| Περίληψη: | In this thesis, the problem of resource allocation is investigated in multiuser, multi-antenna downlink wireless systems in which spatial multiplexing is employed in the physical layer. The thesis consists of two main parts; in the first, the interest is focused on optimizing system's performance in terms of users' transmission rate. Under this context, a low-complexity but highly performing user selection algorithm is presented for the flat-fading channel, when zero-forcing beamforming is employed at the BS and the aim is to maximize system's throughput. For the more interesting case where the transmission is performed simultaneously over a number of parallel subchannels, two fairness-aware resource allocation problems are investigated in the sense that certain QoS constraints are considered. Typically, this kind of problems fall within the NP class because of the integer nature of the involved user selection procedure. Hence, several near-optimal and heuristic solutions are proposed. In the second part of the thesis, the concept of physical layer security is integrated into the resource allocation procedure and the secrecy rate becomes the critical quantity of the considered problems. The considered setup consists of a multiuser downlink system in which a passive eavesdropper tries to wiretap the message of one or more users. First, the problem of secrecy rate balancing is investigated for a SISO downlink system in which transmission is performed over a set of parallel subchannels. Under the assumption that each subchannel is occupied exactly by one user, several optimal, near-optimal and heuristic solutions are proposed for several different problem settings. By assuming a multiantenna BS, the resource allocation problem is further enriched with spatial multiplexing within each subchannel. Thus, the thesis is completed with the investigation of two such problems in which secrecy-rate QoS constraints are also taken into consideration.
|
|---|
