How Copenhagen Airport started making decisions in real-time

The slime-mold Physarum polycephalum stretches meters across the terrain, re-organises on the fly in the search for food, finds the shortest way through complex mazes – and re-creates human-made networks to a fault.

In 2010, Japanese and British researchers placed oat flakes in such a way as to trace the cities surrounding Tokyo. The researchers watched as the Physarum polycephalum spread out a network “comparable in efficiency, reliability, and cost to the real-world infrastructure of Tokyo’s train network.”

How did it do it? After all, the Physarum has no brain or nervous system.

In comparison to nature’s wonders, human-built networks seem fragile and embarrassing.

During the pandemic, supply networks broke down, and the Russia-Ukraine war caused energy networks to destabilise.

The need to reform the way we have built our networks have however not been driven primarily by single events, but by the steadily growing complexity and pace of our world.

An airport long ago

Back in 1925, when Copenhagen Airport was inaugurated, planes arrived when they wanted to, and the pilot took off when he felt like it. Grass on the runway was kept short by sheep who were shepherded away before a landing.

It was a small world. As the number of yearly passengers grew from 72.000 to 40 million, the world of the airport expanded. Whatever happened at any point in the world was increasingly important to whatever happened at Copenhagen Airport. And whatever happened at any small corner of the airport was increasingly important to whatever happened at every other corner of the airport.

Software engineers built custom extensions and branches on top of the underlying database system to alleviate the problem. One system handled timestamps for check-ins, one system handled timestamps for inbound flights, and so on. It worked – sort of. The only problem was that each flight had a whopping 600 attributes, and there was little time for ground handlers and flight attendants to make timestamps for everything. Even if they had the time, the attributes would still be scattered across various systems. And if the entire system broke down, nobody would know how to fix it due to its complexity.

In 2018, Copenhagen Airport decided to develop a system that would eventually be named AIRHART. In time, it was to completely replace the underlying legacy system – a dusty old IBM system from the 90s called CATS.

AIRHART went into shadow production on June 9th last year. On April 17th, the old legacy system is decommissioned, and AIRHART goes live.

On that day, Copenhagen Airport will be the most digital, the most resilient, and the most integrated airport in the world.

Instead of a complex system of apps that screams for a central coordinator, there is one app that handles all data. Instead of small worlds of data silos, real-time data is distributed and – just as importantly – made actionable across the entire network of stakeholders.

Like the slime mold, Copenhagen Airport has become significantly more responsive to its environment – and more resilient to change.

And the technology can be used in every sector.

PULSE: a general-purpose version of AIRHART

The IT company behind AIRHART at Copenhagen airport is Netcompany. Netcompany developed PULSE as a general-purpose platform. In 2023, businesses are not confined to their own walls but extend far into the world.

Without an operating system that can leverage that complexity, businesses will find it harder to build resilience.

Use-case one: Securing a stable energy grid

When you eat, blood is sent from other parts of your body to the abdominal region. That happens automatically because the body knows that your stomach needs the blood more than other parts of your body.

In a perfect world, Europe’s electricity grid functions just like that.

Unfortunately, right now, the grid mostly resembles a body suffering from both anemia and blood clots. Energy outages and grid congestions are rampant, and the causes are multiple. One reason is those ways to intelligently send electricity to where it is most needed based on real-time sensor data and forecasts about supply and demand are badly underutilised.

That makes the grid vulnerable to surges in demand and supply.

Heading toward 2050, global energy demand is set to double. Further, the use of variable sources of energy is set to increase by 50 % by 2035.

As we move into the future, a flexible, resilient grid will become paramount.

PULSE has already proven its ability to stream large amounts of real-time data across airport stakeholders, and it can be used to stream energy data in real-time as well.  In collaboration with the platforms’ users – DERs, aggregators, companies, and even private heat pumps, car batteries, and electric kettles – we can develop a grid that is future-proof.

Use-case two: Shrinking into excellence

The energy grid might be vulnerable to shocks, but not any more than the supply chain.

In 2015, GoPro was having a bad time. Shares had fallen 89 % from their high in September 2014, and on the shelves sat $57 million worth of inventory that couldn’t be sold. GoPro had expanded and as a result, stacked up their inventory. However, as demand fell unexpectedly, that inventory turned into a liability. GoPro had not been able to accurately predict demand during an expansion of the company, and the result was near bankruptcy.

What if companies could accurately predict demand in real-time, and thereby reduce the need for large inventory buffers and their associated risks?

Byrne Hobart of Capital Gains calls that maneuver “shrinking into excellence”. It involves greater visibility up and down the supply chain and “keeping a closer eye on demand, too; the more you know about what your customers will be doing in the near future, the more you know what you need to be doing right now.”

The only problem is that accurate information about the supply- and demand chains are not easy to come by. Less than 40% of logistics firms have introduced digital platforms and data analytics over the past three years. Less than a third use Internet of Things technology to keep track of their inventory and goods in real-time. If we want a resilient supply chain, that needs to change, and PULSE is the platform on which that change happens.

Use-case three: Hospitals

Predicting demand is just as important in healthcare as in inventory management. When an unexpected rush hour in the emergency department occurs, an avalanche of questions arises: at what exact time does the ambulance arrive? Who is the patient, and what is their status? Are there rooms available for the patient? Equipment? Doctors? Are there other hospitals that might be suited to better handle this patient?

Such questions need to be answered collectively, in real-time. Today, hospitals do their best when a rush arrives and discuss crowding events afterward in monthly meetings.

A platform such as PULSE could help hospitals more accurately predict patient flows ahead of time and plan for various scenarios and risk profiles. It can manage the automatic booking of rooms and equipment based on real-time information, and it can facilitate lifesaving, real-time data sharing between various medical entities.

At this point, the obvious question is:

If indeed all sectors of our economy could benefit from a real-time data platform equipped with AI superpowers, and the technology already exists, then why isn’t it already implemented everywhere?

Back to the future

In 1879, Thomas Edison created the first working light bulb. 20 years later, only 3 % of American households had electricity. Another 20 years later, that number was 50 %.

Why the long wait? According to computer scientist Ajay Agrawal, in his new book Power and Prediction (2022), electricity experienced a boom not when it was seen as a mere reduction in fuel costs, but when it ”provided a way of decoupling energy use from its source. That freed users from the constraint of distance, leading to a cascade of factory and workflow design improvements.”

Analogously, great change today will not come from using AI and real-time data as slightly more efficient ways of doing the same things as we have always done, such as improving fraud detection or weather prediction.

Rather, great change will come from reimaging how businesses work considering the rising complexity of our world.

Such radical systems-changes are not quick fixes, but they bring about change.

PULSE is not today’s light bulb, but today’s broad electrification.

Benefits only arrive with broad implementation. If implemented maximally, the whole supply chain decides instead of one supplier making decisions. Instead of one nurse deciding, all of healthcare decides, in real-time, collectively.

It might sound like a mirage. At Copenhagen Airport, however, it isn’t a mirage. The system is right there, working.

As William Gibson said: ”the future is already here. It’s just unevenly distributed”.