The Impact of Potential Crowd Behaviours on Emergency Evacuation: An Evolutionary Game Theoretic Approach

Ibrahim, Azhar Mohd, Venkat, Ibrahim, De Wilde, Philippe (2019) The Impact of Potential Crowd Behaviours on Emergency Evacuation: An Evolutionary Game Theoretic Approach. Journal of Artificial Societies and Social Simulation, 22 (1). ISSN 1460-7425. (doi:10.18564/jasss.3837)

PDF - Author's Accepted Manuscript
Download (253kB) Preview
[img]
Preview
Official URL
http://dx.doi.org/10.18564/jasss.3837

Abstract

Abstract—Crowd dynamics have important applications in evacuation management systems relevant to organizing safer large scale gatherings. For crowd safety, it is very important to study the evolution of potential crowd behaviours by simulating the crowd evacuation process. Planning crowd control tasks via studying the impact of crowd behavioural evolution towards evacuation simulation could mitigate the possibility of crowd disasters that may happen. During a typical emergency evacuation scenario, conflict among agents occurs when agents intend to move to the same location as a result of the interaction of agents within their nearest neighbours. The effect of the agent response towards their neighbourhood is vital in order to understand the effect of variation of crowd behaviours towards the whole environment. In this work, we model crowd motion subject to exit congestion under uncertainty conditions in a continuous space via computer simulations. We model bestresponse, risk-seeking, risk-averse and risk-neutral behaviours of agents via certain game theory notions. We perform computer simulations with heterogeneous populations in order to study the effect of the evolution of agent behaviours towards egress flow under threat conditions. Our simulation results show the relation between the local crowd pressure and the number of injured agents. We observe that when the proportion of agents in a population of risk-seeking agents is increased, the average crowd pressure, average local density and the number of injured agents get increased. Besides that, based on our simulation results, we can infer that crowd disaster could be prevented if the agent population are full of risk-averse and risk-neutral agents despite circumstances that lead to threat consequences.

Item Type: Article
DOI/Identification number: 10.18564/jasss.3837
Uncontrolled keywords: Simulation of dynamic systems; Multi-agent systems; Agent-based models; Dynamic games
Divisions: Faculties > Sciences > School of Computing > Computational Intelligence Group
Depositing User: Philippe De Wilde
Date Deposited: 06 Dec 2018 16:05 UTC
Last Modified: 09 Jul 2019 09:07 UTC
Resource URI: https://kar.kent.ac.uk/id/eprint/70670 (The current URI for this page, for reference purposes)
De Wilde, Philippe: https://orcid.org/0000-0002-4332-1715
  • Depositors only (login required):

Downloads

Downloads per month over past year