CSL Professor Petros Voulgaris is part of a collaborative grant from the National Science Foundation, titled “Distributed Quasi-Newton Methods for Nonsmooth Optimization.” The grant aims to expand some of the most efficient, recently developed methods for convex optimization to the case of highly structured and distributed scenarios.
Optimization, which finds the most effective mathematical function for a system, is a workhorse algorithm behind many of the advances in smart devices or cloud-based applications.
“As data gets larger and more distributed, new ideas are needed to maintain the speed and accuracy of optimization,” said Voulgaris, professor of aerospace engineering and ECE affiliate.
The focus of the grant is on large, distributed, and streaming data sets, so that the resulting techniques and embedded systems implementations can support optimization for cyber-physical systems, such as robotics, infrastructure, health care, manufacturing systems, and the Internet of Things.
“This grant will help deliver methods that will outperform current state-of-the-art in terms of speed of computations, scalability with big data sizes, the ability for systems to deal with uncertainty, and work over many networks in real-time,” said Voulgaris.
The team expects these new methods to benefit a variety of applications: signal processing, control, machine learning, and robotics.