Information flow in a Boolean network model of collective behavior

Research output: Contribution to journalArticle

Abstract

In animal groups, leaders have often been proposed to be those individuals who possess additional knowledge about their surroundings, such as the location of a food source or a potential predator. Understanding how this information propagates through the group to shape collective response is an important step to elucidate the evolutionary basis of leadership. In this paper, we study a Boolean model of collective behavior, in which a single leader interacts with a group of followers in a binary decision making process. Through an analytical treatment of the associated Markov chain, we establish closed-form solutions for the transition probability matrix and the stationary distribution, as functions of the noise in the decision making process and the size of the group. We leverage these expressions to quantify information transfer within the group, measured through the information-theoretic construct of transfer entropy. We find that information transfer depends nonlinearly on the group size and noise. For low noise intensities, the system is nearly deterministic, such that no information is shared within the group; an equivalent effect is observed for large noise intensities, which mask the information transfer. We determine the existence of critical noise intensities at which the leader maximizes information transfer to a follower or followers maximize information sharing between each other for a given group size. These analytical findings suggest that noise might have a positive role in collective behavior, facilitating the transfer of knowledge within the group, from leaders to followers.

Original languageEnglish (US)
JournalIEEE Transactions on Control of Network Systems
DOIs
StateAccepted/In press - Oct 21 2017

Fingerprint

Boolean Model
Boolean Networks
Collective Behavior
Information Flow
Network Model
Decision making
Information Transfer
Markov processes
Masks
Animals
Entropy
Decision Making
Maximise
Transition Probability Matrix
Knowledge Transfer
Leadership
Information Sharing
Predator
Stationary Distribution
Closed-form Solution

Keywords

  • Control systems
  • Information theory
  • leadership
  • Markov chain
  • transfer entropy

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Signal Processing
  • Computer Networks and Communications
  • Control and Optimization

Cite this

@article{77738a2dfff74ac9936cfa63a741ffa1,
title = "Information flow in a Boolean network model of collective behavior",
abstract = "In animal groups, leaders have often been proposed to be those individuals who possess additional knowledge about their surroundings, such as the location of a food source or a potential predator. Understanding how this information propagates through the group to shape collective response is an important step to elucidate the evolutionary basis of leadership. In this paper, we study a Boolean model of collective behavior, in which a single leader interacts with a group of followers in a binary decision making process. Through an analytical treatment of the associated Markov chain, we establish closed-form solutions for the transition probability matrix and the stationary distribution, as functions of the noise in the decision making process and the size of the group. We leverage these expressions to quantify information transfer within the group, measured through the information-theoretic construct of transfer entropy. We find that information transfer depends nonlinearly on the group size and noise. For low noise intensities, the system is nearly deterministic, such that no information is shared within the group; an equivalent effect is observed for large noise intensities, which mask the information transfer. We determine the existence of critical noise intensities at which the leader maximizes information transfer to a follower or followers maximize information sharing between each other for a given group size. These analytical findings suggest that noise might have a positive role in collective behavior, facilitating the transfer of knowledge within the group, from leaders to followers.",
keywords = "Control systems, Information theory, leadership, Markov chain, transfer entropy",
author = "Maurizio Porfiri",
year = "2017",
month = "10",
day = "21",
doi = "10.1109/TCNS.2017.2765835",
language = "English (US)",
journal = "IEEE Transactions on Control of Network Systems",
issn = "2325-5870",
publisher = "IEEE CONTROL SYSTEMS SOCIETY",

}

TY - JOUR

T1 - Information flow in a Boolean network model of collective behavior

AU - Porfiri, Maurizio

PY - 2017/10/21

Y1 - 2017/10/21

N2 - In animal groups, leaders have often been proposed to be those individuals who possess additional knowledge about their surroundings, such as the location of a food source or a potential predator. Understanding how this information propagates through the group to shape collective response is an important step to elucidate the evolutionary basis of leadership. In this paper, we study a Boolean model of collective behavior, in which a single leader interacts with a group of followers in a binary decision making process. Through an analytical treatment of the associated Markov chain, we establish closed-form solutions for the transition probability matrix and the stationary distribution, as functions of the noise in the decision making process and the size of the group. We leverage these expressions to quantify information transfer within the group, measured through the information-theoretic construct of transfer entropy. We find that information transfer depends nonlinearly on the group size and noise. For low noise intensities, the system is nearly deterministic, such that no information is shared within the group; an equivalent effect is observed for large noise intensities, which mask the information transfer. We determine the existence of critical noise intensities at which the leader maximizes information transfer to a follower or followers maximize information sharing between each other for a given group size. These analytical findings suggest that noise might have a positive role in collective behavior, facilitating the transfer of knowledge within the group, from leaders to followers.

AB - In animal groups, leaders have often been proposed to be those individuals who possess additional knowledge about their surroundings, such as the location of a food source or a potential predator. Understanding how this information propagates through the group to shape collective response is an important step to elucidate the evolutionary basis of leadership. In this paper, we study a Boolean model of collective behavior, in which a single leader interacts with a group of followers in a binary decision making process. Through an analytical treatment of the associated Markov chain, we establish closed-form solutions for the transition probability matrix and the stationary distribution, as functions of the noise in the decision making process and the size of the group. We leverage these expressions to quantify information transfer within the group, measured through the information-theoretic construct of transfer entropy. We find that information transfer depends nonlinearly on the group size and noise. For low noise intensities, the system is nearly deterministic, such that no information is shared within the group; an equivalent effect is observed for large noise intensities, which mask the information transfer. We determine the existence of critical noise intensities at which the leader maximizes information transfer to a follower or followers maximize information sharing between each other for a given group size. These analytical findings suggest that noise might have a positive role in collective behavior, facilitating the transfer of knowledge within the group, from leaders to followers.

KW - Control systems

KW - Information theory

KW - leadership

KW - Markov chain

KW - transfer entropy

UR - http://www.scopus.com/inward/record.url?scp=85032447033&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85032447033&partnerID=8YFLogxK

U2 - 10.1109/TCNS.2017.2765835

DO - 10.1109/TCNS.2017.2765835

M3 - Article

AN - SCOPUS:85032447033

JO - IEEE Transactions on Control of Network Systems

JF - IEEE Transactions on Control of Network Systems

SN - 2325-5870

ER -