FC 2016

Optimal Feedback Control of Computing Systems under Uncertainty

Abstract: The proliferation of data and the advances in statistical and machine learning methods create tremendous opportunities for optimal feedback control under uncertainty to address computing systems at scale. We consider a general approach moving from data to stochastic models of uncertainty within a computing systems based on statistical methods and then to optimal feedback control of the computing system based on these stochastic models of uncertainty. Two different examples will be used to illustrate instances of our general approach, both involving dynamic computer resource allocation problems. In each case, we devise an optimal feedback control policy that includes easily and efficiently implementable algorithms for governing dynamic resource allocation over time. Numerical experiments demonstrate the significant benefits (e.g., high-performance, energy-efficiency) of our approach over previously published results.

About the Speaker: Mark S. Squillante is a Distinguished Research Staff Member and the Area Head of Stochastic Processes, Optimization and Control within the Mathematical Sciences Department at the IBM Thomas J. Watson Research Center. He also serves as Director of the Center for Optimization under Uncertainty Research across IBM Research. He has been an adjunct faculty member at Columbia University and a Member of the Technical Staff at Bell Telephone Laboratories, and has held visiting positions at various academic institutions. His research interests concern mathematical foundations of the analysis, modeling and optimization of the design and control of complex systems under uncertainty. He is a fellow of ACM and IEEE, and the author of more than 250 technical papers and more than 30 issued or filed patents. His work has been recognized through The Daniel H. Wagner Prize (INFORMS), 8 best paper awards, 11 keynote/plenary presentations, 14 major IBM technical awards, and 26 IBM invention awards. He serves as Chair of IFIP W.G. 7.3 and as Chair of the Computer Performance Foundation, and currently serves on the editorial board of ACM Transactions on Modeling and Performance Evaluation of Computing Systems, Performance Evaluation, and Stochastic Models. He received a Ph.D. degree from the University of Washington.