He had it all figured out, Bas Kerkhof thought four years ago, when he worked as a Site Manager at Rhenus’ facility in Kraaiven. The warehouse, packed chock-full with products from an online pet supply store, was strictly organized. The only possibility to be more efficient would be to work even harder. At least, that was what he supposed.
However, this was only until Pipple convinced him that improvements could be obtained by employing algorithms. Although, convincing… Kerkhof: “I could not imagine that there would be additional room for improvement. ‘Just try’, was what I thought.” At the time, he did not know that Rhenus would be Pipple’s first client.
Beyond spreadsheets
The Excel spreadsheets, constructed by the engineers at Rhenus, were indeed put together well. However, using mathematics, artifical intelligence and programming languages, it should be possible to go a few steps further. This would lead to 5% savings, Pipple promised.
In the end, the savings turned out to be at least twice as large. The crux? Smarter picking routes and different order combinations. The model, built by Pipple, was thereafter also applied to other European sites. Kerkhof’s wager paid off twice over.
Kerkhof’s successor, Rutmer Mourits, decided to hire Pipple for a follow-up project:’Batching 2.0.’ Mourits: “Because of the shorter routes, the order pickers collected the products in the warehouse more efficiently and therefore worked faster. Nevertheless, we were still facing congestion at the drop-off station, i.e. the conveyor belt.” Once again, the engineers at Rhenus and the econometricians at Pipple pulled together to tackle the problem.
Self-learning planning software
This joined-up effort led to the development of a dynamic planning model with a self-learning nature. As a consequence, a planner will eventually become redundant. Mourits: “It works like a navigation system: the model constantly calculates which routes navigate the picker as quickly as possible to the final destination without running into queues.” A self-learning, autonomous AI-planner as employed here is revolutionary in our branch, Kerkhof argues.”The large majority of order fulfillment companies still relies on ordinary sorting algorithms.”
Truly, we are partners on a joint mission and already have been for almost four years now.
Rhenus is already working with the new software in the real-life operational environment while the algorithm is still constantly being tweaked. “This is the best way to quickly make progress”, Mourits explains. “Entrepreneurship is all about taking risks. “It is a learning process, which he would not want to go through with any party other than Pipple. “We are the client and Pipple is the purveyor, however it does not feel like that. Truly, we are partners on a joint mission and already for almost four years now. This can also be clearly noticed from the chemistry between our engineer Henny Zhang and the consultants at Pipple. They speak the same language and trust on each other’s expertise.” Kerkhof: “Pipple is pragmatic and venturous, which totally agrees with our culture.”
