Optimal resource allocation and performance modeling in wireless networks in the presence of delay or fading.

This thesis addresses a number of problems in optimal resource allocation and performance analysis in wireless networks characterized by delay and fading. In the first part we address the problem of optimal bandwidth allocation among heterogeneous users sharing the same medium for transmission/reception. We limit our attention to centralized schemes where a controller allocates time slots to users according to their buffer state. The state information is delayed due to physical nature of the system. The objective is to minimize the total expected packet holding cost over a finite or infinite horizon. We first study the properties of an optimal policy for a two-user system and prove that under certain conditions a threshold type policy is optimal and a multiple slot allocation can be reduced to the problem of a single slot allocation. We then consider a myopic policy and study its optimality for an arbitrary number of users in scenarios where feedback delay is either due to propagation delay or duty cycling of the devices. In the second part we study the problem of optimal resource allocation in terms of transmission power and buffer size for a single user in a fading channel. This is formulated as a joint optimal power control (transmission rate) and admission control (acceptance rate) problem. Transmission power, packet buffering and packet dropping all incur costs to the user. The goal is to minimize the total cost over a finite or infinite horizon. We derive a number of properties of the optimal policy. In the last part of this thesis we study the performance of TCP connection splitting proxy over a satellite channel via detailed performance modeling and analysis on file transfer latency. We identify conditions under which using a proxy provides significant or marginal performance gain by investigating factors including initial window size, congestion level of the proxy, and the level of asymmetry between the segments segregated by the proxy. We also discuss how these conditions affect the deployment and provisioning of such systems.Ph.D.Applied SciencesElectrical engineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/125350/2/3192627.pd