Abstract
The Hajj pilgrimage in Makkah, Saudi Arabia, represents the world's largest annual mass gathering. Ensuring the safety of millions of pilgrims is a paramount concern for Saudi authorities, leading to significant infrastructure projects like the redesign of the Aljamarat Bridge in Mina. While the new multi-level structure has improved crowd flow, the risk of unexpected hazards necessitates robust emergency planning. This paper presents a Discrete Event Simulation (DES) model developed using ExtendSim to evaluate pilgrim evacuation scenarios. Validated against data from three million pilgrims, the simulation demonstrates that a strategic evacuation via six dedicated towers can facilitate a safe, smooth exit process in approximately 10 minutes, preventing dangerous crowd buildup.
Introduction: The Challenge of Mass Gatherings
The Hajj is a week-long pilgrimage involving rituals at several sites, including the Grand Mosque, Mina, Arafat, and Muzdalifah. The Aljamarat Bridge, where pilgrims perform the "stoning of the devil" ritual, has historically been a critical point for crowd management due to high densities and the potential for stampedes. Although the bridge was redesigned into a five-level facility to mitigate these risks, the possibility of fires, gas leaks, or other emergencies remains. Effective evacuation planning requires tools that can predict crowd behavior under stress, accounting for the physical constraints of the bridge and the diverse demographics of the pilgrims.
Methodology: Discrete Event Simulation
To model these complex dynamics, researchers utilized Discrete Event Simulation (DES) via the ExtendSim software. Unlike agent-based models that simulate individual micro-behaviors, DES focuses on the logical flow of entities (pilgrim groups) through a sequence of activities.
Model Components
The simulation incorporates several layers of complexity:
- Structural Logic: The model replicates the five-level bridge, featuring 11 entry points, 12 standard exits, and 6 specialized evacuation towers.
- Multi-Attribute Entities: Pilgrims are assigned specific attributes, including age (which dictates walking speed), region of origin (linked to historical incident data), and fitness levels (categorized as fit, tired, or very tired).
- Stochastic Processes: To account for real-world variability, ritual durations (stoning and prayer) are modeled using triangular distributions. Walking times include random components to reflect heterogeneity within the crowd.
Data and Validation
The model was calibrated using secondary data from the Institute of Hajj and Umrah Research and the Ministry of Hajj. Validation involved comparing simulation outputs against statistics and video recordings from the 2015, 2016, 2017, and 2019 Hajj seasons. The results for ritual duration and level utilization showed high accuracy, with less than a 1% difference from real-world data observed in the 2019 season.
Key Insights and Results
The simulation provided critical insights into both daily operations and emergency scenarios:
Operational Efficiency
The study confirmed that the current strict scheduling system is highly effective. Under normal conditions, the entire three-million-person event can be processed in 22 hours, with individual groups completing the ritual in 12–14 minutes. Furthermore, the model suggests that distributing pilgrims more evenly across all five levels could further optimize capacity, potentially reducing ritual duration by an additional 0.4 minutes per group.
Evacuation Performance
The primary focus of the study was the efficiency of emergency evacuations. The researchers tested a strategy directing pilgrims to six dedicated evacuation towers rather than standard exits.
Key findings include:
- Speed: In a worst-case scenario with 100,000 pilgrims on the bridge, evacuation to the nearest towers takes approximately 10 to 10.5 minutes.
- Flow Management: The tower-based approach prevents the intersection of opposing flows, allowing for a "balanced" use of facilities and maintaining a smooth process without crowd buildup.
- Resilience: While random (Poisson) arrival patterns increased the instantaneous number of people on the bridge compared to scheduled arrivals, evacuation times remained stable as long as all six towers remained operational.
Conclusion and Future Directions
The DES model proves to be a robust tool for Hajj safety planning, validating the effectiveness of current scheduling and highlighting the life-saving potential of dedicated evacuation towers. However, the authors note that the model currently assumes all towers are functional and that pilgrims follow instructions perfectly. Future research will explore the impact of tower closures and incorporate "irrational" human behaviors or communication delays to further refine emergency protocols. Additionally, future work will focus on the first day of the stoning ritual, which presents a unique challenge due to a tighter 12-hour time window.
Summary updated