Wireless Network Coding: Analysis, Control Mechanisms, and Incentive Design

Publication Year:
2014
Usage 218
Abstract Views 168
Downloads 50
Repository URL:
http://hdl.handle.net/1969.1/152682
Author(s):
Hsu, Yu-Pin
Tags:
wireless network coding; network control; game theory; incentive design
thesis / dissertation description
The access to information anywhere and anytime is becoming a necessity in our daily life. Wireless technologies are expected to provide ubiquitous access to information and to support a broad range of emerging applications, such as multimedia streaming and video conferencing. The need to support the explosive growth in wireless traffic requires new tools and techniques that maximize the spectrum efficiency, as well as minimize delays and power consumption. This dissertation aims at novel approaches for the design and analysis of efficient and reliable wireless networks. We plan to propose efficient solutions that leverage user collaboration, peer-to-peer data exchange, and the novel technique of network coding. Network coding improves the performance of wireless networks by exploiting the broadcast nature of the wireless spectrum. The new techniques, however, pose significant challenges in terms of control, scheduling, and mechanism design. The proposed research will address these challenges by developing novel network controllers, packet schedulers, and incentive mechanisms that would encourage the clients to collaborate and contribute resources to the information transfer. Our contributions can be broadly divided into three research thrusts: (1) stochastic network coding; (2) incentive mechanism design; (3) joint coding and scheduling design. In the first thrust we consider a single-relay network and propose an optimal controller for the stochastic setting as well as a universal controller for the on-line setting. We prove that there exist an optimal controller for the stochastic setting which is stationary, deterministic, and threshold type based on the queue length. For the on-line setting we present a randomized algorithm with the competitive ratio of e/(e-1). In the second thrust, we propose incentive mechanisms for both centralized and distributed settings. In the third thrust, we propose joint coding and scheduling algorithms for time-varying wireless networks. The outcomes of our research have both theoretical and practical impact. We design and validate efficient algorithms, as well as provide insights on the fundamental properties of wireless networks. We believe these results are valuable for the industry as they are instrumental for the design and analysis of future wireless and cellular networks that are more efficient and robust.