On the optimal formulation of resource management in wireless networks

Abstract

Resource management is of great importance in wireless networks, as it can improve the scarce wireless spectrum usage, enhance the network capacity, and help to provide QoS support for various services. In this thesis, we discuss resource management in both CDMA cellular networks and wireless ad-hoc networks. For the CDMA cellular network, we firstly present a comprehensive survey on resource management in such a system and then introduce two major aspects of work we have done so far: firstly, admission control for the downlink multi-services MC-CDMA cellular networks. In this work we quantitatively demonstrate the significant impact from specific statistical factors, which are derived from stochastic behavior of mobile subscribers and channel conditions, and design the admission strategy accordingly for connection-level traffic and the background data transmission; secondly, adaptive cell sectoring for non-uniform traffic to minimize the total transmission power in a CDMA cellular network. We make use of a dynamic programming approach to formulate and solve the optimal sectoring problem. In addition, to reduce the computational complexity incurred from the optimal calculation and to prevent sector boundaries going across high-density regions, we propose a cluster-based sectoring algorithm with bounded complexity under high density traffic distribution. For the wireless ad-hoc network, we first look into the recent work on through-put capacity maximization and energy conservation in such an infrastructure-free system, where cross-layer joint optimization could be applied to achieve more efficient resource utilization. We then introduce mobility assisted routing in a mobile ad-hoc network (MANET). In this work, we explore node mobility in the search for packet delivery routes under a deterministic mobility model, while optimizing a variety of QoS criteria. We consider networks operating in both interference-free and interference-limited regimes. In the interference-free case, we present efficient polynomial time algorithms to find optimal routes. In the interference-limited case, we give a mathematical formulation of the most energy efficient routing for multiple packets, and present an efficient heuristic algorithm. Our recent work mainly focuses on joint optimization in multi-channel multi-radio wireless multi-hop networks. We formulate the optimization under a deterministic model, and we seek to minimize overall system activation time to satisfy a given end-to-end traffic demand, subject to the number of channels, the multi-access interference restriction in the system and the radio interface constraint at each node. The exact solution to such an optimization problem is prohibitively complex due to the combinatorial complexity, particularly with the deployment of multi-radio and multi-channel. We then develop an efficient column generation based approach to solve this problem. Lastly, we discuss our future work in wireless mesh networks. Seeking efficient resource management strategies, which comply with the specific demands of this emerging area, would be the main focus of our study.