Article Preview
Top1. Introduction
The regular gatherings of people in large public places (i.e., hospitals, shopping centers, or university campuses) raise several issues, including safety in the case of abnormal situations like a fire or terrorist attack. Due to the limited use of indoor global positioning systems (GPS), a smooth evacuation of a building is a challenge. The process and the timing of an evacuation impact injuries and their severity.
Links between people include social relationships, such as family, friends, or colleagues. These individuals tend to walk together in a group (Xie et al., 2020). In an emergency, some people are more likely to take on the leadership role. These leaders are often independent individuals who can maintain their composure and help others in an emergency (Pelechano & Badler, 2006). On the other hand, some people rely on others to help them out of an emergency rather than make their own decisions in issues like emergency exits. Here, the evacuee’s behaviours in social groups differ from that of a single individual (Haghani et al., 2019).
Crociani et al. (2018) and Von Krüchten and Schadschneider (2017) investigated how evacuation decisions are made within groups, how decisions within groups impact the evacuation process and the impact of individual choices of evacuees. This information may play a significant role in improving the accuracy of evacuation models. Furthermore, in various models, such as the follower-leader model (Fang et al., 2016), the follower will follow the leader with no specific personality. Other models include Cellular Automata (CA). The use of square meshes in CA has subjected models to errors due to an incorrect choice of map projection or directional bias (Clarke, 2014). It aims to generate quantitative or discrete results about pedestrian and crowd movements (Chenet et al., 2020; Kalogeiton et al., 2015; Mirahadi et al., 2019; Yang et al., 2005). Social Force Model (SFM) is widely used in crowd evacuation simulations. First, it considers the impact of other users or components on user movements. Second, it defines the impacts as social forces (Haghani et al., 2019; Yıldız&Çağdaş, 2020). Agent-Based Model (ABM) consists of a set of autonomous and intelligent agents that can perceive and interact with their environment to solve problems, achieve goals, or execute tasks (Janssen et al., 2020). These features make it a promising technique to address emergency evacuation problems.
Motivated by the above, the authors propose a dynamic and smart evacuation process in a building model with multi-exit of various sizes.
The approach was made possible by hybridizing the ABM, SFM, and psychological behaviour with emotions and norms (BEN). The goal is to replicate our world's interconnected networks. The approach exploits network communication messages based on the message queuing telemetry transport (MQTT) protocol to inform a group leader about hazards through sensors in the building. The leader can then use expert system rules (ESR) to choose an exit depending on congestion, preference, and hazard progression. Simulations are done using the GAMA platform, including effective social agent behaviour and modeling (https://www.qgis.org/en/site/about/index.html).