Smarter algorithms prevent factory robots from colliding
Chinese e-commerce warehouses shipped more than 83 billion packages in 2020, nearly 23 times more than a decade ago. This boom in online shopping is creating huge logistical challenges for retailers like JD.com, which sells everything from books to home appliances, providing same-day delivery to 470 million customers in China who place orders before. 11 a.m., whether they’re buying something big or small, common or hard to get.
The solution to this huge logistical challenge could be robotic automation, according to a team of researchers from JD.com, Shanghai University of Finance and Economics, Chicago Booth and University of Southern California. . In their work on JD.com, the researchers helped the company improve the efficiency of its fulfillment warehouses by adopting a facility management model in which workers stay at stations while robots bring them products to pack. The results could one day help companies deliver products efficiently by drones and robots, they write.
JD.com is comparable to the best-known Alibaba in terms of revenue ($ 114 billion in 2020 for JD.com compared to $ 109 billion for Alibaba). But JD.com manages its own inventory and fulfillment while Alibaba operates more like an online marketplace, similar to eBay or Etsy. To move goods in cartons and on trucks and planes, JD.com has largely automated its warehouses and manages its own shipments. (Imagine if Amazon owned FedEx.)
In its use of robots, JD.com overturns the conventional human-driven distribution center model. In a standard warehouse, people are rushing around on foot or by forklift, grabbing and packing products as orders arrive. This is called a picker-to-parts model. JD.com uses a parts picking system, where swarms of football-sized robots race up the floors of distribution centers to bring racks of materials from shelves to humans waiting to grab and drop. ” package the products. The system increases productivity, preserves worker safety and reduces costs, the researchers write.
However, to make the parts picker work, each of the hundreds of warehouse robots must solve the complex problem of knowing which shelves to bring to which packers once every five seconds, without colliding with each other. JD.com first tried to achieve this by using standard artificial intelligence software to control the robots, but “commercial software has proven to be too slow,” says Linwei Xin of Chicago Booth, one of the researchers. involved. Because of this, the company found that many of its automated warehouses were not sending enough packages to justify the construction costs.
About five years ago, researchers began to develop algorithms to improve the functioning of warehouse robots. They divided the problem into discrete layers, or challenges. For example, an integrated management layer links fulfillment to other systems such as order processing. Each layer is defined by multiple algorithms, and all are continuously monitored and updated to help automated centers cope with changing demands and conditions.
The algorithms created by the researchers allow the company’s warehouses to function properly even when orders accelerate up to 10 times normal, as they did during the worst of the coronavirus pandemic, according to the research. While many conventional factories have slowed down operations or have been shut down completely to prevent the spread of the virus during the height of the recent pandemic, JD.com’s warehouses have been able to operate using the algorithm-based coin picking system.
JD.com saw a decrease in its execution expense ratio (which measures the costs of executing sales) to 6.5% in 2020 from 7.2% in 2016. In addition, in 2020, 90% of Orders sold by JD.com, rather than by third-party sellers, were delivered the same day or the day after they were placed, the researchers report.
Xin sees potential outside the warehouse, perhaps to control drone or robot delivery of product to urban areas, or to deliver items such as vital medical supplies to hard-to-reach areas.
However, “100% automation is not ideal,” the researchers write. Even the most automated warehouses still need people to do surveillance, maintain robots, and package non-standard items, they note. Having said that, they have yet to find a point where investing in robots will yield diminishing benefits.