Friday, March 3, 2017

Decision-Making Under Deep Uncertainty With Applications to Energy

I arrived back from Washington DC about 1AM this morning after a late flight. I was in DC to speak and take part in a fascinating international expert panel on Financial Supernetworks.

Today, I had a lot to catch up on but was very much looking forward to hearing my College of Engineering colleague, Professor Erin Baker, speak this afternoon in our UMass Amherst INFORMS Speaker Series. Her presentation title is below.
This Speaker Series is organized by the award-winning UMass Amherst INFORMS Student Chapter.

The students had brought food - pizza plus pastries and fresh fruit and rinks, which were served prior to Professor Baker's lecture. She was introduced by the very energetic and hard-working Chapter President, doctoral student in Management Science at the Isenberg School, Pritha Dutta.
Professor Baker's outstanding talk presented a theoretical framework to handle "deep uncertainty," as when experts are in conflict or models are in conflict. She focused in her talk on the former. The very elegant theoretical framework consisting of alternatives, uncertainty, and preferences, along with expert beliefs used in a Bayesian way provide us what she and her collaborators are calling "belief dominance" in contrast to stochastic dominance and Pareto dominance. The framework eliminates bad solutions and from the set of the best ones identifies the robust ones.

Her research focuses on decision analysis and energy, including renewables (and she is head of a big NSF IGERT project on wind energy). She is an active member of INFORMS.

The application of the theoretical framework was to energy R&D portfolios since there is disagreement among experts on the best solutions for climate change.
Part of the motivation for the research that she presented today was that dynamic decision-making under uncertainty and learning has been criticized as lacking external consistency. It does make a lot of sense that in robust decision-making (and multiple stakeholders) one should start with a small number of alternatives and share with the stakeholders.  Stakeholders appreciate transparency. Her application to a portfolio of energy investment choices, such as nuclear, biomass, etc., at 3 different levels, was vividly illustrated in different colors highlighting belief dominance.

There was an excellent Q&A that followed after Professor Baker's talk and I snapped the group photo below).
Many thanks to Professor Baker for speaking today especially since you had even taught a class shortly before!