Argonne National Laboratory

MACSER: It's about the math

February 23, 2018

The DOE-funded MACSER project held its kickoff meeting at Argonne National Laboratory Feb. 9, 2018. MACSER is a new multi-institutional center funded by the U.S. Department of Energy Advanced Scientific Computing Research program to address a grand challenge: how to quantify the occurrence and features of rare, high-impact events in complex energy and environment systems.

The meeting opened with an introduction by Mihai Anitescu, senior computational mathematician in Argonne’s Mathematics and Computer Science Division and director of the MACSER project. He outlined the project’s technical objectives, which include advancing foundational mathematics and statistics; formulating flexible mathematical models of complex systems; and designing and operating energy systems that can withstand and recover from high-impact events.

But equally significant, Anitescu stressed, are the organizational issues: How do MACSER researchers ensure a balance of high-risk and low-risk activities, and how are such activities integrated? Anitescu emphasized two ideas: automation and innovation. On the one hand, project reporting, evaluation, and especially accomplishment tracing should be done as automatically as possible. On the other hand, interactions should be about new ideas, new demos, and new directions that cannot be done in isolation.

“We also need to make ourselves more visible, clarifying the importance of math In tackling complex problems in operations research and energy systems,” said Anitescu.

To this end, the meeting included presentations on various mathematical challenges – for example, uncertainty quantification, machine learning, resilience, and dynamic optimization –with the aim of stimulating collaborations among principal investigators. One area of special interest to Argonne researchers is how space-time statistics can help overcome challenges from the power grid – for instance, determining the best location of wind farms to minimize vulnerability to high-impact events.

The meeting also provided an opportunity for MACSER members to discuss use cases, develop documents identifying clear mathematical formulations, and assign tasks and timelines for the next two to three years.