Running an airport - or for that matter, any major infrastructure - in emergency mode when things go wrong often involves making compromises.
How maths can prevent a repeat of the Heathrow airport chaos
In 2008, a chain of events that began with Queen Elizabeth II of the United Kingdom cutting a ribbon led to almost a week of chaos at Heathrow Airport. Passengers were stranded as mountains of luggage piled up, and hundreds of flights were cancelled.
The inauguration of British Airways' Terminal 5 at Heathrow Airport had been thoroughly planned, but a series of flaws and glitches in the infrastructure and logistics led up to an unpredictable state of confusion.
This cascade effect is becoming ever more prevalent throughout today's high-speed economy. It was involved in triggering the global banking crisis in 2008, and in the massive blackout that hit hundreds of millions of Indian homes this summer.
The internet, and most particularly social networks, have now brought it to everyone's doorstep. When videos go viral on YouTube, or rumours spread on Facebook, the very same cascade effect is involved.
What do these complex systems have in common?
The same dense web of connections that makes them work, as a side effect, also makes them vulnerable. Delays and disruptions, but also rumours and viruses, can spread through them with consequences that are often all but impossible to predict.
Much of our work has focused on aviation, where delayed flights cost airlines, and ultimately businesses and passengers, billions of dollars every year.
But making sure that flights are on time and delays don't propagate is not only good for business. Running an airport - or for that matter, any major infrastructure - in emergency mode when things go wrong often involves making compromises. In such situations, concerns for safety and the environment tend to fade in the face of more immediate challenges.
At Ecole Polytechnique Federale de Lausanne (EPFL), we worked with a regional European airline to improve their flight schedule, designing it to be robust to the small disruptions that strike on a regular basis. By translating the problem into a mathematical model, we could use optimisation algorithms to determine the schedule that best satisfied our demands.
Over decades, airlines have tried to increase their efficiency by adding more flights on tighter schedules, minimising the time aircraft spend at foreign airports. Our findings, however, suggested trading in short-term efficiency for increased robustness.
With our models, we were able to identify where flight schedules should be padded to make them more reliable so that, in the long run, savings from less frequent disruptions would more than offset increased costs.
But planes will continue to break down from time to time, and flight crew will continue to fall ill at short notice. In such an event, our algorithms propose recovery strategies that ensure operations bounce back as quickly as possible and that disruptions do not spread.
In collaboration with EPFL Middle East, we are currently streamlining operations at Saqr Bulk Port in Ras Al Khaimah. To do so, we are optimising the allocation of docking sites for cargo ships and the loading, unloading, transfer and storage of bulk goods, such as cement, coal, grain, or petroleum products.
Container ports, where all goods are handled in identical containers, have been extensively optimised. Bulk ports, in use since the earliest days of shipping, have, however, received much less attention.
What works for container ports does not necessarily apply to bulk ports, mainly because the infrastructures used to handle goods are not the same.
A freighter carrying coal, for instance, has to dock near conveyor belts dedicated to transporting it to appropriate stockyards.
Health and safety impose additional constraints. Toxic goods and grains must be kept a safe distance apart.
The same is true for hazardous raw materials.
By addressing the challenges at Saqr Port, we are developing a blueprint for the efficient design of bulk ports around the world, making them robust to glitches and increasing their resilience so that they can recover quickly if something does go wrong.
More importantly, we are developing new tools and algorithms that will contribute to making airports, power grids, and other infrastructure we depend on more reliable.
Prof Michel Bierlaire is director of the Transport and Mobility Laboratory at EPFL in Lausanne, Switzerland.