Two Agents Acting as One

Malte Harder; D. Polani; C.L. Nehaniv

Two Agents Acting as One

Malte Harder
, D. Polani
, C.L. Nehaniv

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

4 Citations (Scopus)

51 Downloads (Pure)

Abstract

We consider two agents, each equipped with a controller. When they achieve a joint goal configuration, their coordination can be measured informationally. We show that the amount of coordination that two agents need to configure in
a certain way depends on the amount of information they obtain from their environment. Furthermore the environment imposes a coordination pressure on the agents that depends on the size of the environment. In a second scenario we introduce a shared centralized controller which leads to a synchronisation
of the agents’ actions for suboptimal policies. However, in the optimal case this intrinsic coordination vanishes and the shared centralized controller can be split into
two individual controllers.

Original language	English
Title of host publication	Procs of Artificial Life XII
Editors	Harold Fellermann, Mark Dorr, Martin M. Hanczyc
Publisher	MIT Press
Pages	599-606
Number of pages	8
ISBN (Print)	0-262-29075-8, 978-0-262-29075-3
Publication status	Published - 2010
Event	Artificial Life XII - Odense, Denmark Duration: 19 Aug 2010 → 23 Aug 2010

Conference

Conference	Artificial Life XII
Country/Territory	Denmark
City	Odense
Period	19/08/10 → 23/08/10

Access to Document

904689Final published version, 958 KB

Cite this

@inproceedings{87acfcc3d7c448d7a40b0fcdaab13095,

title = "Two Agents Acting as One",

abstract = "We consider two agents, each equipped with a controller. When they achieve a joint goal configuration, their coordination can be measured informationally. We show that the amount of coordination that two agents need to configure in a certain way depends on the amount of information they obtain from their environment. Furthermore the environment imposes a coordination pressure on the agents that depends on the size of the environment. In a second scenario we introduce a shared centralized controller which leads to a synchronisation of the agents{\textquoteright} actions for suboptimal policies. However, in the optimal case this intrinsic coordination vanishes and the shared centralized controller can be split into two individual controllers.",

author = "Malte Harder and D. Polani and C.L. Nehaniv",

year = "2010",

language = "English",

isbn = "0-262-29075-8",

pages = "599--606",

editor = "Harold Fellermann and Mark Dorr and Hanczyc, \{Martin M.\}",

booktitle = "Procs of Artificial Life XII",

publisher = "MIT Press",

address = "United States",

note = "Artificial Life XII ; Conference date: 19-08-2010 Through 23-08-2010",

}

TY - GEN

T1 - Two Agents Acting as One

AU - Harder, Malte

AU - Polani, D.

AU - Nehaniv, C.L.

PY - 2010

Y1 - 2010

N2 - We consider two agents, each equipped with a controller. When they achieve a joint goal configuration, their coordination can be measured informationally. We show that the amount of coordination that two agents need to configure in a certain way depends on the amount of information they obtain from their environment. Furthermore the environment imposes a coordination pressure on the agents that depends on the size of the environment. In a second scenario we introduce a shared centralized controller which leads to a synchronisation of the agents’ actions for suboptimal policies. However, in the optimal case this intrinsic coordination vanishes and the shared centralized controller can be split into two individual controllers.

AB - We consider two agents, each equipped with a controller. When they achieve a joint goal configuration, their coordination can be measured informationally. We show that the amount of coordination that two agents need to configure in a certain way depends on the amount of information they obtain from their environment. Furthermore the environment imposes a coordination pressure on the agents that depends on the size of the environment. In a second scenario we introduce a shared centralized controller which leads to a synchronisation of the agents’ actions for suboptimal policies. However, in the optimal case this intrinsic coordination vanishes and the shared centralized controller can be split into two individual controllers.

M3 - Conference contribution

SN - 0-262-29075-8

SN - 978-0-262-29075-3

SP - 599

EP - 606

BT - Procs of Artificial Life XII

A2 - Fellermann, Harold

A2 - Dorr, Mark

A2 - Hanczyc, Martin M.

PB - MIT Press

T2 - Artificial Life XII

Y2 - 19 August 2010 through 23 August 2010

ER -

Two Agents Acting as One

Abstract

Conference

Access to Document

Fingerprint

Cite this