| The evolution of mobile wireless communication, triggered by an ever-increasing need for higher capacity, has fueled the interest in finding more efficient ways of managing the resources available to the system, the most important of which are radio spectrum and power.;In the context of Frequency- and Time-Division Multiple Access (FDMA/TDMA) systems, Dynamic Channel Assignment (DCA) allows for more efficient use of the available spectrum and eliminates the burden of costly frequency planning. Nonetheless, the impact of user motion on the performance of DCA systems is a topic that has not received enough attention. In this dissertation, we compare the performance of a number of representative distributed fixed-power DCA algorithms in different mobility scenarios. A novel algorithm well suited to mobility environments is proposed and studied in depth. In addition, the degradation associated with blind-slot constraints, hardware limitations and imperfect signal and interference estimation is also quantified.;Power, on the other hand, is also a capacity-constraining factor in systems inherently limited by interference. This, in turn, has led to the development of a variety of power control (PC) schemes. PC can help reduce interference thus allowing for smaller reuse distances and a potentially large increase in capacity. Again, not enough effort has been devoted to studying the impact of mobility on the performance of PC schemes. We quantify such impact as well as the degradation associated with imperfect signal and interference estimation in systems implementing distributed PC.;Despite the intense activity in both the DCA and the PC arenas, little has been done aimed at trying to effectively integrate them. We investigate the integration of DCA and PC in mobility environments and we show that the capacity gains of DCA and PC are not additive because of some significant redundancy between the two. Nonetheless, we create a family of integrated algorithms which achieve capacities significantly higher than those of a system with only DCA or PC. Furthermore, these algorithms are fully distributed and adaptive, in the sense that their capacity can be traded for complexity. With respect to a system without any DCA or PC, several-fold capacity increases are obtained. |
