Wireless Relay Networks with Stochastic Arrivals

Open Access
Ciftcioglu, Ertugrul Necdet
Graduate Program:
Electrical Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
November 02, 2011
Committee Members:
  • Aylin Yener, Dissertation Advisor
  • Aylin Yener, Committee Chair
  • William Kenneth Jenkins, Committee Member
  • Thomas F Laporta, Committee Member
  • Guohong Cao, Committee Member
  • Randall A Berry, Committee Member
  • Randall A Berry, Special Member
  • Relay networks
  • Cross-layer design
  • Stochastic Network Optimization
  • Two-Way Relaying
  • Quality-of-Information
The ever-growing demand for wireless services has called for efficient multiuser transmission schemes, most notably being relay assisted transmission, where intermediate nodes serve as relays that may cooperate with source(s) or other relay(s) forming wireless ad hoc networks. Significant benefits in terms of transmission rate, reliability and coverage has been demonstrated via cooperative communication techniques. To that end, we focus on the impact of stochastic arrivals in the general setting where pairs of nodes exchange information via multiple intermediate relays. Throughout this thesis, we consider and address issues as scheduling, rate allocation, stability regions and quality of information for wireless networks. The premise of this thesis is that resource allocation and scheduling among users has to be performed in a cross-layer manner for multiuser networks subject to randomness and Quality-of-Information(QoI)-based performance assessment. Consequently, we both address transmission policies, and characterize theoretical bounds for multiuser networks with various physical layer capabilities, input traffic and utility metrics. We start by considering a pair of nodes with stochastic traffic flows who wish to communicate in a bi-directional communication scenario using intermediate relays in two-hop fashion. We first develop the throughput optimal control policy by tailoring the backpressure algorithm to the problem at hand. Two main alternatives for network operation are either to have queues at the relays, i.e., hop-by-hop scheduling, or no queuing at the relays, i.e., immediate forwarding. For the case of static channels, we show the stability regions of these two approaches are identical. On the other hand, for time-varying channels, we show that hop-by-hop scheduling outperforms immediate forwarding. Next, we turn our emphasis in considering some other very critical issues in wireless networks such as minimizing consumption of resources and distributed operation. Each use of the relay incurs a unit cost, for example, representing the required transmission energy. We consider both centralized and decentralized algorithms. The results illustrate new cost-delay trade-offs depending on different levels of cooperation and queue information availability in the network. We then consider the case where users within the network are capable of compression of received data within the network. We demon- strate how compression decisions on distortion-sensitive traffic are effected and carried out along with rate allocation for networks with attributes such as stochastic tra±c and capability of performing network coding. We proceed with Quality-of-Information (QoI)-aware networking, where a new set of attributes are used to characterize value of information. We propose rate allocation and attribute optimization for multiuser chan- nels to maximize network utility based on QoI metrics. We next propose rate allocation and link scheduling for QoI-based utility maximization for a multihop relay network with time-varying channels. Lastly, we study the exploration-exploitation tradeoff for multi- ple sensor devices which aim to maximize QoI utility-minus-operational costs, when the attribute of freshness of received information is taken into account.