Georgia Tech Supply Chain and Logistics Institute

Georgia Tech Supply Chain and Logistics Institute Contact information, map and directions, contact form, opening hours, services, ratings, photos, videos and announcements from Georgia Tech Supply Chain and Logistics Institute, College & University, 765 Ferst Drive NW, Atlanta, GA.

08/28/2026

Supply chain utopia isn't a new idea.

Ask a room of leaders to describe the ideal system and you'll get some version of what's in this image. You'd have gotten roughly the same answer in 1995.

The goal hasn't moved. What's moved is everything standing between us and it.

Compute got cheap. Data got fast. The distance between sensing, deciding and executing is collapsing and physical AI is starting to close the last stretch, the part where a decision has to turn into a pallet that actually moves.

Which means the constraint was never the vision. It was the friction.

And that surfaces a question the field has mostly been able to avoid:

If the system can sense, decide, and execute on its own, what is management actually for?

Chris Gaffney and Zaid Duwayri take that on September 3. Part 1 of three.

What would you add to the list? Or take off it?

Somewhere in your operation, a planner just deleted a set of routes your company paid six figures to generate.They didn'...
08/25/2026

Somewhere in your operation, a planner just deleted a set of routes your company paid six figures to generate.

They didn't do it out of stubbornness.

They did it because the routes were wrong in ways the optimizer couldn't see:

→ Service times in the model don't match what actually happens at the dock
→ The sequence ignores which driver knows which customer
→ The math is feasible on paper and unworkable at 6:15 a.m.

So the planner rebuilds it by hand. Every day. And the software gets blamed for underdelivering on savings it was never set up to find.

Here's the uncomfortable part: no commercial route optimization system can actually produce an optimal set of routes. Not one. The problem is computationally out of reach, and every vendor is selling you a very good approximation while using the word "optimized."

That gap (between what the system optimizes and what the operation needs) is where most last-mile cost hides.

LastMile-IQ was built to answer questions like this one in the open. No gated whitepaper, no discovery call. Georgia Tech researchers and last-mile practitioners, answering the questions that actually come up on the floor.

The full answer to this one is in the archive, along with 50+ others on routing, day assignment, delivery windows, and evaluating last-mile technology.

Link in the comments. 👇

If your planners override the optimizer regularly — how often? Curious whether it's most days or most routes.

Your fall recruiting calendar has one Georgia Tech-shaped hole in it. 30+ organizations already have a table at Supply C...
08/25/2026

Your fall recruiting calendar has one Georgia Tech-shaped hole in it.

30+ organizations already have a table at Supply Chain Day. There's still room for yours.

Wednesday, September 9 | 10am–2pm | GT Exhibition Hall

One room. Four hours. Hundreds of Georgia Tech students who chose supply chain on purpose. They're looking for full-time roles, internships, co-ops, and capstone projects.

You're looking for people who can already think in constraints, flows, and tradeoffs.

New this year: we've expanded with the Center for Health & Humanitarian Systems, so organizations working in healthcare operations, analytics, med device, payers, health tech, and global health are at the table too.

Amazon Robotics, CVS Health, GE Aerospace, Hapag-Lloyd, Manhattan Associates, Michelin, RELEX, Wellstar, Smurfit Westrock, and two dozen more are confirmed.

Registration closes Monday, September 1.

If your fall hiring plan has a Georgia Tech line item on it, this is the four hours that fills it. Link in the comments.

08/21/2026

Ten days left on the countdown.

Through August 31, Georgia residents can take Georgia Tech supply chain courses at 50% off, underwritten by the GA-AIM grant from the U.S. Department of Commerce EDA.

Not a sale. A workforce investment in the state that runs the third-busiest container port in the country.

Warehousing. Logistics. Network design. Planning. The courses our teams actually get asked for.

Use code SCL-GAAIM at registration.

Outside Georgia? Code SCL-Cert takes 17% off certificate programs. Sending three or more from one organization? Email [email protected]

Registration link in the comments. Not sure which course fits your team? Drop your role below and I'll point you.

The best hour of supply chain thinking you'll get this month shouldn't have to compete with a 12:00 meetingStarting Sept...
08/20/2026

The best hour of supply chain thinking you'll get this month shouldn't have to compete with a 12:00 meeting

Starting September 3, Lunch & Learn goes on-demand!
No registration form. No noon block to defend. No missing it because a load was late or a meeting ran long.

