Finite neighborhood binary games: A structural study

Jijun Zhao, Ferenc Szidarovszky, Miklos N Szilagyi

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

The purpose of this study is to present a systematic analysis of the long-term behavior of the agents of an artificial society under varying payoff functions in finite neighborhood binary games. By assuming the linearity of the payoffs of both cooperating and defecting agents, the type of the game is determined by four fundamental parameters. By fixing the values of three of them and systematically varying the fourth one we can observe a transition from Prisoner's Dilemma to Leader Game through Chicken and Benevolent Chicken Games. By using agent-based simulation we are able to observe the long-term behavior of the artificial society with different and gradually changing payoff structure. The difference between different games is explored and the effect of the transition from one game to the other on the society is investigated. The results depend on the personality types of the agents. In this study greedy and Pavlovian agents are considered. In the first case, we observe the most significant change in trajectory structure between Prisoner's Dilemma and Chicken Games showing significant difference in the behavioral patterns of the agents. Almost no changes can be observed between Benevolent Chicken and Leader Games, and only small change between Chicken and Benevolent Chicken. The trajectories change from always converging to regularly oscillating patterns with systematically altering amplitude and central values. The results are very similar whether the agents consider themselves as members of their neighborhoods or not. With Pavlovian agents no significant difference can be observed between the four games, the trajectories always converge and the limits smoothly and monotonically depend on the value of the varying parameter.

Original languageEnglish (US)
JournalJASSS
Volume10
Issue number3
StatePublished - Jun 2007
Externally publishedYes

Fingerprint

prisoner
leader
Values
Trajectories
personality
simulation
Society

Keywords

  • Agent-based simulation
  • Equilibrium
  • N-person games
  • Structure analysis

ASJC Scopus subject areas

  • Social Sciences(all)

Cite this

Finite neighborhood binary games : A structural study. / Zhao, Jijun; Szidarovszky, Ferenc; Szilagyi, Miklos N.

In: JASSS, Vol. 10, No. 3, 06.2007.

Research output: Contribution to journalArticle

Zhao, J, Szidarovszky, F & Szilagyi, MN 2007, 'Finite neighborhood binary games: A structural study', JASSS, vol. 10, no. 3.
Zhao, Jijun ; Szidarovszky, Ferenc ; Szilagyi, Miklos N. / Finite neighborhood binary games : A structural study. In: JASSS. 2007 ; Vol. 10, No. 3.
@article{0dabc3cc726a49de851dabc90ed02adb,
title = "Finite neighborhood binary games: A structural study",
abstract = "The purpose of this study is to present a systematic analysis of the long-term behavior of the agents of an artificial society under varying payoff functions in finite neighborhood binary games. By assuming the linearity of the payoffs of both cooperating and defecting agents, the type of the game is determined by four fundamental parameters. By fixing the values of three of them and systematically varying the fourth one we can observe a transition from Prisoner's Dilemma to Leader Game through Chicken and Benevolent Chicken Games. By using agent-based simulation we are able to observe the long-term behavior of the artificial society with different and gradually changing payoff structure. The difference between different games is explored and the effect of the transition from one game to the other on the society is investigated. The results depend on the personality types of the agents. In this study greedy and Pavlovian agents are considered. In the first case, we observe the most significant change in trajectory structure between Prisoner's Dilemma and Chicken Games showing significant difference in the behavioral patterns of the agents. Almost no changes can be observed between Benevolent Chicken and Leader Games, and only small change between Chicken and Benevolent Chicken. The trajectories change from always converging to regularly oscillating patterns with systematically altering amplitude and central values. The results are very similar whether the agents consider themselves as members of their neighborhoods or not. With Pavlovian agents no significant difference can be observed between the four games, the trajectories always converge and the limits smoothly and monotonically depend on the value of the varying parameter.",
keywords = "Agent-based simulation, Equilibrium, N-person games, Structure analysis",
author = "Jijun Zhao and Ferenc Szidarovszky and Szilagyi, {Miklos N}",
year = "2007",
month = "6",
language = "English (US)",
volume = "10",
journal = "Journal of Artificial Societies and Social Simulation",
issn = "1460-7425",
publisher = "University of Surrey",
number = "3",

}

TY - JOUR

T1 - Finite neighborhood binary games

T2 - A structural study

AU - Zhao, Jijun

AU - Szidarovszky, Ferenc

AU - Szilagyi, Miklos N

PY - 2007/6

Y1 - 2007/6

N2 - The purpose of this study is to present a systematic analysis of the long-term behavior of the agents of an artificial society under varying payoff functions in finite neighborhood binary games. By assuming the linearity of the payoffs of both cooperating and defecting agents, the type of the game is determined by four fundamental parameters. By fixing the values of three of them and systematically varying the fourth one we can observe a transition from Prisoner's Dilemma to Leader Game through Chicken and Benevolent Chicken Games. By using agent-based simulation we are able to observe the long-term behavior of the artificial society with different and gradually changing payoff structure. The difference between different games is explored and the effect of the transition from one game to the other on the society is investigated. The results depend on the personality types of the agents. In this study greedy and Pavlovian agents are considered. In the first case, we observe the most significant change in trajectory structure between Prisoner's Dilemma and Chicken Games showing significant difference in the behavioral patterns of the agents. Almost no changes can be observed between Benevolent Chicken and Leader Games, and only small change between Chicken and Benevolent Chicken. The trajectories change from always converging to regularly oscillating patterns with systematically altering amplitude and central values. The results are very similar whether the agents consider themselves as members of their neighborhoods or not. With Pavlovian agents no significant difference can be observed between the four games, the trajectories always converge and the limits smoothly and monotonically depend on the value of the varying parameter.

AB - The purpose of this study is to present a systematic analysis of the long-term behavior of the agents of an artificial society under varying payoff functions in finite neighborhood binary games. By assuming the linearity of the payoffs of both cooperating and defecting agents, the type of the game is determined by four fundamental parameters. By fixing the values of three of them and systematically varying the fourth one we can observe a transition from Prisoner's Dilemma to Leader Game through Chicken and Benevolent Chicken Games. By using agent-based simulation we are able to observe the long-term behavior of the artificial society with different and gradually changing payoff structure. The difference between different games is explored and the effect of the transition from one game to the other on the society is investigated. The results depend on the personality types of the agents. In this study greedy and Pavlovian agents are considered. In the first case, we observe the most significant change in trajectory structure between Prisoner's Dilemma and Chicken Games showing significant difference in the behavioral patterns of the agents. Almost no changes can be observed between Benevolent Chicken and Leader Games, and only small change between Chicken and Benevolent Chicken. The trajectories change from always converging to regularly oscillating patterns with systematically altering amplitude and central values. The results are very similar whether the agents consider themselves as members of their neighborhoods or not. With Pavlovian agents no significant difference can be observed between the four games, the trajectories always converge and the limits smoothly and monotonically depend on the value of the varying parameter.

KW - Agent-based simulation

KW - Equilibrium

KW - N-person games

KW - Structure analysis

UR - http://www.scopus.com/inward/record.url?scp=34547284217&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34547284217&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:34547284217

VL - 10

JO - Journal of Artificial Societies and Social Simulation

JF - Journal of Artificial Societies and Social Simulation

SN - 1460-7425

IS - 3

ER -