Optimal Joint Resource Allocation In Multi-Channel Relay-Assisted Wireless System

In recent decades, wireless communication is always an important branch in communication research area. To satisfy users request for the quality of service (QoS) and expand the wireless service area, engineers have designed the relay based wireless networks, such as Ad-hoc, mesh. To reduce the cost and improve the wireless service coverage and system throughput, relay has become one of the key characteristics in 4G (fourth generation) wireless communication stan-dards. Meanwhile, because of the excellent work to eliminate the inter symbol interference (ISI), orthogonal frequency division multiplexing (OFDM) is also adopted as a key technology in 4G wireless communication standards.Although there are many developed new technologies, the conflict between the limited re-sources, such as transmission power, subchannels and bandwidth, and the increasing QoS re-quirement, such as real time video conference, real time online game, becomes more and more urgent. As an effective way to relieve this conflict, resource allocation algorithms design is al-ways the hot topic for scientists and engineers. The optimal resource allocation can improve the efficiency of the resource utilization, then satisfy users QoS simultaneously. In 4G standards, OFDM and relay deployment complicates the resource allocation scheme design. Compared with the traditional single-hop cell system, in multi-channel multi-relay system, we not only need to consider the power allocation and subchannel allocation, but also need to study relay selection and subchannel pairing. In OFDM based multi-relay system, for each subchannel of the source at the first hop, a particular subchannel is selected at the relay, which helps source accomplish its transmission in the second hop. This method is called subchannel pairing. Considering joint power and subchannel allocation, relay selection and subchannel pairing to design the optimal resource allocation scheme is the main subject of this dissertation. Most papers focused on the resource allocation in deterministic model in which users have instantaneous data rate requirement in each time slot. This model comply with the real time transmission request. However, when QoS is non-real time, such as email, FTP file transmission service, we only need to satisfy their average transmission rate during the transmission process. What is the difference on resource allocation algorithms between real time model and non-real time model? The answer is that for non-real time services, the model is stochastic. How to design the joint resource allocation in stochastic model is another research topic in this dissertation.In single hop system, frequency reuse is employed to improve the spectrum utilization ef-ficiency. However, frequency reuse also brings the inter cell interference(ICI). To reduce the interference, power control has to be adopted. Comparing to the non-relay system, the presence of relay would complicate the power control scheme to do the interference coordination. The performance of user served by relay not only coupled with that of other users who utilize the same frequency band, but also related to the time. This is because that relay-assisted user per-formance is decided after two time slots. On the other hand, the relay deployment would bring the intra cell interference (ITCI), this is because different users served by different relays in the same cell could utilized the same frequency band as long as they are far enough. How to design the power control scheme to do the interference coordination in multi-channel, multi-relay and multi-cell system is another focus in this dissertation.We focus on the multi-channel relay system in this dissertation. Firstly,in Chapter II, we design the joint optimal power allocation, subchannel allocation, relay selection and subchan-nel pairing in multi-channel, multi-relay, single pair of source-destination node system. Since the optimal algorithm needs overall channel state information (CSI) feedback with higher com-putational complexity in central controller, it is impractical. Hence, in we design the distributed scheme and distributed scheme with limited feedback respectively. In the system we do not study the users QoS requirement. Hence in Chapterâ…¢, we design the opitmal joint resource allocation scheme when users have real time and non-real time transmission data rate requirements, which are called deterministic and stochastic models, respectively. The resource allocation scheme in stochastic model are more efficient than that in deterministic model because it can exploit the multi user diversity gain in temporal domain. Finally, in Chapter IV, we consider how to do the interference coordination by power control in multi-channel, multi-cell relay system. From the analysis, we observe Geometric Programming (GP) could be used to design the optimal power control scheme. However, the scheme is centralized and need heavy signaling, it is impractical and only can be treated as an theoretical upper bound of the performance. Hence, by Game Theory we design the semi-distributed and full-distributed schemes respectively.