Helping the Homeless

Last week, my Multiobjective Decision Analysis class presented their class project to the Daily Planet, an agency in Richmond, Virginia, that provides a number of services for the homeless and underpriveledged.

Working on the project was eye-opening for me and the students. As an optimization person, it was refreshing to work on a problem where “optimal” had little meaning–the challenges in the project were different than challenges we are are used to facing. It was satisfying to see the students develop so much respect for the case managers.

The big challenge for me was to manage expectations and motivate the students to move forward. Half of the semester was used to determine what the problem was to begin with. That was frustrating. I learned the hard way that students will drag their feet and resist making decisions in such an open-ended project. Eventually, the class headed in the right direction and worked closely with the Daily Planet to make a set of good recommendations.

Over the course of the semester, the students engaged me in a discussion of what “real” OR is. Most of them thought that creating and solving mathematical programs is real OR, whereas dealing with more ethereal issues of defining a problem and selling a solution is not real OR. Not surprising, since most OR curriculums focus on teaching how to model and solve well-defined problems rather than recognize OR problems. I think the students came around a little bit. Eventually.

The students’ final presentation was well-received by the Daily Planet. We didn’t get to make very many quantitative recommendations due to the lack of data that has been collected. The agency was thrilled to hear that many of the ideas they came up with over the years but never enacted really are good ideas.

The coolest thing about this project? My students truly learned what the “science of better” is all about. They had all heard the OR slogan, but never knew what it meant.