Abstract
S.C. Borst, M.G. Markakis, I. Saniee (2014) “Nonconcave Utility Maximization in Locally Coupled Systems, with Applications to Wireless and Wireline Networks,” IEEE/ACM Transactions on Networking, 22(2).
Today’s communication networks, such as 4G cellular networks or the Internet, are expected to support large traffic volumes and an increasingly complex mixture of best-effort (e.g., web browsing, emails, file transfers) and delay-sensitive services (e.g., voice/video over IP, web TV), with resources that are fundamentally limited. Thus, there is need to make the most efficient use of these resources, and achieve optimal performance across several network layers, different technologies, and various network nodes. On the other hand, the massive size of today’s networks makes the implementation of any sort of centralized optimization procedure extremely difficult. Hence, a key challenge, and the central thrust of our work, is to devise algorithms that are “simple enough” to be applicable, and yet offer guaranteed performance to a diverse population of users/services.
In this paper we propose an iterative algorithm that operates in a decentralized and asynchronous fashion (hence, it is suitable for complex, large-scale networks), requires limited computational effort per iteration, and has provable convergence and optimality properties. Case studies and numerical experiments are also presented to illustrate the applicability of the proposed method to realistic scenarios.
Download this working paper in PDF format
Mihalis G. Markakis is an assistant professor in the Department of Economics and Business, at Universitat Pompeu Fabra and PhD from the Laboratory for Information and Decision Systems, at MIT. His research interests are in modeling, analysis, and optimization of stochastic systems and their applications to Operations Research and Management Science. His teaching is concentrated on the Master of Science in Management (specialization in Business Analytics) at the UPF Barcelona School of Management.
Authors
Mihalis G. Markakis
Assistant professor in the Department of Economics and Business
Pompeu Fabra University
