INFORMS: What inspired you to research this particular topic?

ZWART: There were two driving factors. First of all, I have a tendency to care about problems that are overlooked by others. One of the few results that is almost always true is that the delay of a user in a stochastic system scales inverse proportionally with the ‘system slack,’ i.e., as 1/(1‑rho). However, Shortest-Remaining-Processing time (SRPT) does not, and I want to understand exactly why this happens in size-based scheduling disciplines. The second driving factor was because scheduling is typically studied by two separate communities and our paper was a unique opportunity to blend techniques from the stochastic and deterministic worlds.

INFORMS: Did any of your results surprise you?

ZWART: The result in itself was not surprising to me, more the fact that it was possible to combine techniques from disparate areas.

INFORMS: What is the most important takeaway you hope readers will learn from your paper?

ZWART: There are two in my opinion. (i) It is known that SRPT minimizes delay. The question is, is it possible to find a blind scheduling policy (which is a policy that does not know the sizes of jobs in advance) that matches the growth rate of SRPT in heavy traffic? The answer is no, but we construct a randomized policy that is in some sense optimal. (ii) In our proof we use a bound from the world of competitive analysis in deterministic scheduling. Though this is a bound designed to work in adversarial settings, it is still good enough in the stochastic setting.

INFORMS: Can you give us a real-world example that your technique of combining competitive analysis and applied probability can be applied?

ZWART: This is somewhat speculative, but I suppose the scheduling discipline we construct to show how close to SRPT one gets can be further tweaked to make it applicable in practice, e.g., in data centers.

INFORMS: Tell us about the process of writing this paper.

ZWART: I visited Adam Wierman in 2007 at Caltech and he told me about Nikhil Bansal’s idea to look at this problem. Adam showed me some rough notes from Nikhil, but it was not clear how to proceed. At that time I was at Georgia Tech and Nikhil at IBM.

Due to various events in each of our lives both Nikhil and I ended up in the Netherlands, sharing our professional time between Eindhoven and Amsterdam. We wrote a proposal together on this topic that got funded, and Bart Kamphorst was hired as PhD student.

In the first few months we made everything work assuming the fourth moment of the job size was finite, but we felt this was not good enough. Bart then went to follow a route to make our theorem work under a second moment condition. This was a long and lengthy route that turned into a completely separate paper, but the results were not necessary in the end. More than a year later, we found a relatively simple argument — all we needed was a technical result on finite moments of busy periods, which we suspected should be in the Russian literature somewhere; Vitali Wachtel directed us to the right reference. This led to a proof that I think is very clean. We submitted the paper on Christmas Eve in 2014, seven years after I first saw Nikhil’s notes — which dated back to 2004.

INFORMS: Why was it important for you to publish in Mathematics of Operations Research?

ZWART: Mathematics of Operations Research (MOR) is, in my vision, a top journal, just like Operations Research and Management Science. It ranges the entire spectrum of methods that operations researchers use. As I am on the theory end myself, it is always an honor to publish in MOR. I am also very proud of the fact that seven of my PhD students published a paper in MOR. For this paper in particular, we felt that our second take-away message made MOR the right journal.

INFORMS: How do you define “analytics”?

ZWART: (i) A buzzword that is probably gone in 10 years. (ii) A blessing as it provides our community with an opportunity to complete the missing link between data and decisions. As much as I enjoy combining techniques from stochastic O.R. and deterministic scheduling, I would enjoy seeing more collaboration between operations researchers and statisticians.

INFORMS: How do you keep yourself up-to-date on the latest research in your field?

ZWART: As an area editor in Operations Research, many bright researchers have sent me their best papers for the past nine years and I was able to read what other experts thought of those works. I learned a lot from this and will probably be one of the things I will miss about that job. Other than that, I am looking at arXiv on a daily basis.

INFORMS: What about your career might surprise us?

ZWART: I failed Probability 101 on my first attempt.

INFORMS: Are you currently doing any research? If so, can you tell us a little bit about what you’re working on?

ZWART: I am working on various things. The topic I am most passionate about is the modeling, analysis, and control of electricity networks with novel sources of randomness, such as sun, wind, and plug-ins like electric vehicles.

INFORMS: You coauthored a paper that won the 2016 INFORMS Revenue Management & Pricing Section Best Publication Award. Tell us how this has affected your career.

ZWART: It did not affect my career in the way the Erlang Prize did back in 2008, which probably had a more usual impact that a prize has. Still I believe the prize impacted my career in the following way: we did not work on this topic at all in the Netherlands until Arnoud den Boer became a PhD student of Rob van der Mei and myself on this topic. The prize gave the recognition of the quality of our work. Even though I already knew it, it was giving confirmation that, with a strong and dedicated PhD student, one can switch to a new area and make substantial contributions after a few years. It is helpful to think about this when your paper or proposal about your latest idea is rejected, and continue following the long-term vision.

INFORMS: When you’re not using your OR/MS superpowers to try to make the world a better place, what are some of the ways you like to spend your time?

ZWART: I am happily married with two kids who are two and four years old. If they are taking a nap and I am not too exhausted, I am using the Dutch bike paths to go out for a run. I also love to play checkers on the 10x10 board.

INFORMS: Which INFORMS event do you most look forward to attending?

ZWART: The Annual Meeting. So many people I know come together there. It is nice to see them, even if it is only to say “Hi” for a few seconds.

INFORMS: What would you name the autobiography of your life?

ZWART: Running to Stand Still.

INFORMS: Tell us something that not many people know about you.

ZWART: I hold a master title in 10x10 checkers.

INFORMS: What is the best advice you can give to students in your field?

ZWART: Dare to follow your own ideas.