Amazon develops algorithm to improve collaboration between robots and humans

Amazon has identified a financially beneficial way for robots and humans to coexist, and it’s saving the online enterprise half a billion dollars per year. Using robots to bring shelves of inventory to associates to pick customer orders cuts down the distance traveled by robots and decreases the company’s storage footprint.

A common fear is that artificial intelligence (AI) and robotics will put Americans out of work, but that may not be the case—new research in the INFORMS Journal on Applied Analytics is using the science of operations research (O.R.) to enable humans and robots to work together more effectively.

The study, “Algorithm for Robotic Picking in Amazon Fulfillment Centers Enables Humans and Robots to Work Together Effectively,” shows the work done to redesign the robotic picking algorithm used in Amazon Robotics (AR) fulfillment centers (FCs).

“Rather than associates going to the product shelves to pick for a customer order shipment or stow new inventory, robots bring shelves of inventory to associates who are at workstations either picking or stowing items,” says Russell Allgor of Amazon.

The picking algorithm decides which specific units of inventory on which pods should be picked to fulfill customer order shipments. The algorithm reduced the distance traveled by pods by 62% without negative operational impact. This has now been implemented in all Amazon Robotics FCs.

“This improvement reduced the number of drives required in AR FCs by 31%, which amounted to half a billion dollars in savings,” continued Allgor.

Allgor, alongside his co-authors Tolga Cezik and Daniel Chen, both of Amazon, says that by implementing AR FCs as the standard for new fulfillment centers, Amazon has also reduced their storage footprint by about 29% when compared with non-AR FCs. This translates to more efficient use of space and a more productive human workforce.