How an algorithm predicts travel problems before they emerge

train-delayThere is seldom anything as frustrating and stressful in our working lives as the commute to and from work each day.  Studies have shown that many of us, especially the younger ones among us, would be quite happy to take a pay cut in return for a shorter commute.

Whilst there have been a tremendous number of new apps and services to try and better inform us as to potential travel problems and delays, many of these tend to be real-time alerts, so whilst they’re useful, they don’t really allow us to circumvent the problems before they arise.

For instance, at the recent Rugby World Cup, many train operators revealed that they were caught unaware by the sheer volume of passengers they were required to transport, and long delays ensued.

Predicting problems before they occur

A Swedish project, called Stockholmstag, aims to change that by providing a smart warning.  They’ve developed an algorithm that analyzes behavior on the Swedish train network and is capable of predicting delays on the network around two hours before they will occur.

The aim is to both provide train operators with enough time to put on extra services, and also to alert passengers of problems looming to allow alternative plans to be made.

The algorithm was created by the mathematician Wilhelm Landerholm and uses historical data to see how small delays can ripple throughout the train network.

This can often result in small initial delays causing substantial problems throughout the rest of the network as connecting services become delayed and so on.

The hope is that the algorithm will allow operators to make the kind of adjustments required to modify service levels before problems emerge.

The tool is expected to go live later this year, and will hopefully prove to be sufficiently successful that it is rolled out in other parts of the world too.

Related

Facebooktwitterredditpinterestlinkedinmail

One thought on “How an algorithm predicts travel problems before they emerge

Leave a Reply

Your email address will not be published. Required fields are marked *

Captcha loading...