Dynamic evolution model of pedestrian cooperation behavior based on coordination game. (1st April 2021)

• Record Type: Journal Article
• Authors: Niu, Yunyun; Chen, Yulin; Yuan, Bo; Xiao, Jianhua
• Is Part Of: Expert systems with applications. Volume 167(2021)
• Publication Date: 2021-04-01
• Subjects: Evacuation process -- Dynamic network -- Coordination game -- Evolution of cooperation -- Replicator dynamics; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2020.114173
• Languages: English
• ISSNs: 0957-4174

Abstract

Abstract: Game theory was widely used to model the decision making strategy of pedestrians in the process of emergency evacuation. However, the dynamic evolution of cooperation relationships during this process has rarely been studied. Here we propose a spatial evacuation model of pedestrians with replicator dynamics and coordination game to address this problem. The novelty of this paper is to study the evolution of pedestrians' cooperative behavior based on the coevolution of coordination game and evacuee interactive network. A pedestrian game network is established with different strategy sets guiding two types of pedestrians, i.e., cooperators and defectors. The dynamic network theory is used to analyze the changing social relationship between pedestrians, while the coordination game is adopted to model the decision-making process of pedestrians. Rules initiated from replicator dynamics are implemented to update strategies during the evacuation process. The simulation results show that the strategy evolution during evacuation vary with the panic level, the proportion of initial cooperators, and the pedestrian flow density. Especially, a higher proportion of initial cooperators make the system evolve toward cooperation, while high panic level and high pedestrian density make the system evolve toward defection. The effect of crucial parameters on evacuation efficiency is also discussed. Highlights: A new evacuation model with the replication dynamics and game theory is established. Pedestrian evacuation time is independent of the maximum average payoff. Pedestrian evacuation time is positively correlated with the game scope. A higher proportion of initial cooperators make the system evolve toward cooperation. High panic level and high pedestrian density make the system evolve toward defection.