Monday, March 3, 2014 - 8:45am to 9:45am
KEC 1007

Speaker Information

Mingyi Hong
Research Assistant Professor
Department of Electrical and Computer Engineering
University of Minnesota


To cope with the growing demand for wireless data and to extend service coverage, the next generation wireless networks will increasingly rely on the use of low power micro/pico base stations and relays. These access nodes will be densely deployed to form small cells in order to provide hotspot coverage and traffic off-loading. However, due to frequency sharing and limited infrastructural resource, interference and network congestion become major performance limiting factors.

In the first part of the talk, we discuss several recent advances in joint interference management and network provisioning for future cellular networks. We show that proper provisioning of such network involves careful design of various aspects such as physical layer signal processing as well as higher layer base station clustering, network traffic flow control etc. We provide a general cross-layer formulation of the problem which can include all these aspects of the design into consideration. The resulting problems are usually nonconvex, nonsmooth, and of very large sizes, hence are very difficult to deal with. We then discuss a family of algorithms based on successive convex approximation (SCA) that is able to effectively solve such large-scale network provisioning problems.

In the second part of this talk, we extend the above SCA approach and present a general framework, referred to as the block successive upper-bound minimization method of multipliers (BSUM-M), which is able to deal with a wide range of large-scale engineering problems well beyond wireless communication/signal processing. The algorithm is simple, flexible, massively parallelizable, and can deal with problems with millions or billions of variables in a highly efficient manner. We discuss different properties of BSUM-M, and show how specializations of such framework can be used to address key issues arising in emerging applications such as machine learning and smart energy systems.

Speaker Bio

Mingyi Hong is a research Assistant Professor with the Department of Electrical and Computer Engineering, University of Minnesota. He received his B.E. degree in Communications Engineering from Zhejiang University, China, his M.S. degree in Electrical Engineering from Stony Brook University (SBU) and Ph.D. degree in Systems Engineering from University of Virginia (UVa) in 2005, 2007 and 2011, respectively. His current research is focused on the design and analysis of next generation wireless networks. He is also interested in the theory of modern large-scale optimization and its applications in signal processing, big data analytics and management of smart energy systems.