Optimized resource allocation in wireless systems

University of Technology, Sydney. Faculty of Engineering and Information Technology.Modern wireless systems rely to a great extent on the judicious distribution of available resources (e.g. power, bandwidth) to meet an ever increasing demand of better quality-of-service (QoS). Scarcity of these resources with time, coupled with the tremendous growth in numbers of users, network throughput, and applications, have resulted in making the problem of optimal resource allocation extremely important especially in wireless networks. Generally, optimization problems posed in the resource allocation framework are nonconvex and thus render it difficult to find an optimal solution. Previous studies on this subject have reported only numerically cumbersome and non-tractable solutions. This dissertation attempts to exploit the hidden convexity of the resource allocation problems under some given performance criteria such as minimum mean square error (MMSE) or signal-to-interference-plus-noise ratio (SINR) and then successfully finds tractable optimization formulations. The first research problem deals with the optimal power allocation and sensor assignment in linear and nonlinear networks for static and dynamic target tracking. The proposed method casts power allocation as a semi-definite program (SDP) while sensor selection is solved via d.c. (difference of convex functions/sets) programming. The second problem considers optimal beam-forming and source power allocation in relay-assisted multiuser communication. This problem is further extended to include multiple-antenna systems to exploit spatial diversity in modern cellular communication by jointly optimizing source precoding and relay processing matrices. Supremacy of the proposed d.c. programming based iterative algorithm over existing methods is demonstrated via extensive simulations