Learning character behaviors using agent modeling in games

Richard Zhao, Duane Szafron

Research output: Chapter in Book/Report/Conference proceedingConference contribution

5 Citations (Scopus)

Abstract

Our goal is to provide learning mechanisms to game agents so they are capable of adapting to new behaviors based on the actions of other agents. We introduce a new on-line reinforcement learning (RL) algorithm, ALeRT-AM, that includes an agent-modeling mechanism. We implemented this algorithm in BioWare Corp.'s role-playing game, Neverwinter Nights to evaluate its effectiveness in a real game. Our experiments compare agents who use ALeRTAM with agents that use the non-agent modeling ALeRT RL algorithm and two other non-RL algorithms. We show that an ALeRT-AM agent is able to rapidly learn a winning strategy against other agents in a combat scenario and to adapt to changes in the environment.

Original languageEnglish (US)
Title of host publicationProceedings of the 5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009
Pages179-185
Number of pages7
StatePublished - Dec 1 2009
Event5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009 - Stanford, CA, United States
Duration: Oct 14 2009Oct 16 2009

Publication series

NameProceedings of the 5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009

Other

Other5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009
CountryUnited States
CityStanford, CA
Period10/14/0910/16/09

Fingerprint

Learning algorithms
Reinforcement learning
Modeling
Experiments
Reinforcement Learning
Experiment
Night
Scenarios

All Science Journal Classification (ASJC) codes

  • Artificial Intelligence
  • Visual Arts and Performing Arts

Cite this

Zhao, R., & Szafron, D. (2009). Learning character behaviors using agent modeling in games. In Proceedings of the 5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009 (pp. 179-185). (Proceedings of the 5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009).
Zhao, Richard ; Szafron, Duane. / Learning character behaviors using agent modeling in games. Proceedings of the 5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009. 2009. pp. 179-185 (Proceedings of the 5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009).
@inproceedings{148f89a8b5cd4b808b3920a2162bd031,
title = "Learning character behaviors using agent modeling in games",
abstract = "Our goal is to provide learning mechanisms to game agents so they are capable of adapting to new behaviors based on the actions of other agents. We introduce a new on-line reinforcement learning (RL) algorithm, ALeRT-AM, that includes an agent-modeling mechanism. We implemented this algorithm in BioWare Corp.'s role-playing game, Neverwinter Nights to evaluate its effectiveness in a real game. Our experiments compare agents who use ALeRTAM with agents that use the non-agent modeling ALeRT RL algorithm and two other non-RL algorithms. We show that an ALeRT-AM agent is able to rapidly learn a winning strategy against other agents in a combat scenario and to adapt to changes in the environment.",
author = "Richard Zhao and Duane Szafron",
year = "2009",
month = "12",
day = "1",
language = "English (US)",
isbn = "9781577354314",
series = "Proceedings of the 5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009",
pages = "179--185",
booktitle = "Proceedings of the 5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009",

}

Zhao, R & Szafron, D 2009, Learning character behaviors using agent modeling in games. in Proceedings of the 5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009. Proceedings of the 5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009, pp. 179-185, 5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009, Stanford, CA, United States, 10/14/09.

Learning character behaviors using agent modeling in games. / Zhao, Richard; Szafron, Duane.

Proceedings of the 5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009. 2009. p. 179-185 (Proceedings of the 5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Learning character behaviors using agent modeling in games

AU - Zhao, Richard

AU - Szafron, Duane

PY - 2009/12/1

Y1 - 2009/12/1

N2 - Our goal is to provide learning mechanisms to game agents so they are capable of adapting to new behaviors based on the actions of other agents. We introduce a new on-line reinforcement learning (RL) algorithm, ALeRT-AM, that includes an agent-modeling mechanism. We implemented this algorithm in BioWare Corp.'s role-playing game, Neverwinter Nights to evaluate its effectiveness in a real game. Our experiments compare agents who use ALeRTAM with agents that use the non-agent modeling ALeRT RL algorithm and two other non-RL algorithms. We show that an ALeRT-AM agent is able to rapidly learn a winning strategy against other agents in a combat scenario and to adapt to changes in the environment.

AB - Our goal is to provide learning mechanisms to game agents so they are capable of adapting to new behaviors based on the actions of other agents. We introduce a new on-line reinforcement learning (RL) algorithm, ALeRT-AM, that includes an agent-modeling mechanism. We implemented this algorithm in BioWare Corp.'s role-playing game, Neverwinter Nights to evaluate its effectiveness in a real game. Our experiments compare agents who use ALeRTAM with agents that use the non-agent modeling ALeRT RL algorithm and two other non-RL algorithms. We show that an ALeRT-AM agent is able to rapidly learn a winning strategy against other agents in a combat scenario and to adapt to changes in the environment.

UR - http://www.scopus.com/inward/record.url?scp=77955168244&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77955168244&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:77955168244

SN - 9781577354314

T3 - Proceedings of the 5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009

SP - 179

EP - 185

BT - Proceedings of the 5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009

ER -

Zhao R, Szafron D. Learning character behaviors using agent modeling in games. In Proceedings of the 5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009. 2009. p. 179-185. (Proceedings of the 5th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2009).