|
Fans of the Matrix series of movies might like this one. In those movies, the bad guys the human heroes fought against were called agents. Well, actually the agents were artificial constructs that represented the real bad guy - the computer system that kept humans imprisoned in a persuasive but totally false version of reality.
If a recent article from Line56 (Supply Chain Agents, Demir Barlas, November 5) is to be believed, something very much like agent technology may soon hit the supply chain. The idea is to equip supply chains with artificial intelligence systems that will be able not only to proactively monitor supply chain performance but also independently take corrective action to, for example, detect a forthcoming material shortage that could affect production and delivery, and not just order the necessary material but even go to an alternative supplier if necessary.
It's not all that leading-edge, actually. The article quotes Steven Kickert, CTO of supply chain execution company HighJump, who points out that NASA has used this kind of AI for the Mars probe, programming it to schedule its activities so that it maximized the amount of time it spent facing the sun. It's just that this kind of thing hasn't yet been applied to the supply chain. And I bet your average supply chain has a heck of a lot more interconnected, interdependent variables than prioritizing work so that you spend as much time as possible facing a particular way. (I mean, everybody does that, right?)
As applied to the supply chain, this agent stuff would represent a departure in that existing intelligent supply chain technologies - the precursors of agents, if you will - can't do much thinking at the moment. As Barlas explains, because they're limited-input, they're also limited-output. In plain English, existing intelligent supply chain tools are pretty much a matter of programming a tool to detect a given condition or set of conditions and issue a predetermined response in each case, very often with the goal of getting a person involved. A true agent would not only do the monitoring and detection, but once a 'trigger' condition has been identified, it would then use complex rules and algorithms to determine and execute the best response independently.
Kickert imagines a future in which large numbers of such agents would attend to large, geographically distributed supply chains encompassing different types of facilities. He admits neither his own company nor any other is ready to put anything like this on the market yet, but he says the technology does exist and it's just a matter of time. He says the supply chain VP of one large retail customer is already actively interested.
No word on whether this person is a movie fan.
Supply Chain 2020
In September, supply-chain experts from 12 European companies attended the first meeting of the Supply Chain 2020 European Advisory Council meeting in Madrid. The meeting was sponsored by the MIT Center for Transportation & Logistics and the Zaragoza Logistics Center.
Among other items, the Council plans to model the outcome of various selected scanarios to understand how supply chains might respond. They haven't decided what all the scenarios will be yet, but one might be the rise of oil prices to US$100 a barrel, the spread of RFID and tougher environmental legislation.
Bears watching. See http://web.mit.edu for more.
Andrew Brooks
Editor
|
