The effect of dispersal and neighbourhood in games of cooperation

S. Számadó, Ferenc Szalai, I. Scheuring

Research output: Contribution to journalArticle

23 Citations (Scopus)

Abstract

The prisoner's dilemma (PD) and the snowdrift (SD) games are paradigmatic tools to investigate the origin of cooperation. Whereas spatial structure (e.g. nonrandom spatial distribution of strategies) present in the spatially explicit models facilitates the emergence of cooperation in the PD game, recent investigations have suggested that spatial structure can be unfavourable for cooperation in the SD game. The frequency of cooperators in a spatially explicit SD game can be lower than it would be in an infinitely large well-mixed population. However, the source of this effect cannot be identified with certainty as spatially explicit games differ from well-mixed games in two aspects: (i) they introduce spatial correlations, (ii) and limited neighbourhood. Here we extend earlier investigations to identify the source of this effect, and thus accordingly we study a spatially explicit version of the PD and SD games with varying degrees of dispersal and neighbourhood size. It was found that dispersal favours selfish individuals in both games. We calculated the frequency of cooperators at strong dispersal limit, which in concordance with the numerical results shows that it is the short range of interactions (i.e. limited neighbourhood) and not spatial correlations that decreases the frequency of cooperators in spatially explicit models of populations. Our results demonstrate that spatial correlations are always beneficial to cooperators in both the PD and SD games. We explain the opposite effect of dispersal and neighbourhood structure, and discuss the relevance of distinguishing the two effects in general.

Original languageEnglish
Pages (from-to)221-227
Number of pages7
JournalJournal of Theoretical Biology
Volume253
Issue number2
DOIs
Publication statusPublished - Jul 21 2008

Fingerprint

Game
Prisoners' Dilemma
Spatial distribution
Spatial Correlation
Spatial Structure
spatial distribution
Population
Prisoner's Dilemma Game
Prisoner Dilemma
Concordance
Spatial Distribution
Numerical Results
Decrease
Interaction
Model
Range of data
Demonstrate

Keywords

  • Cooperation
  • Finite population
  • Hawk-doves
  • Prisoner's dilemma
  • Snowdrift game
  • Spatial structure

ASJC Scopus subject areas

  • Agricultural and Biological Sciences(all)

Cite this

The effect of dispersal and neighbourhood in games of cooperation. / Számadó, S.; Szalai, Ferenc; Scheuring, I.

In: Journal of Theoretical Biology, Vol. 253, No. 2, 21.07.2008, p. 221-227.

Research output: Contribution to journalArticle

@article{75032bd57c674d04b8212fd749247954,
title = "The effect of dispersal and neighbourhood in games of cooperation",
abstract = "The prisoner's dilemma (PD) and the snowdrift (SD) games are paradigmatic tools to investigate the origin of cooperation. Whereas spatial structure (e.g. nonrandom spatial distribution of strategies) present in the spatially explicit models facilitates the emergence of cooperation in the PD game, recent investigations have suggested that spatial structure can be unfavourable for cooperation in the SD game. The frequency of cooperators in a spatially explicit SD game can be lower than it would be in an infinitely large well-mixed population. However, the source of this effect cannot be identified with certainty as spatially explicit games differ from well-mixed games in two aspects: (i) they introduce spatial correlations, (ii) and limited neighbourhood. Here we extend earlier investigations to identify the source of this effect, and thus accordingly we study a spatially explicit version of the PD and SD games with varying degrees of dispersal and neighbourhood size. It was found that dispersal favours selfish individuals in both games. We calculated the frequency of cooperators at strong dispersal limit, which in concordance with the numerical results shows that it is the short range of interactions (i.e. limited neighbourhood) and not spatial correlations that decreases the frequency of cooperators in spatially explicit models of populations. Our results demonstrate that spatial correlations are always beneficial to cooperators in both the PD and SD games. We explain the opposite effect of dispersal and neighbourhood structure, and discuss the relevance of distinguishing the two effects in general.",
keywords = "Cooperation, Finite population, Hawk-doves, Prisoner's dilemma, Snowdrift game, Spatial structure",
author = "S. Sz{\'a}mad{\'o} and Ferenc Szalai and I. Scheuring",
year = "2008",
month = "7",
day = "21",
doi = "10.1016/j.jtbi.2008.02.037",
language = "English",
volume = "253",
pages = "221--227",
journal = "Journal of Theoretical Biology",
issn = "0022-5193",
publisher = "Academic Press Inc.",
number = "2",

