Statistics show AI successfully cutting picking routes

Recent statistics have shown that artificial intelligence (AI) is helping to reduce the distance being travelled in warehouses.

FM Logistics reports that after deploying the AI-based Google AlphaEvolve at a logistics site in Poland, the yearly distance travelled by warehouse workers was cut by 15,000 kilometres, a reduction of about 10%. This was achieved without investing in improving the warehouse infrastructure or buying more forklifts and picking equipment from a warehouse equipment supplier.

Order picking is a major cost in the warehouse, especially in e-commerce where pickers collect many items from different areas of the warehouse for a single order. FM Logistics has 17,700 picking locations. Even small errors in planning picking routes result in extra distances walked by pickers and longer vehicle journeys. Though the order picking process was efficient, FM Logistics wanted to improve it.

Before using AlphaEvolve, the company built an order-grouping algorithm that improved picking performance. AlphaEvolve was developed by Google DeepMind and used AI to refine FM Logistics algorithms and create new ones that find solutions to complex optimisation problems. FM Logistics was one of the first companies to take part in the AlphaEvolve early-access programme.

Each change that AlphaEvolve suggested was tested in a simulation model. Not every suggestion applied to the live warehouse. The combination of AI-suggested changes and the company’s operational knowledge arrived at solutions that were both practical and effective. The results of this partnership were shorter picking routes, and a decrease in operator travel of 15,000 kilometres a year,

FM Logistics says that this demonstrates that improving warehouse efficiency does not always require investment in new equipment.

You may also interested in: