Guest Commentary: CompSust09
The following is a guest commentary about CompSust09 by Bob Day, Assistant Professor of Operations and Information Management at the University of Connecticut, and winner of the Dantzig Dissertation Award and the INFORMS Computing Society Prize for his research in combinatorial auctions.
Regarding CompSust09, the 1st International Conference on Computational Sustainability, the most positive aspect of the conference was that we had a lot of people from different disciplines in the same room, hearing and seeing each other’s research presentations and posters. We got to hear what sustainability meant to those focused on species management, and those concerned with maintaining sustainable human communities. We saw an abstract economic model for long-term social planning, trying to find an optimal schedule for the consumption of renewable and nonrenewable resources over the next 11 generations. (Just to name a few.)
My overall goal in attending (I presented nothing) was to gather information about what was out there, and to get an idea of how I might contribute to sustainability causes through future research. This conference did a great job of opening my eyes to what is out there, and definitely helped me narrow the field of where I might want to go with future research in sustainability.
As an OR person myself, I was most comfortable (and impressed) with the presentations focused on algorithms and practical computational techniques. So the highlights for me were the talks by Carlos Guestrin (Carnegie Mellon), Steve Phillips (AT&T Labs), Warren Powell (Princeton), and David Shmoys (Cornell). Not only were these insightful talks from big people in OR and CS, but it was also quite encouraging to think that they were making connections with other sustainability researchers in need of “computational help.” This was a bit of a theme among the conference (other than the curse of dimensionality) that people from an application-focused research area could benefit from a little added computational expertise, perhaps by learning from or partnering with someone from an algorithmic background. It was also enlightening to see the level of sophistication from the ecological discussions (for example) in terms of posing interesting optimization problems that require advanced computational technique.
But despite an overall positive experience at the conference, there were a few things that left me feeling dissatisfied. The main problem was a lack of diversity in the participants in terms of discipline. Compared to the literature that is out there (for example what we can link to from greenor.org), there was a disproportionate representation of biological/ecological areas (roughly two-thirds of the conference), and a relative lack of representation from the energy-sector and industrial applications focused on sustainability. This disparity was reflected in the participation by government agencies: only the U. S. Geological Survey had a noticeable presence, with no participation from the EPA, the DOE, or even the USDA, for example. I expected to hear more about gas emissions, or about mobilizing the economy for a change into a more efficient, and cleaner industrial model, but there were only one or two talks along these lines. As business faculty, I have been hearing a lot about green-supply-chain management, green product design, green manufacturing, etc. But really nothing like that, and aside from just one or two talks, I found the more industrial applications totally lacking.
So even though this conference was enjoyable and informative overall, perhaps a wider and more balanced group of participants could enhance future CompSust conferences, making it a more true representation of everything both computational and focused on sustainability. The problem is probably a reflection of this being the FIRST conference in the CompSust stream, and one that I hope the organizers can fix in years to come. (And could awards be given for best computational sustainability research paper, best research application? These things tend to help stimulate interest and improve quality.)
My only other problem was learning how much there is still left to be done in the field of sustainability, and it is frustrating to know how little we can do today to save coral populations, or clean up the atmosphere, for example. We are still really far from making optimal decisions, as we are behind on both the accuracy and sophistication of our models and in our ability to manage hard optimization problems. Fortunately, I saw a lot of hope from both sides at CompSust09, and it seems that the CompSust series will provide fertile opportunities for collaboration among natural scientists and computer scientist/mathematician types like you and me, for years to come.