AI logistics – predicting orders

  • AI logistics
  • AI logistics
  • AI logistics

About This Project

Order today, receive tomorrow. This is the standard we accept today. For logistics service provider CB, this norm creates new challenges. They were searching for an answer to their daily question: “To what customers are we expected to deliver tomorrow and what quantity?”. Mediaan accepted the challenge of answering this transportation question, using artificial intelligence (AI). The result is a model that predicts orders per delivery address which increased planning accuracy by 80%.

ORDER before 11 PM, receive it the next day

The biggest challenge is to get thousands of orders, the latest arriving at 11:00 pm, to be delivered the next day by the most suitable carrier. This requires not only smart and efficient scheduling, but also knowing how to deal with uncertainties and doubts.

Although we make the basic planning for the next delivery day at four o'clock every night, we will know the exact information of location and how much needs to be delivered by 01:00 am the next night. However, we cannot wait for that, because we will no longer get the connections with external carriers, which means we will let down our customers. In order to make the right choices, the best possible prognosis is very important for our delivery process. At 07:00 pm and 11:00 pm we will rebalance, based on the latest information. The data of those different measuring moments ultimately determine what we transport as CB itself and which external carriers we use. - Arjan De Jong, Process and Implementation Manager at CB

AI logistics

The forecasts of Arjan de Jong and his team rely on available current data, experiences and historical data. The company is setting up the automated forecast itself.

Because we are dealing with a huge amount of information and because the logistical process is complex and vulnerable due to time pressure, we have reached out to Mediaan. The aim was to develop a model that enables us to determine whether we can use AI to get a clear answer to the logistics question: to what customers do we expect tomorrow to deliver how much volume? Using machine learning algorithms, Mediaan has developed a AI model that allows us to get answers based on our historical order data. - Arjan De Jong, CB

IMPROVEMENT OF 80%

The mediaan forecast shows an improvement of more than 80% in predicting the number of addresses CB needs to deliver to. The AI model itself looks for patterns in the data that influence the forecast and, based on this, predicts the drop size per delivery address. This enables CB to further realize an optimization in cost, quality and efficiency. The power of the Microsoft cloud enables them to deliver a specific customized forecast in no more than half an hour for approximately 2,500 addresses.

PROOF OF CONCEPT

This model has been developed as a ‘proof of concept’ in three weeks by four Medianers: Axel Goblet, David Roschewitz, Marcel Pereboom and Marc Pont. A mix of young and talented data scientist and an experienced IT Architect. The developed model is designed so that CB can easily retrieve new features from the data and add it to the prediction. For example, orders per region and the relationship between bookshops and products.

In the next few weeks, we will discuss how to proceed with this promising technology. As an innovative logistics service provider, we are constantly looking for improvements, optimizations and innovations in order to better help our customers. - Arjan De Jong, CB

CB

Whoever thinks of CB, remembers a company that has been successful due to the logistics of paper and digital books for decades. Their main customers are consumers, webshops, publishers, bookstores and crossdock transportation for third parties clients.

In our warehouse in Culemborg, there are approximately 40 million books and related articles in stock. Every day, we pick almost 250,000 orders. The transport of B2C orders, ie directly to the consumer, we do in cooperation with partners. We deliver 2,500 B2B orders throughout the Benelux. On an annual basis, this is half a million. The distribution of all those orders is largely been done by CB itself. Almost 4o percent, especially smaller volume shipments, we outsource to transport partners like Audax logistics, postNL or DHL. In addition, only a few people know that we also support our customers with billing, administration and debt management. And when a book is out of stock, we can print books with our print-on-demand service. Due to our logistics expertise and our delicate transport network, we have also become active in Healthcare & Fashion. - Arjan de Jong, CB
Customer

Centraal Boekhuis (NL)

More info

Stefan Daelemans on +31455718355.
Or send an e-mail to info@mediaan.com

Category
Data & Intelligence, Logistics, Machine Learning