Live Long and ProspOR

For my final blog post, I thought it might be nice to do some sort of review of the 15+ weeks we spent learning about public sector operations research. As such, I have compiled a summary of five things I learned over this past semester that were particularly important to me:

  1. Equity is important in public sector operations research. This is something that perhaps shouldn’t have been surprising to me, but was. I was intrigued by the idea that it’s possible to build characterizations of equity into a mathematical model and that such a model can be constructed in a myriad of ways. Some ways of modeling equity are certainly better than others, as we saw in Sam’s presentation of the paper “Modeling effectiveness-equity trade-offs in public service delivery systems,” but the inclusion of an equity measure is critical in many public sector problems such as assigning children to schools. [1]
  1. Bilevel optimization is cool (and hard). While talking about applications of bilevel optimization, we learned about some interesting ways it can be used to solve tricky problems with multiple decision makers. In Eli’s presentation about stochastic network interdiction, we saw how network interdiction problems work like a Stackelberg game where the leader tries to take links out of a network so the follower has to find other routes. [2] I also presented a paper about nuclear-weapons project interdiction that used bilevel optimization in a project management setting where the leader tried to make the follower’s tasks take as long as possible. [3]
  1. There are a lot of different ways to do facility location. First of all, when deciding where to locate a facility, what’s our goal? Do we want ambulances as close to as many people as possible? Do we want to place distribution centers so that all customers are “covered” (i.e., within some predefined distance radius)? Once we know what we want, how do we model it? Should we use a p-median model? Should we include queuing – and therefore randomness – in the model? Once we finally have a model, we have to solve it, which is often hard since facility placement requires binary decisions about whether to build a particular facility or not. I presented a paper that provided a novel approach to problems with random budgets, such as facility location problems, that involved building a two-stage stochastic program. [4]
  1. Presentations work a lot better when you have a moderator. I got to do something this semester I’ve never done before: moderate someone’s presentation. Not only was it fun, it encouraged me to read the paper (or at least the slides) before watching the presentation and really think about what made sense and what didn’t, because it was my job to come up with several questions for the person I was moderating. I think it’s a great idea to have assigned student moderators and if I teach a class someday I’ll definitely have my students moderate for each other.
  1. It’s possible to do good with good OR. Early in the semester, we discussed the idea of “doing good with good OR,” which basically boiled down to the question of whether our strategic problem solving was actually helping anyone. While there have certainly been times OR has done more harm than good, I’d say our ultimate answer was a resounding “Yes!” We saw many examples of the ways OR had been applied to solve real-world public sector problems; I thought it was encouraging to know our work has the potential to make life better for someone at the end of the day, whether it’s sending children to the best school or placing an ambulance where it has the greatest potential to save someone’s life (see [5] for one great example).

Anyway, all this is to say it was a great semester. Thanks for reading, and here’s to doing even better with even better OR in the future!

(PS: I’m very sorry for the horrible pun that is this post title.)

[1] Mandell, Marvin B. “Modelling effectiveness-equity trade-offs in public service delivery systems.” Management Science 37.4 (1991): 467-482.
[2] Cormican, Kelly J., David P. Morton, and R. Kevin Wood. “Stochastic network interdiction.” Operations Research 46.2 (1998): 184-197.
[3] Brown, Gerald G., et al. “Interdicting a nuclear-weapons project.” Operations Research 57.4 (2009): 866-877.
[4] Koç, Ali, and David P. Morton. “Prioritization via stochastic optimization.”Management Science 61.3 (2014): 586-603.
[5] Hutton, David W., Margaret L. Brandeau, and Samuel K. So. “Doing good with good OR: supporting cost-effective hepatitis B interventions.” Interfaces41.3 (2011): 289-300.

Advertisements

3 thoughts on “Live Long and ProspOR

  1. Such a warm end-of-semester post.. I would love to add a vote for #4. Thank you for being the good moderator and excellent question generator.

    Liked by 1 person

  2. Great idea for a final blog post. I’m kind of jealous I didn’t think of it.

    I would also add:

    6. Blogs are fun

    I’ve never written a blog before, but I thought these assignments really forced us to think through the content and ensure that we were writing in a way that was easily understandable and concise.

    7. 30-minute presentations aren’t that bad

    The longest presentation I’d given prior to this class lasted 15 minutes, so I wasn’t sure how a 30-minute presentation would go. With that said, I thought the 30-minute presentations turned out great. While they do require a lot more preparation, I felt this led to a much deeper understanding of the material. In comparison, 15-minute presentations now seem like a cakewalk.

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s