Martin Andersen says: “The growing accumulation of data in today’s world results in complex problem sets and difficult computational challenges for data scientists. The existing optimization tools used to analyze such large data sets often require excessive computational resources (time and memory), and this imposes limitations on scale and scope, forcing data scientists to rely on approximations or oversimplified models. This project will develop a new generation of scalable methods that are fast enough to solve current and emerging problems without sacrificing performance, while still remaining user-friendly for data scientists. These new methods will have a wide interdisciplinary reach across applications within power systems, identification of biological processes, materials analysis and more. The project’s focus on the underlying mathematical algorithms and the hidden mathematical structure in problem sets will generate new knowledge within the field of optimization and result in tools ready for implementation by researchers and companies alike.”