An agent-based simulation system for concert venue crowd evacuation modeling in the presence of a fire disaster

Highlights

• Paper presents a prototype agent based simulation system for concert venue crowd evacuation.

• Prototype developed provides basis for testing of multiple disaster scenarios at virtually no cost.

• Prototype is unique in literature as it is designed to simulate crowd evacuation of concert venues.

• Prototype is highly configurable allowing users to define seats, pathways, and other parameters.

• Prototype presented in this paper is designed specifically for the real-world applications.

Abstract

A key activity in emergency management is planning and preparation for disaster. If the right safety measures are implemented beforehand, harmful effects can be significantly mitigated. However, evaluation and selection of effective measures is difficult due to the numerous scenarios that exist in most emergency environments coupled with the high associated cost of testing such scenarios. An agent-based system employs a computational model of autonomous interacting agents in an environment with the purpose of assessing the emergent behavior of the group. This paper presents a prototype of a computer simulation and decision support system that uses agent-based modeling to simulate crowd evacuation in the presence of a fire disaster and provides for testing of multiple disaster scenarios at virtually no cost. The prototype is unique in the current literature as it is specifically designed to simulate a concert venue setting such as a stadium or auditorium and is highly configurable allowing for user definition of concert venues with any arrangement of seats, pathways, stages, exits, and people as well as the definition of multiple fires with fire and smoke dynamics included.

Introduction

Of paramount importance to emergency managers is the question of how to prepare for as yet unseen disasters. Proper safety measures can literally mean the difference between life and death for large groups of affected people. However, emergency situations and their associated safety measures are highly specific to the environment in which they exist and there are generally numerous scenarios that must be considered. The cost of testing these multiple scenarios is oftentimes prohibitive (Jain & McLean, 2008). Thus, evaluation and selection of effective safety measures for emergency preparedness is quite difficult and is often left to the subjective judgment of an emergency manager.

Computer modeling and simulation seeks to remedy this problem by allowing for testing of multiple environment-specific scenarios at low cost. Agent-based systems use a computational model of autonomous agents that move and interact with each other and their environment. Such systems use a bottom-up modeling approach in which system control is decentralized and governed only by the behavior of the agents (Borshchev & Filippov, 2004). Agent-based modeling is the preferable technique for simulation of systems with a large number of active objects (e.g., people, business units, animals, etc.) that are dependent on the order/timing of events for the following reasons: (1) it allows for the capture of highly complex dynamics, (2) it can be implemented with little or no knowledge of the global interdependencies and/or aggregate effects of the system, and (3) it is easier to build upon as model changes generally require local not global adjustments (Borshchev & Filippov, 2004). The development of agent-based systems for emergency planning and preparedness remains an open research area as there exist a multitude of disaster environments that have yet to be addressed (Jain & McLean, 2008).

This paper presents a prototype of an Agent-based Decision Support System (ABS) for the simulation of crowd evacuation in the presence of a fire disaster for venues that are specifically intended for mass gatherings such as stadiums and auditoriums. The goal of the system is to allow for multiple scenario testing and decision support for the planning and preparedness phase of emergency management with regards to fire disasters at concert venues. The system is designed for emergency managers, police, and any administrators who are charged with fire disaster mitigation planning for concert venues. Users of the system can benefit by evaluating the effects of potential safety measures such as restrictions on the maximum number of people, wider pathways, additional exits, and fewer seats on crowd evacuation dynamics. The system is unique as it is specifically designed to simulate evacuation of a concert venue setting rather than an urban roadways or building evacuation setting as is prevalent in the literature. High densities of people and relatively limited exit routes and exit points are common characteristics of concert venues and their combination make such venues a significant concern for emergency managers. Additionally, the ABS system is highly configurable allowing for user definition of a concert venue with any number and arrangement of seats and bleachers, aisles and path ways, stages and playing fields, exits, and people and also allows for the definition of multiple fires with dynamics of fire spreading and smoke production included. The contribution of this study is twofold:

• It provides an agent-based system that is specifically designed for crowd evacuation simulation of concert venues during a fire disaster.

• The system is built for customization and provides to the user the ability to define the layout and structure of the concert venue to be simulated. This allows the user to replicate the venue of concern and provides decision support for the planning and preparedness phases of emergency management.

The rest of this paper is organized as follows: Section 2 provides a brief survey of the current research on agent-based systems for crowd evacuation modeling, Section 3 gives a description of the prototype ABS system, Section 4 details experiments conducted using the system to simulate disaster scenarios for simulated replicas of actual concert venues, and Section 5 discusses future work necessary to enhance and transition this prototype system into a viable commercial software system.