Analysis and Control of Stochastic Networks: From Data Centers to Transportation Systems
February 17, 2017, Webb 1100
Modern large-scale stochastic systems face much demand and processing variability, and a key challenge is the design of efficient control and scheduling policies that are robust to these uncertainties. In this talk, I will present several robust scheduling policies for stochastic models with applications to emerging sectors such as data centers and intelligent transportation systems. In the first part, I present a robust scheduling policy with performance guarantee, for a novel stochastic model of job scheduling in data centers, where jobs are represented as directed acyclic graphs (DAG). I will then visit the long-standing open problem on the stability of of the longest-queue-first scheduling policy for multiclass open queueing networks, and resolve this problem for an important special case. In the second part of the talk, I focus on transportation networks. I develop the first exact analysis of fixed-time control for urban networks, and briefly mention a few opportunities and challenges in exploiting feedback control policies.
Ramtin Pedarsani is an assistant professor in the ECE department at UCSB. He obtained his Ph.D. in Electrical Engineering and Computer Sciences from UC Berkeley in 2015. He received his M.Sc. degree at EPFL in 2011 and his B.Sc. degree at the University of Tehran in 2009. His research interests include stochastic systems, information and coding theory, and transportation systems. He is the recipient of the best paper award in the IEEE International Conference on Communications (ICC) in 2014.