Network scheduling
A system that can handle 170 million parcels. Swiss Post, facing increasing e-commerce fulfilment needs and volume spikes, is digitally optimising its operations.
It’s not easy. Night after night, Swiss Post transports approximately 500,000 letters and parcels nationwide. The corporation’s parcel centres, with their own sidings and rail-mounted gantry cranes, are at the heart of its operations, together with its logistics network of parcel centres and 66 parcel bases.
Hundreds and hundreds of swap bodies have to be moved at the same time, docks have to be correctly assigned to lorries and the shunting units and gantry cranes have to be synchronised.
This complex logistics system currently offers a maximum capacity of 135 million parcels a year. As ever, however, the devil is in the detail. “In the 1990s we operated t a steady pace that spread the load throughout the day,” recalls Stefan Luginbühl, the head of transport and sorting at Swiss Post.
That’s long gone. Today, fluctuating volumes from hour to hour, day to day and month to month have to be managed, as do evening peaks brought on by the rise of e-commerce solutions, including next-day delivery. Then there are also orders tailored by location and time, which make standardised logistics processes difficult. So how to optimise activities? “We use Inform’s ‘Yard Management System’ (YMS), with its intelligent algorithms, in all three parcel centres, to manage swap body and trailer movements in the yard,” Luginbühl explained.
Artificial intelligence improves organisation
The system has been in use since 1999 and has been continuously developed. It spans the entire process, as fully automated entry and exit gates use transponder scanners to detect incoming vehicles and their loads.
The networked dispatching system recognises incoming swap bodies and automatically prints out a work order at the entry gate, indicating the best unloading location for the swap body as well as the subsequent loading location for the next transport order.
The decision concerning which swap body to move and when it is made based on a number of parameters, including the queue of incoming trucks, dock occupancy and loading spaces in the yard, train departure times and each swap body’s priority. This process minimises empty runs and selects the shortest possible shunting routes.
Manual override in place if needed
Humans, however, remain in control of the situation. “Dispatchers can also change the recommendation if their experience shows that the solution isn’t optimal,” Luginbühl explains. Rarely, however, does the dispatcher have to manually intervene in this process. Drivers confirm the successful execution of a transport order by pressing a button, which then transmits the information to the transport management system by radio.
When asked what yard logistics would look like without optimising algorithms, Luginbühl points to some old magnetic boards. “We’d probably still be moving wagon labels from A to B, and compiling lists.” Instead, it’s all fully automated.