What's Your StORy?

How do you define “analytics”?
Using math, data, and computers to help make decisions.

What prompted you to enter the field?
In the early 1980s, engineering school curriculums were less specialized than they are now. We had time to sample courses from several engineering disciplines. I enjoyed my first course in operations research and also in computer science. The concepts were broadly applicable and sharpened my thinking. Soon after, I became intrigued by supply chain problems. Often they were simple to ask: how many of which parts to make when? And yet challenging to answer. This reminded me of an earlier love—chess—that has few rules but the many permutations require heuristic approaches.

Tell us about your experiences at IBM.
In 1984, I joined IBM’s semiconductor manufacturing business in East Fishkill, NY, to help design and develop a new software system for production control. Mostly, it was trying to understand what the users were doing manually and then designing software to automate it or—when their method was ill-defined—developing a method that seemed logical. It wasn’t until a decade later that I first used linear programming for planning IBM’s semiconductor division supply chain. By that time, the problems had grown more complex, and key members of the user community were comfortable using computers to make calculations that they could not perform personally on their hand-held calculators. In the years that followed, computing power and user comfort with it continued to increase. We saved IBM hundreds of millions of dollars by re-engineering the semiconductor supply chain planning and decision-making process.

Why did you leave IBM after 27 years to become an assistant professor?
In 2008, our IBM department downsized deeply. I got a job in another IBM department that I didn’t enjoy as much as when I was working on supply chain problems. Academia provided the freedom to work on any problem I wanted and the opportunity to teach and mentor young people.

You have more than 40 U.S. patents—tell us your favorite!
Material requirements planning (MRP) systems create plans that can be infeasible indicating actions that should have been done in the past. Though infeasibilities are undesirable in most contexts, in the context of a site MRP run, the site’s delivery commitments have already been made so the infeasibilities indicate where expediting is required. On the other hand, for a new IBM technology (with double-speed sorted product), linear programming (LP) was more appropriate for wisely allocating assets. I spent weeks trying to figure out how to make MRP smarter or to get an LP to indicate the expediting needs. Neither approach worked. Pressure for me to deliver was growing along with my frustrations. Finally, it occurred to me how to blend both LP and MRP technologies. This led to my first and favorite patent (U.S. Patent 5,943,484).

Tell us about your experiences with the Edelman competition – from being an Edelman laureate, to editing the Interfaces Edelman special issue.
Being a competitor, coach, judge, chair, and special issue editor for 17 Edelman Award competitions has provided me with a front-row view of the world’s best work in O.R. and analytics. It’s fascinating and inspiring.

What have you learned recently?  [can use either example below]
In New Guinea, Médecins Sans Frontières prototyped drone deliveries of medical supplies in which the drone stopped to recharge its battery. This approach extends the range of drone deliveries and involves a new set of problems in which O.R. can help: designing and using a network of drone battery recharging stations.

Amazon’s 1,800 square feet minimarket opened today (Monday, January 22, 2018). Using cameras extensively, the store has no shopping carts and no checkout lines.

What advice would you give to your younger self?
Volunteer for activities that are important and interesting—even though they’re not your job. You’ll learn more, enjoy it, make contributions, and you’ll often be recognized for it. Also,  working half-days on Saturdays will boost your productivity 20-30% (due to uninterrupted time) and you’ll still have plenty of weekend time for other matters (including relaxing/exercise).

Why have you decided to give back to INFORMS by volunteering as a mentor in the new Mentor Match program?
It’s satisfying to help others. I’ve found mentoring productive and especially when I’ve been on the protégé side of the relationship. The one-to-one interactions are pleasant because you can focus on one person at a time. When mentoring, I always learn something.

What are your other interests?
Church, kayaking, college hockey, spending time with my wife.