quickly in real time. SARSA reinforcement learning is used for a machine learning model that controls combat units. Various other helping functions are distributed to agent classes that aid in the AI in the different areas of the game. These agents are categorized as strategy, economy, combat, and intelligence. By using these parts in tandem VikingBot aims to use less training resources and still be able to play games at a high enough level to beat a human player.

Document Type

Event

Start Date

18-5-2020 12:00 AM

End Date

22-5-2020 12:00 AM

Department

Computer Science

Genre/Form

student projects, posters

Type

Image

Keywords

Machine Learning, AI, StarCraft

Rights

Copying of this document in whole or in part is allowable only for scholarly purposes. It is understood, however, that any copying or publication of this document for commercial purposes, or for financial gain, shall not be allowed without the author’s written permission.

Language

English

Format

application/pdf

Share

COinS
 
May 18th, 12:00 AM May 22nd, 12:00 AM

VikingBot: The StarCraft Artificial Intelligence

VikingBot is an automated AI that plays StarCraft by using a combination of machine learning and artificial intelligence. High level strategies are planned using the Brown-UMBC Reinforcement Learning and Planning (BURLAP), library which implements planning algorithms and provides interfaces for defining a domain and models of that domain for planning. For the planning, we used the BURLAP implementation of the sparse sampling algorithm because the time complexity is independent of the size of the state space, and we have to plan quickly in real time. SARSA reinforcement learning is used for a machine learning model that controls combat units. Various other helping functions are distributed to agent classes that aid in the AI in the different areas of the game. These agents are categorized as strategy, economy, combat, and intelligence. By using these parts in tandem VikingBot aims to use less training resources and still be able to play games at a high enough level to beat a human player.

 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.