A Study On Resource Allocation Algorithms For Small Cell Networks

With the rapid development of the mobile Internet, more and more mobile devices come into our daily life, and people's demand for data traffic is rapidly increasing. Due to the lack of radio spectrum resources, we are facing a severe spectrum shortage problem. To address this problem, the concept of small cells has been introduced in mobile cellular networks. Research in both industry and academia has shown that the deployment of small cells in traditional cellular networks offers advantages of larger system throughput, improved spectrum efficiency and reduced overload of macro base stations in the long run. On the other hand, the deployment of small cells would introduce more interferences in cellular networks. Such interferences can be effectively mitigated through efficient spectrum resource allocation. Therefore, the design of efficient spectrum resource allocation algorithms becomes an important research topic in small cell networks.This thesis studies the spectrum resource allocation problem in small cell networks, and proposes three spectrum resource allocation algorithms for different system scenarios, including a high spectrum efficiency resource allocation algorithm, a distributed and dynamic resource allocation algorithm, and a fairness-baesd distributd and dynamic resource allocation algorithm.Firstly, a high spectrum efficiency resource allocation algorithm is proposed for maximizing the capacity of a cellular system. The algorithm considers a general system scenario with a single macro cell and multiple small cells. In the system, spectrum resource allocation is performed by the macro base station in a centralized control manner. The number of macro users is fixed and small cells share the spectrum resources of macro users. The algorithm consists of two allcation phases:on-demand resource allocation and specturm-reuse resource allocation? The objective of the first phase is to ensure users'basic communication demands, while the objective of the second phase is to improve the spectrum utilization of the system. To mitigate the inter-cell interferences, the algorithm introduces the concept of a protected area and does not consider those small cell users within the protected area of a macro user. Meanwhile, to mitigate the intra-cell interferences, it introduces the concept of an interference table To ensure that the spectrum resources allocated to a small cell user is not contained in the interference tables of the user's neighbor small cells. Simulation results show that the proposed algorithm can significantly improve the spectrum resource utilization while ensuring the system throughput compared to a traditional random resource allocation algorithm.Secondly, a distributed and dynamic resource allocation algorithm is proposed for a system scenario with a high density of small cells, where a centralized control mode would bring a large amount of computational burden and control overhead. The algorithm considers a dynamic scenario where mobile users arrive in and leave a small cell dynamically. To support the algorithm, the macro base station and each small cell base station need to maintain a couple of interference tables, SITable and WITable, and a couple of resource status tables, SRSTable and WRSTable. Based on the inteference tables and resource status tables, each base station performs RB allocation for its users. For each new user, those RBs that would cause a strong interference are not considered in allocation. Only those RBs that would not cause any interference or only cause a weak interference are considered. Simulation results show that the proposed algorithm can effectively increase the average data rate in the system and achieve better user satisfaction ratio as compared to a random resource allocation algorithm.Finally, the thesis proposed a fairness-based resource allocation algorithm to improve the user fairness. Simulation results show that the proposed algorithm can effectively increase the user fairness in the system as compared to the distributed and dynamic resource allocation algorithm, while maintaining the system throughput.