Agent-based Crowd Simulation, Analysis and Optimization for Pedestrians flow during Hajj
Hajj is one of the largest mass gatherings where Muslims from all over the world gather in Makah each year for pilgrimage. Such large gatherings always have risk of confrontation with disasters, either natural or man-made. In order to perform Hajj rituals, Pilgrims move around the various religious sites. In the past few years, thousands of casualties have occurred due to stampede or chaos during different Hajj rituals, especially during the Circumambulation of Kaba (Tawaf) and stoning the devil in Mina. We have to analyze and propose an appropriate evacuation strategy to make these events safe and risk free and to alleviate the risk of causalities. However, it is a daunting research problem to identify optimal course of actions during emergency disaster management when several constraints are involved, especially when dealing with a crowd of massive size.
We propose an Agent-based Crowd Simulation & Analysis framework that incorporates Anylogic space markup and pedestrian library to design spatial environment and simulate large-scale crowd movements to evaluate different evacuation strategies for mass gatherings. Hence provides a run time environment for analyzing complex situations, e.g., emergency evacuation strategies. The key features of proposed framework includes (i) design of spatial environment and large-scale crowd simulation model (ii) provides interoperability to integrate simulation model with external modules to run optimizations (iii) ability to implement and integrate different optimization algorithms with simulation model in order to determine the optimal evacuation plan. We present a case study of Hajj scenario as a proof of concept and a test bed for identifying and evaluating optimal strategies for crowd evacuation.
Our proposed framework is composed of three layers:
1- Simulation Layer
2- Interface Layer
3- Optimization and Analysis Layer
Simulation Layer
In this layer, we propose the use of Anylogic simulation environment using which a modeler can design spatial environments for any large crowd. The Space markup library provides necessary elements for developing a real-scale graphics model of a continuous space. It also provides import feature for AutoCAD floor plans or GIS shape files. Using this feature the spatial geometry can be modeled precisely up to the scale of a real-world environment. Finally, the modeler is required to create a logical model of crowd using the Anylogic Pedestrian library (APL). APL provides friendly interface for robust development of pedestrian models in the form of flowcharts. It permits making pedestrian buildings models e.g. subway stations and concert halls. The library can import previously created agent population and the spatial environment and implements the flow of pedestrian as per physical rules defined by the modeler.
We present a case study of Hajj scenario as a proof of concept and a test bed for identifying and evaluating optimal strategies for crowd evacuation.
Figure given below shows the spatial environment of Masjid-Al-Haram, where black border lines are denoting the walls and 12 entrances and exits are depicted in green color.
Interface Layer
This layer enables the integration & interoperability of Simulation layer with external programs. We develop a Java based middleware that interfaces with the Simulation layer and allows the user to expose input parameters of the Simulation model as public variables which can be configured outside the simulation layer.
Optimization and Analysis Layer
The main purpose of this layer is to allow users to implement and integrate optimization algorithms. This layer allows flexibility and independence for using any optimization approach and utilizes the interface layer for interacting with Simulation model in a variety of ways
