This talk considers how to restore electric power systems after a natural or man-made disaster. Such a restoration is extremely challenging from a computational standpoint as it combines a complex logistics problem with activation problems over a complex physical system. We present a four-step approach which allows us to separate the logistic and restoration aspects, while ensuring an overall high quality restoration. The approach, which heavily exploits hybrid optimization, is validated on benchmarks using realistic power system data. The experimental results show significant improvements over the practice in the field and the critical role of hybrid optimization to find high-quality solutions in reasonable time. (Joint work with Carleton Coffrin and Pascal van Hentenryck)
Russell Bent received his PhD in Computer Science from Brown University in 2005 and joined LANL as a technical staff member that year in the infrastructure analysis group (DSA-4). He is currently the leader of the optimization research and development team of DSA-4. He is a Co-PI on a laboratory directed research and development (LDRD) grant on the topic of the smart grid, Co-PI on a project for uncertainty quantification of networks, and Co-PI for the DOE-OE project Grid Science. He is also responsible for textit{RestoreSims}, a project for resource management, planning, and distribution for disasters both for the National Infrastructure Simulation and Analysis Center (NISAC). Russell’s publications include deterministic optimization, optimization under uncertainty, infrastructure modeling and simulation, disaster planning, constraint programming, vehicle routing and scheduling, supply chain, algorithms, and simulation. In the past ten years Russell has published 1 book and over 40 articles in artificial intelligence and operations research. A full list of his publications can be found at http://public.lanl.gov/rbent/