Artificial intelligence in city logistics

“Sustainable city logistics is a hot topic. But above all, logistics is a cost factor. If you add an environmentally compatible solution for the last few kilometres of the journey to the delivery costs, no-one wants to pay that,” says Jean-Paul Rival, one of the four founders of French start-up Urbismart. “Therefore, if you want to deal with city logistics properly, you have to rethink the entire supply chain.” As a result, Urbismart has developed a model for a shared truck platform that breaks down the boundaries between shippers, logistics service providers, hauliers and even local authorities – in short, all of the players in the supply chain. In doing so, they are seeking to create a supply chain that satisfies all interests.

Describing the Urbismart business model, Rival says, “We buy transport and sell it to the shippers with a margin that generates income. We manage everything in such a way that the logistics providers can operate highly efficiently and thus profitably, while the shippers incur minimal costs. To this end, we combine the delivery quantities of the participating shippers, look for the best routes to the customers being supplied, and ensure that the trucks are loaded optimally.” In hubs on the outskirts of the city, either cross docking takes place, or the deliveries are prepared in detail by merging the B2B and B2C shipments. This gives the transport companies access to some of the deliveries which are currently entrusted to express couriers or the postal service. This, in turn, increases the efficiency of the driver making deliveries to the street in question – especially if he uses his empty truck to pick up return deliveries, swap bodies, clothes hangers or cardboard boxes for recycling.

Specifically, this means that tractor-trailers cover the journey from large ware­houses to cross-docking hubs on the outskirts of the city, where just one vehicle is loaded with all of the consignments for one street in the city centre – goods shops and parcels for private internet customers. Accordingly, there is just one delivery truck for each street. In return, the authorities should guarantee a parking space for the vehicle from which the deliveries are to be made. There is neither noise nor pollution as long as the truck remains parked there. “We need dynamic optimisation in real time,” explains Jean-Paul Rival. “On the one hand, you have hundreds of shippers, and on the other, hundreds of cities. In between, there are thousands of transport companies and tens of thousands of recipients. This is big data – a huge set of data that is constantly changing and requires quick responses.” New situations occur all the time as orders are changed, trucks break down or drivers fall ill, or when the current traffic situation or the weather disrupt the route calculations. “This amount of information and data can no longer be handled by conventional software, let alone people in front of a screen. We need a more powerful machine here – and this is precisely where artificial intelligence comes in.”

“Shippers are faced with increasingly complicated issues because of multi-channel deliveries,” says Jean-Paul Rival. They lack experience in collating and delivering all of the orders from shops and also from online buyers. At the same time, they want to keep down their costs. The first victims of this financial pressure are the transport companies: there are more than 30,000 such businesses in France, and many of them are close to bankruptcy, according to Rival. At the same time, local authorities want to limit traffic and thus reduce noise and pollution – while shop deliveries should still be guaranteed. “We therefore wanted to combine the demand and the services of all players in order to exploit the relevant synergies. If the interaction between the various shippers, service providers and delivery channels is well organised, then everyone benefits,” says Rival. He is certain that shippers could thus achieve significant savings, as transport companies would improve vehicle capacity utilisation. Moreover, there would be fewer delivery vehicles on city centre roads.

The concept was put into practice and further improved during a test in Bordeaux, which lasted several months. “The companies that took part were enthusiastic without exception,” reports Rival. Deliveries to the city centre were made in cooperation with Libner, a truck bodybuilder, which has developed a small electric delivery vehicle that is carried on board a truck. In a car park close to the city centre, a pallet loaded with goods is pushed onto the electric vehicle, which is then set down by the truck and begins making deliveries along the road in question. When the pallet is empty, the next one is taken from the truck, and the deliveries continue. Rival explains, “The test in Bordeaux has proven that we can manage supply chains such as this with a variety of shippers and hauliers.”

So far, Urbismart remains a more or less virtual undertaking. To position itself as a player on the market, the start-up needs to have its own offices, IT and staff, and acquire as many customers, that is shippers and hauliers, as possible. “To get going in practice, we initially need at least 1.5 million euros,” admits Rival. “At the moment, we’re still looking for investors.”