Cooperative Resource Allocation in Wireless Communication Networks

The concept of cooperation where two or more parties work together to pursue a common goal, is applicable in almost every aspect of today's life. For instance, in the upcoming car-to-car communications, the vehicles exchange information regarding their current status and potential threats on the road in order to avoid accidents. With the evolution of the wireless communication systems and the advent of new services and devices with more capabilities, the demand for higher data rates is ever increasing. In cellular networks, the achievable data rates of the users are limited by the inter-cell interference, which is caused by the simultaneous utilization of the time/frequency resources. Especially, the data rates of the users located at the vicinity of neighboring base stations is affected by the inter-cell interference. Hence, in this dissertation, cooperation in cellular communication downlink networks is investigated, where the base stations coordinate their operation in order to mitigate the impact of co-channel inter-cell interference. Thus, the constantly increasing user demand can be satisfied. Cooperative resource allocation schemes are derived, where practical conditions and side constraints regarding the available channel state information at the base stations are taken into account. Cooperation in the form of power control and joint time/frequency scheduling is mainly studied. In the former type of cooperation, the base stations dynamically adjust their own transmit powers to cause less inter-cell interference to the users connected to neighboring base stations. In the case of cooperative scheduling, the available time/frequency resources are jointly allocated by the base stations in order to trade off user throughput and inter-cell interference. The cooperative scheduling schemes apply two special cases of the power control approach, where the base stations either serve their connected users with maximum transmit power, or abstain from transmitting data, i.e., muting, in order to reduce the interference caused to users served by neighboring base stations. One major contribution of this work is the formulation of the cooperative resource allocation problems by considering the availability of channel state information at the transmitter in form of data rate measurement reports, which follows standard compliant procedures of current mobile networks such as LTE and LTE-Advanced. From a system perspective, two parameters are considered throughout this dissertation in order to derive the proposed cooperative schemes. These parameters are the cooperation architecture and the traffic model characterizing the demand of the connected users. In the case of the cooperation architecture, centralized and decentralized schemes are studied. In the former, a central controller performs the cooperative schemes based on global knowledge of the channel state information, and in the latter, the cooperative decisions are carried out independently per base station based on local information exchanged with adjacent base stations. It is expected that the centralized architecture provides the best performance, however, the gap with respect to the decentralized approaches reduces significantly under practical network assumptions, as demonstrated in this work based on numerical simulations. With respect to the traffic model, the user demand is characterized by full-buffer and non-full-buffer models. The first model is applied in order to assess the performance of the proposed cooperative schemes from a capacity enhancement perspective, where all users constantly demand as much data as possible. On the other hand, the non-full-buffer model represents a more practical network scenario with a dynamic utilization of the network resources. In the non-full-buffer model case, the proposed schemes are derived in order to improve the link adaptation procedures at the base stations serving users with bursty traffic. These link adaptation procedures, establish the transmission parameters used per serving link, e.g., the transmit power, the modulation and the coding schemes. Specifically, a cooperative power control scheme with closed-form solution is derived, where base stations dynamically control their own transmit powers to satisfy the data rate requirements of the users connected to neighboring base stations. Moreover, centralized and decentralized coordinated scheduling with muting is studied to improve the user throughput. For the centralized case, an integer linear problem formulation is proposed which is solved optimally by using commercial solvers. The optimal solution is used as a benchmark to evaluate heuristic algorithms. In the case of decentralized coordinated scheduling with muting, a heuristic approach is derived which requires a low number of messages exchanged between the base stations in order to coordinate the cooperation. Finally, an integer linear problem is formulated to improve the link adaptation procedures of networks with user demand characterized by bursty traffic. This improvement results in a reduction of the transmission error rates and an increase of the experienced data rates. With respect to non-cooperative approaches and state-of-the-art solutions, significant performance improvement of the achievable user throughput is obtained as the result of applying the proposed cooperative schemes, especially for the users experiencing severe inter-cell interference