Power-down mechanisms are well known and are widely used to save energy; such mechanisms are encountered on an everyday basis. Power management and energy-efficient computing are ever more important, given the growth in the number of information technology devices. In fact, it is surprising that only a few algorithmic techniques exist. And many approaches are simplistic or ad-hoc, and do not work well in practice nor do they have good theoretical underpinning. We study power down management in the framework of online competitive analysis. Recent results were obtained in collaboration with researchers at the University of Electro-Communications, Tokyo. For two-state on-off systems a novel decrease and reset algorithms is analyzed. The method automatically attunes itself to the frequency of requests, and gives better performance for real world inputs than standard algorithms. Work is under way to generalize results to systems with multiple states. The work presented in this talk was funded by the by the Panasonic Corporation of Japan and the National Science Foundation (IIA 1427584).

Biographical Sketch: Professor Bein holds a Ph.D. (Dr. rer. nat.) from the University of Osnabrueck, Germany. He has been on the faculty of Duke University, the University of New Mexico, the University of Texas at Dallas, and was on the technical staff of American Airlines, before joining the faculty at UNLV. At UNLV he is a Professor of Computer Science and co-Director of the Center for Information Technology and Algorithms.