Gal A. Kaminka: Publications

Sorted by DateClassified by Publication TypeClassified by TopicGrouped by Student (current)Grouped by Former Students

Context-dependent joint-decision arbitrarion for computer games

Gal A. Kaminka, Jared Go, and Thuc D. Vu. Context-dependent joint-decision arbitrarion for computer games. In Proceedings of the Agents in Computer Games Workshop, 2002.

Download

[PDF]185.3kB  [gzipped postscript]67.5kB  

Abstract

Multi-agent teams benefit from the application of explicit teamwork mechanisms. These alleviate the workload on the designer by automating key components of teamwork, such as communication content and timing decisions, conflict resolution, and joint decision making. However, existing mechanisms make architectural commitments to context-free methods of joint decision making: The agents use the same procedure (e.g., negotiations, voting, reference to rank) to decide on the next step to take, regardless of the status of their task, or the options available. We hypothesize that teams will benefit instead from context-dependent arbitration, where the selection of a joint-decision procedure depends on the status of the task and agents involved. To examine this hypothesis, this paper presents SCORE (Synchronous CoORdination Engine), a prototype teamwork and coordination executable model, that allows the human designer of a team to specify different joint-decision procedures and the conditions under which they are to be used. We evaluate SCORE and context-dependent arbitration in a complex multi-agent 3D virtual environment, using the GameBots interface. We empirically show that, all else being equal, using context-dependent arbitration results in performance which is superior to that resulting from context-free arbitration scheme.

Additional Information

BibTeX

@InProceedings{aicg02, 
  author = 	 {Gal A. Kaminka and Jared Go and Thuc D. Vu}, 
  title = 	 {Context-dependent joint-decision arbitrarion for computer games}, 
  OPTcrossref =  {}, 
  OPTkey = 	 {}, 
  booktitle = {Proceedings of the Agents in Computer Games Workshop}, 
  OPTpages = 	 {}, 
  year = 	 {2002}, 
  abstract = {Multi-agent teams benefit from the application of explicit teamwork 
mechanisms. These alleviate the workload on the designer by automating 
key components of teamwork, such as communication content and timing 
decisions, conflict resolution, and joint decision making. However, 
existing mechanisms make architectural commitments to context-free 
methods of joint decision making: The agents use the same procedure 
(e.g., negotiations, voting, reference to rank) to decide on the next 
step to take, regardless of the status of their task, or the options 
available. We hypothesize that teams will benefit instead from context-dependent 
arbitration, where the selection of a joint-decision procedure depends 
on the status of the task and agents involved. To examine this hypothesis, 
this paper presents SCORE (Synchronous CoORdination Engine), a prototype 
teamwork and coordination executable model, that allows the human 
designer of a team to specify different joint-decision procedures 
and the conditions under which they are to be used. We evaluate SCORE 
and context-dependent arbitration in a complex multi-agent 3D virtual 
environment, using the GameBots interface. We empirically show that, 
all else being equal, using context-dependent arbitration results 
in performance which is superior to that resulting from context-free 
arbitration scheme.}, 
  wwwnote = {}, 
  OPTeditor = 	 {}, 
  OPTvolume = 	 {}, 
  OPTnumber = 	 {}, 
  OPTseries = 	 {}, 
  OPTaddress = 	 {}, 
  OPTmonth = 	 {}, 
  OPTorganization = {}, 
  OPTpublisher = {}, 
  OPTnote = 	 {}, 
  OPTannote = 	 {} 
} 

Generated by bib2html.pl (written by Patrick Riley ) on Mon Nov 16, 2020 22:25:46