MHI Solutions Improving Supply Chain Performance
When did you first notice the shift toward artificial intelligence (AI)? Perhaps during the COVID‑19 pandemic? Or maybe even earlier than that—as far back as 10 years ago or more? The answers to these questions will likely vary according to your distribution center or warehouse.
But regardless of where your distribution center or warehouse is based or which products it stores and processes, three facts are clear. AI has arrived throughout the entire material handling industry. It’s here to stay. And it’s directly improving order orchestration and fulfillment considerably.
Personally, Taoufik “TK” Haddadi, sales manager at MHI member Ryson International Inc., observed a shift toward meaningful AI integration between 2015 and 2017. At that time, cloud‑based warehouse and transportation management systems began to incorporate machine learning modules that could learn from data rather than simply follow programmed rules.
“Before that, most systems were rule‑based and pretty rigid,” he said. “The shift to cloud infrastructure was key because it gave these systems the computing power and data access that they needed to be effective.”
Since then, AI has positively influenced order orchestration in three primary ways. First, it has transformed how orders move throughout the supply chain, as plants and warehouses now have “intelligent routing,” according to Haddadi. As a result, AI can analyze which distribution center or warehouse should fulfill an order based on their carrier performances, delivery timelines, inventory levels and shipping expenditures.
Second, AI systems can now predict inventory needs by analyzing purchasing patterns, seasonal trends and external factors, including local events and weather. In turn, they offer real‑time demand forecasting, enabling your distribution center or warehouse to position inventory before it’s ever needed.
“Finally, when things go wrong, which they always do, AI has become much better at exception handling,” Haddadi added. “Instead of flagging every issue for human review, modern systems can adapt and make decisions autonomously, such as automatically rerouting an order if a warehouse is at capacity or a carrier is experiencing delays.”
In recent years, Joel Thomas, head of intralogistics USA for MHI member Siemens Digital Industries, has witnessed another critical positive regarding AI and order orchestration. Distribution centers’ and warehouses’ data sets are more organized than ever before, as they can be combined, reviewed by operators and used to drive improvements. A wide array of data iteration and configuration options can also be implemented, yielding more information on order orchestration efficiency than ever before.
“Through the use of simulation, we’re able to combine the real and digital worlds,” he said. “Data can be tied in with the real world, brought back into edge computing and contextualized. Through simulation and real‑world dynamics, we can link all data together and learn what we can do to improve efficiency.”
Raj Senguttuvan, senior vice president of product for MHI member Roboteon, has especially noticed a rise in vision system usage over the last few years, leading to an enhancement in order orchestration. Used for depalletizing, palletizing, picking and placement, among other tasks, vision systems are powered by deep learning vision algorithms. Through these algorithms, vision systems can automatically pick items and place them in appropriate locations, irrespective of the items’ features, textures and types.
“Even if there is a slight misalignment, a vision system can adjust itself in real time and place items in the right bins,” he said. “Even grippers—from soft to two‑finger, for example—can be changed automatically and in real time.”
Looking ahead to the future, with regard to AI and order orchestration, Haddadi thinks even routine orders will be fully autonomous, from the moment an item is purchased until it’s delivered. Additionally, dynamic rerouting will become the “standard.”
“Imagine your order automatically shifting to a different fulfillment center or carrier mid‑process because of a hurricane, a supplier delay or traffic congestion,” Haddadi said. “The system will continuously optimize in real time, too.”
Over the last five‑plus years, AI has also enhanced order fulfillment in various ways. In particular, picking routes have been optimized, according to Haddadi. In the past, warehouse and distribution center employees wandered randomly to collect items for orders. But that’s no longer the case, as AI can calculate whichever path is the most efficient for item collection, especially when it comes to multiple orders.
“AI also helps with slotting, which is deciding where to place products,” Haddadi added. “Fast‑moving items get positioned closer to packing stations and AI learns and adjusts, based on actual picking patterns.”
Therefore, warehouses and distribution centers can “effectively bypass labor shortages,” according to Senguttuvan. He has discovered that a “winning approach” is a combination of traditional picking by human employees and AI. Through this combination, warehouses and distribution centers can fully leverage their human talent (without relying solely on it), while gaining the speed and scalability that humans simply can’t provide.
Click here to read the full article.