Resource allocation in wireless networks.

Abstract

In this dissertation, two types of wireless communication networks are considered: cellular (single-hop) networks and multihop wireless networks. In emerging and future generations of wireless communication systems, heterogeneous communication services will be provided over a single cellular network. Satisfying the diverse service requirements of different users through a single radio band, however, is a challenging issue due to the scarcity of radio resources, user mobility, and the time-varying, unreliable nature of wireless channels. For cellular networks, the problem of radio resource sharing in the context of DS-CDMA wireless networks is studied. The primary concerns are the efficiency of radio resource usage and QoS support among users. The approach to resource sharing problems is via formulation of static nonlinear constrained optimization problems, the solutions of which are used to develop dynamic resource allocation mechanisms. Optimal solutions are derived, and simple autonomous and cooperative implementations of the algorithms are provided. Extensive real-world simulation is used to evaluate their performance over time. For multihop wireless networks, the design of a robust routing protocol, or modifications of conventional routing algorithms, to efficiently adapt to time-varying network topologies arising from user mobility has been a critical issue. An equally important feature of multihop wireless networks is the controllability of its point-to-point connections or topology, because these are primarily determined by the transmission powers of the users. In this work, an active approach of controlling the network topology is taken, rather than a passive approach which attempts to resist degradation of wireless channels. A transmission power allocation algorithm based on graph-theoretic results is provided, and it guarantees certain desirable features of network topology characteristics, such as k-connectivity. The performance and complexity of the algorithm is evaluated via analysis and simulation.