Date on Master's Thesis/Doctoral Dissertation


Document Type

Doctoral Dissertation

Degree Name

Ph. D.


Electrical and Computer Engineering

Degree Program

Electrical Engineering, PhD

Committee Chair

Li, Hongxiang

Committee Co-Chair (if applicable)

Faul, Andre

Committee Member

Faul, Andre

Committee Member

Inanc, Tamer

Committee Member

Kumar, Anup

Author's Keywords

multicast; network coding; queueing stability; lyapunov


In this dissertation queue stability in wireless multicast networks with packet erasure channels is studied. Our focus is on optimizing packet scheduling so as to maximize throughput. Specifically, new queuing strategies consisting of several sub-queues are introduced, where all newly arrived packets are first stored in the main sub-queue on a first-come-first-served basis. Using the receiver feedback, the transmitter combines packets from different sub-queues for transmission. Our objective is to maximize the input rate under the queue stability constraints. Two packet scheduling and encoding algorithms have been developed. First, the optimization problem is formulated as a linear programming (LP) problem, according to which a network coding based optimal packet scheduling scheme is obtained. Second, the Lyapunov optimization model is adopted and decision variables are defined to derive a network coding based packet scheduling algorithm, which has significantly less complexity and smaller queue backlog compared with the LP solution. Further, an extension of the proposed algorithm is derived to meet the requirements of time-critical data transmission, where each packet expires after a predefined deadline and then dropped from the system. To minimize the average transmission power, we further derive a scheduling policy that simultaneously minimizes both power and queue size, where the transmitter may choose to be idle to save energy consumption. Moreover, a redundancy in the schedules is inadvertently revealed by the algorithm. By detecting and removing the redundancy we further reduce the system complexity. Finally, the simulation results verify the effectiveness of our proposed algorithms over existing works.