Goodbye to the Hajj Scheduling Project

From 2006 - 2014 and 2016 - 2018, the research team I was part of developed operation plans for the great Islamic pilgrimage towards Mecca. These plans were mathematically optimized for crowd safety and applied by the Hajj authorities. With currently more than two million pilgrims the Hajj is one of the largest mega-events on the planet. 

The Hajj season 2018 was our final contribution to the safety concept of Hajj. Specifically, we are not part of any preparation activities for Hajj 2019 and have no plans to be involved in Hajj planning going forward. In 2019 the central rituals start on Friday, August 9 and end on Wednesday, August 14. 

Hooray! On May 27^th, 2019, I successfully defended my PhD thesis at the University of Hamburg. Finally wrapping up my research project in scientific journals [1] marks the end of an adventurous expedition into an admittedly unfamiliar world. In the sleepless nights along this journey, I have learned a lot about dealing with pressure, about the limits of applying optimization in the real world, and about the attention to detail required to put scientific planning out into the real world. I would think that sometimes researchers, who have mastered the art of solving problems in an abstracted form, tend to underestimate the pitfalls of practical implementation. Especially in the field of crowd management, where even the slightest lack of attention can quickly escalate into a full-blown disaster, only a healthy mix of experienced managers and scientific expertise promises success. And most importantly, both ends need to accept and appreciate their mutual dependence. 

In 2006, the research team I have been part of was welcomed by security forces, crowd managers, and infrastructure engineers to form a joint planning committee seeking to re-think every aspect of Hajj planning.

Our objective was challenging: to plan crowd movements in one of the largest and most difficult to control mega-events on the planet. Set in a desert environment hostile to human life, the Hajj attracts millions of faithful who attempt to perform the required religious rituals once in their lifetime.

We began our planning by digitizing the local pedestrian infrastructure such as roads, walkways, stairs, escalators, and tunnels as well as camps, street blocks, and sanitary facilities.

Based on this digital infrastructure we developed a pre-event simulation of how the crowd flows through the infrastructure and finally designed a mathematical planning model and algorithm to determine safe pilgrim group departure times and routes for the stoning-of-the-devil ritual. Year for year, we were guests in Saudi-Arabia and gained a deeper understanding of the Hajj organizational structure. With growing experience, we refined our model-based planning and improved our understanding of how to structure and process the massive amount of input-data.

The magnitude of the Hajj planning problem is remarkable. During the stoning-ritual more than two million pilgrims approach three pillars to symbolically stone the temptations of the devil. In the pre-2006 era, the collective pilgrim behavior in combination with inadequate crowd management repeatedly caused overcrowding, resulting in large accidents that killed hundreds – and sometimes thousands- of people.

To mitigate the risk of the reoccurring disasters, the infrastructure and the organization of the Hajj underwent a complete overhaul. Our contribution was to support the dispatching operation using mathematical modelling and simulation – basic building blocks of typical applications in Operational Research.

And I would argue, we succeeded in providing safe operation plans for the event in every single year of our engagement. Our plans were implemented and distributed to the Hajji field guides from 2007 until 2014 and again from 2016 until 2018.

The development of safe timetables and routings is very challenging, because the traditional Hajj dates constitute a hard deadline. Any operation plan delivered too late is of zero value. However, some of the fundamental data in our approach, such as the spatial distribution of pilgrim contingents over the tent city or the allocation of metro ticket holders (about 400,000) change even days before Hajj. In the hot planning phase, we tracked required input data night and day in meetings, phone calls and emails – sometimes in spreadsheets, sometimes scribbled with pencil on a map, sometimes recorded via phone in a car. During these times, it was not uncommon to receive a call in the morning following a nightly planning session, to inform us at a moment’s notice about 10 thousand additional pilgrims to be added into the schedules. However, such changes come with operational risks: a significant change in one part of a balanced schedule results in a chain of consecutive changes at other parts of the schedule. But all these updates finally need to arrive at the field guides who also require some training to read and apply the schedules.

Additionally, security forces develop detailed operation plans that prescribe opening and closing times for specific pedestrian passages (roads, tunnels, bridges, escalators etc.) and intended movement directions for specific infrastructure. We incorporated such operational details when planning the crowd flows. Again, some of these details were subject to complete overhaul within one of the many last-minute preparation workshops and needed to be synced with our abstract, mathematical planning approach.

Our cooperation with the Saudi authorities for the stoning-rituals ended in 2015, the year of King Abdullah’s death. In the aftermath of leadership changes following the king’s passing, authorities concerned with crowd planning reshuffled their responsibilities and decided to entrust a different team with the Hajj preparations for 2015.

For the first time since 2006, a massive crowd accident resulted from pilgrim flows unexpectedly crossing a main flow from a side street – a scenario that has been prevented successfully in the years before. Until this day, we have not gained substantial insight into the underlying causes of the 2015 accident, but some press articles contained interesting speculations [2].

After what seems to be one of the largest accidents to date, we were asked to rejoin the planning committee by the Ministry of Hajj in 2016. We have since provided operation plans for the 2016, 2017 and 2018 Hajj season. Again, no crowd disaster occurred in these years.

However, the Hajj season 2018, with more than 2.3 million pilgrims a great success, is our final contribution to the project.

Specifically, we are not part of any preparation activities for Hajj 2019 and have no plans to be involved in Hajj planning going forward.

It has been a very interesting, challenging adventure for our team. We wish the current crowd management all the best to keep the event free of accidents. Hajj 2019 starts on Friday, August 9 and ends on Wednesday, August 14.

  

Further Interesting material around Hajj can be found in [3].

---

 [1]

Haase, K., Kasper, M., Koch, M., & Müller, S. (2019). A Pilgrim Scheduling Approach to Increase Safety During the Hajj. Operations Research, 67(2), 376-406.
https://doi.org/10.1287/opre.2018.1798 

Haase, K., Al Abideen, H. Z., Al-Bosta, S., Kasper, M., Koch, M., Müller, S., & Helbing, D. (2016). Improving pilgrim safety during the Hajj: an analytical and operational research approach. Interfaces/Journal of Applied Analytics, 46(1), 74-90.
https://doi.org/10.1287/inte.2015.0833

[2] 

To stone the devil safely - InsideScience.org
https://www.insidescience.org/news/stone-devil-safely

The 10-Minute Mecca Stampede That Made History – Vanity Fair
https://www.vanityfair.com/news/2018/01/the-mecca-stampede-that-made-history-hajj

 Hajj stampede: 5 things you need to know – CNN
https://edition.cnn.com/2015/09/24/middleeast/hajj-deadliest-disasters/index.html

 [3]

 Johansson, A., Helbing, D., Al-Abideen, H.-Z., Al-Bosta, S. (2008). From crowd dynamics to crowd safety: a video-based analysis. Advances in Complex Systems 11(4), 497-527
- Material: http://www.trafficforum.org/crowdturbulence

 Helbing D., Johansson A. (2009) Pedestrian, Crowd and Evacuation Dynamics. In: Meyers R. (eds) Encyclopedia of Complexity and Systems Science. Springer, New York, NY

 

Mathematik weist Pilgern den Weg - Deutschlandfunk (In German)
Link - Deutschlandfunk

Pilgern mit Plan - 19Neunzehn

Article in German - Link (PDF)

Das größte Fußgängerproblem der Welt in den Griff kriegen - Hamburger Abendblatt

Article in German - Link (PDF)

Organisation von Menschenmassen - Hamburgs Beste 

Video in German - Link (mp4)