Anabel Diaz Calderon

🍺 Can GenAI manage a supply chain? A new project from Carol Long (Harvard), David Simchi-Levi (MIT), Andre Calmon (Georgia Tech), and Flavio C. (Harvard) explores exactly that. They have built a version of the Beer Distribution Game where each role in the supply chain (retailer, wholesaler, distributor, and factory) is played by a GenAI agent. The simulation is now publicly available for anyone to try: Beer Distribution Game Simulator (https://lnkd.in/e6WNTNDj) . 👏 Congratulations and thank you to the authors for doing this important work and sharing it so openly. The results are fascinating. In decentralized settings where each GenAI agent makes its own decision, the bullwhip effect is alive and well. Adding a centralized orchestrator improves performance in some cases, particularly when demand volatility is shared, although not all forms of information sharing helped. Pipeline and downstream visibility should have eased bottlenecks, but in practice the models showed little benefit. Interestingly, advanced models like GPT-5-mini were able to approximate near-optimal strategies even with limited information. For me, two things stand out. First, it is remarkable that LLMs struggle with some of the same challenges as humans in the Beer Game. Information sharing reduces variability, but it does not eliminate the bullwhip effect. Second, it is incredible that a teaching tool created at MIT in the 1960s is still generating new insights today, even when applied to cutting edge technology. This also reminds me of the work that Michael Watson and the Opex Analytics team did several years ago with machine learning based simulations, which showed that other AI techniques (machine learning, for example) might be better suited to this type of decision making. OPEX was later acquired by Llamasoft, which was then acquired by Coupa. I love seeing how the field of system dynamics and the classic Beer Game continues to evolve, from live classroom play to new digital versions such as J. Chris White, DBA’s Beer Game on Steroids. Yet I still believe the best way to begin this learning journey is the way I started as a graduate student at MIT Center for Transportation & Logistics: the original format, live and in person, where peer pressure and performance anxiety make the lessons unforgettable. Today, I run these workshops for my own clients several times per year, and it always leads to real change in perception and strategy. In the end, it seems like there may still be a need for Supply Chain Management professionals. At least for now... 👉 Try the new simulation yourself and let me know what you think: Beer Distribution Game Simulator (https://lnkd.in/e6WNTNDj) ~Mr. Supply Chain® #AlwaysBeLearning #SupplyChain #BeerGame

379

52 commentaren

Backstage in the green room at DPW Amsterdam about to head up and explore: “Can AI ever be circular?” First we’ll unpack AI, in context of supply chains. We’ll explore a self-driving supply chains and the building blocks of math, physi-digi, probabilistic operations and more. AI is in the cloud, and the cloud is a physical place. So… Then, we will look at circularity. Of the 100 billion tons of materials that enter the global economy each year, it all comes through us in supply chain and procurement. I’ll share innovators that help showcase the physical elements: build, power, cool, and maintain. How are you considering the reasonable use of AI? #supplychain #AI #procurement #innovation

124

15 commentaren

Truck makers in the EU face a 15% CO2 reduction target for the first time this year. Our latest analysis looks at the most recent performance of truck makers (using the CO2 of conventional trucks up to 2023, and the sales of ZEVs up to 2024) to see how they have been faring towards this target. One thing stands out: 5 out of the 7 main truck makers have already met their 15% target. First to say, this approach is as conservative as we could be. We don't know what the CO2 performance of trucks will be in 2025, but it will almost definitely be lower than in 2023. Neither do we know yet the ZEV sales in 2025, which will almost definitely be higher than 2024. So who's leading the charge based on this latest data? Scania Group and Volvo Trucks have already reduced their CO2 by 15% in 2023, two full years ahead of schedule. They could both continue to sell the very same trucks they did in 2023 and easily hit the target. Renault Trucks and DAF Trucks NV are also well on track to meet the targets. Both have amassed a truck load of early credits that they can use in 2025 to meet the 15% target, even if they don't improve beyond their 2023 performance. These credits expire in 2026. MAN Truck & Bus SE has not earned as many credits, but a vehicle transfer mechanism - which allows connected manufacturers to effectively pool their emissions - allows them to rely on Scania's early credits, of which there is a lot. Only for two manufacturers - Daimler Truck AG and Iveco Group did we find that some more effort was needed. Both can avail of a vehicle transfer scheme, allowing them to use another manufacturer's ZEVs (limited to 5% of the vehicles they sell in a given year) for their own compliance purposes. If both maximize this flexibility, Daimler will be just 1% from the target, while Iveco will be about 8% from it. From there, Daimler just needs to increase their ZEV sales marginally to 3% to meet the target (it was 2% in 2024). Iveco on the other hand will have to make some serious changes to narrow the gap. The CO2 standards are working, and trucks today emit less than ever before. The targets for 2030 will undoubtedly be a challenge, but ZEV sales have been rising year on year across the EU. Full study available from here: https://lnkd.in/dNx73Fde

103

9 commentaren

This is rail, reimagined for the 21st century. 🌎🚆 Parallel’s autonomous battery-electric vehicles can travel in flexible platoons, reconfiguring quickly to serve different cargo needs and routes. By breaking free from the long-train model, we make short-haul freight by rail possible — offering shippers faster service, lower costs, and cleaner operations. It’s a new operating model for rail: adaptable, efficient, and ready for markets that have always relied on trucks. 📹 See how we’re making it happen! #CleanEnergy #FreightTech #RailPlatooning #ParallelSystems

205

11 commentaren

We’re proud to see Epicenter Alumni Einride continue pushing boundaries in autonomous and electric freight. Their latest milestone at the Port of Antwerp-Bruges – one of the world’s most complex logistics hubs – proves how Europe can lead the way in deploying autonomous freight technology at scale. With operations already running in both Europe and the U.S., Einride’s data-driven electric freight platform is shaping the future of logistics: smoother flows, lower emissions, and smarter use of resources. Way to go, team Einride Linnéa Kornehed Falck Robert Falck 🔗 Read more here: https://lnkd.in/dpsCiJwr

97

“AI isn’t just optimizing logistics — it’s orchestrating the entire supply chain.” #Congratulations to the Nauta team on raising $7M in seed funding to build an #AI-based logistics orchestration platform, with a focus on #importers and expanding to broader #supplychain operations. This round was led by Construct Capital (Rachel Holt & Dayna Grayson) with participation from Amplo, Dynamo Ventures, 8VC, 9Yards Capital, and TMV Soraya Darabii,Marina Hadjipaterass). This incredible group of investors brings deep experience in supply chain, logistics, and tech. We’re thrilled thatRachell will be joining us on stage atManifest: The Future of Supply Chain & Logisticss, sharing her perspective on where AI is taking our industry next. #fundingnews #AI #logistics #supplychain #startups #manifestvegas 🔗 https://lnkd.in/eScPvSf4

25