}

TY - JOUR

T1 - The effect of dispersal and neighbourhood in games of cooperation

AU - Számadó, S.

AU - Szalai, Ferenc

AU - Scheuring, I.

PY - 2008/7/21

Y1 - 2008/7/21

N2 - The prisoner's dilemma (PD) and the snowdrift (SD) games are paradigmatic tools to investigate the origin of cooperation. Whereas spatial structure (e.g. nonrandom spatial distribution of strategies) present in the spatially explicit models facilitates the emergence of cooperation in the PD game, recent investigations have suggested that spatial structure can be unfavourable for cooperation in the SD game. The frequency of cooperators in a spatially explicit SD game can be lower than it would be in an infinitely large well-mixed population. However, the source of this effect cannot be identified with certainty as spatially explicit games differ from well-mixed games in two aspects: (i) they introduce spatial correlations, (ii) and limited neighbourhood. Here we extend earlier investigations to identify the source of this effect, and thus accordingly we study a spatially explicit version of the PD and SD games with varying degrees of dispersal and neighbourhood size. It was found that dispersal favours selfish individuals in both games. We calculated the frequency of cooperators at strong dispersal limit, which in concordance with the numerical results shows that it is the short range of interactions (i.e. limited neighbourhood) and not spatial correlations that decreases the frequency of cooperators in spatially explicit models of populations. Our results demonstrate that spatial correlations are always beneficial to cooperators in both the PD and SD games. We explain the opposite effect of dispersal and neighbourhood structure, and discuss the relevance of distinguishing the two effects in general.

AB - The prisoner's dilemma (PD) and the snowdrift (SD) games are paradigmatic tools to investigate the origin of cooperation. Whereas spatial structure (e.g. nonrandom spatial distribution of strategies) present in the spatially explicit models facilitates the emergence of cooperation in the PD game, recent investigations have suggested that spatial structure can be unfavourable for cooperation in the SD game. The frequency of cooperators in a spatially explicit SD game can be lower than it would be in an infinitely large well-mixed population. However, the source of this effect cannot be identified with certainty as spatially explicit games differ from well-mixed games in two aspects: (i) they introduce spatial correlations, (ii) and limited neighbourhood. Here we extend earlier investigations to identify the source of this effect, and thus accordingly we study a spatially explicit version of the PD and SD games with varying degrees of dispersal and neighbourhood size. It was found that dispersal favours selfish individuals in both games. We calculated the frequency of cooperators at strong dispersal limit, which in concordance with the numerical results shows that it is the short range of interactions (i.e. limited neighbourhood) and not spatial correlations that decreases the frequency of cooperators in spatially explicit models of populations. Our results demonstrate that spatial correlations are always beneficial to cooperators in both the PD and SD games. We explain the opposite effect of dispersal and neighbourhood structure, and discuss the relevance of distinguishing the two effects in general.

KW - Cooperation

KW - Finite population

KW - Hawk-doves

KW - Prisoner's dilemma

KW - Snowdrift game

KW - Spatial structure

UR - http://www.scopus.com/inward/record.url?scp=45649084224&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=45649084224&partnerID=8YFLogxK

U2 - 10.1016/j.jtbi.2008.02.037

DO - 10.1016/j.jtbi.2008.02.037

M3 - Article

C2 - 18440558

AN - SCOPUS:45649084224

VL - 253

SP - 221

EP - 227

JO - Journal of Theoretical Biology

JF - Journal of Theoretical Biology

SN - 0022-5193

IS - 2

ER -