Watch it Thursday. Watch it at 6 a.m. Watch fifteen minutes in the parking lot and the rest on Sunday. Back up and replay the part that actually applies to you. Send it to the two people on your team who need it more than you do.

And when one of these lands on the exact problem sitting on your desk eight months from now, it'll still be there waiting.

Same conversations. Same people worth hearing from. Just built to fit the way you actually work.

First session under the new format: Chris Gaffney and Zaid Duwayri on what's left for leaders to manage once AI both makes and executes supply chain decisions. Part 1 of three.

September 3.

What would you want them to take on in Part 2 and 3? Comments are open.

Machine learning in supply chain has a reputation problem.A lot of teams can talk about it. Fewer can use it to improve ...
08/18/2026

Machine learning in supply chain has a reputation problem.

A lot of teams can talk about it. Fewer can use it to improve a forecast, set inventory policies, or spot a problem before downtime hits.

Machine Learning Applications for Supply Chain Planning is built for practitioners who want ML to show up in planning decisions.

What you will work on:

Forecasting future demand and using it to evaluate inventory policies
Customer segmentation with clustering techniques like k-means
Regression (trees), advanced time series forecasting, and classification methods (decision trees, random forests, neural nets, logistic regression, Bayes classifiers)
Building an anomaly detection model using sensor data from a fictional Cardboard Company case to support proactive production maintenance planning
Applying the work in Python and Power BI, with support resources and a pre course webinar if you need a ramp up

Next session:
Sep 14 to Sep 17, 2026 (Online)

If your planning conversations keep ending with “it depends”, this is the kind of course that helps you turn data into a defensible decision.

52% of companies expect a busier 2026 peak season and many are pulling freight forward now to stay ahead of tariff incr...
08/07/2026

52% of companies expect a busier 2026 peak season and many are pulling freight forward now to stay ahead of tariff increases.

Here's the part that doesn't make the headline: frontloading doesn't remove risk. It relocates it.

Dodge the tariff exposure, and you inherit a different set of costs: inventory carrying, warehouse capacity, tied-up working capital, and a bullwhip that lands on everyone downstream.

The teams that win this peak won't be the ones who moved first. They'll be the ones who could actually model the tradeoff before they moved.

That's the muscle we spend our time building.

08/05/2026

Before you present a recommendation, try to break it.
Not because you doubt your work.
Because someone in that room will.

The most dangerous supply chain decisions aren't the ones with bad analysis behind them. They're the ones where nobody asked: "What would have to be true for this to be wrong?"

That's the Socratic method applied to supply chain and it's one of the most underused tools in the analyst's kit.
Before you walk into the room, ask yourself:
→ If volume runs 10% higher than projected, does this hold?
→ If freight costs spike, does the math still work?
→ If throughput underperforms, what breaks first?

One of those questions usually surfaces the assumption.The single variable that, if it shifts, flips the whole recommendation.

That's not a weakness in your model. That's the most important thing you can tell a decision-maker.

"Here's what we recommend. Here's the one thing we'd watch closely. Here's what we do if it moves."

That's not analysis. That's executive-ready thinking.

08/04/2026

On-time delivery is a game show if you only look at the average.

Door A: 92% on-time, but it swings wildly.
Door B: 88% on-time, but it is consistent.
Door C: 95% on-time, but the misses are brutal.

The lesson: reliability beats a pretty average.
Three simple upgrades to your supplier scorecard:
Track OTIF, not just on-time
Add variability, show the spread, not only the mean
Measure impact, late by one day is not the same as late by two weeks

If you can only add one thing, add a percentile view. P90 lead time or on-time performance tells you what planning actually has to survive.

When you choose suppliers, do you optimize for lower cost, higher average on-time, or lower variability?

Sign up for the GT SCL newsletter for practical supply chain insights: https://bit.ly/47zMtbS

Address

765 Ferst Drive NW
Atlanta, GA
30332

Telephone

+14048942343

Website

Alerts

Be the first to know and let us send you an email when Georgia Tech Supply Chain and Logistics Institute posts news and promotions. Your email address will not be used for any other purpose, and you can unsubscribe at any time.

Shortcuts

Share

Category