Research On Network Selection And Resource Allocation In Heterogeneous Environments

With the rapid development of wireless communication, the contradiction between limited wireless resources and the various needs of new traffic services is increasing. In the mean time, the existing multiple radio access technologies form the heterogeneous environments. In such background, how to make reasonably and effective network selection for users, and take full advantage of heterogeneous network resources are very vital for easing the contradiction between the wireless resources and the needs of users, which also have attracted widespread attention.This dissertation describes three types of the joint radio resource management architectures for heterogeneous wireless networks. Based on the distributed radio resource management architecture, this dissertation focuses on the subjects of network selection and resource allocation.Network selection, also known as radio access technology selection, is the most important part of accessing control and handoff management. For distributed network architecture, this dissertation presents a user-centric network selection strategy with the purpose of load balancing by using a combination of fuzzy logic and neural network technology.This dissertation studies the problem of joint network selection and resource allocation with regard to the uplink and downlink transmission directions, respectively. For uplink transmission, OFDMA-based radio access technologies are considered, this dissertation forms a proportional fairness resource allocation problem with the condition that users may access all available networks in the multi-user multi-access scene. Furthermore, this dissertation then investigates jointly network selection and resource allocation problem in multi-service multi-user multi-access scene, with the purpose of maximizing the proportion fairness of best effort users, and under the condition that delay constraint users must be satisfied. Since the proposed problem is a mixed integer programming problem, the continuous relaxation is applied, and then the dual problem is solved. Therefore, the two-layer iterative solution, which could achieve the near upper bound, is proposed. Considering the complexity of the two-layer iteration solution, this dissertation further proposes a low complexity heuristic algorithm. Although the performance of the proposed heuristic algorithm is worse than the two-layer iterative algorithm, best effort users can still achieve high fairness.For downlink transmission, this dissertation considers variety types of access technology, such as TDMA, CDMA and OFDMA. To make convenience of study, the different resources are mapped to users’data rate. And then, network selection and resource allocation problems are investigated. With the condition that users may get data rates from all available networks, this dissertation forms a utility based resource allocation problem in the multi-user multi-access scene. Furthermore, similarly to uplink, this dissertation then investigates joint network selection and resource allocation problem in multi-service multi-user downlink scene, with the purpose of maximizing the proportion fairness of best effort users, and under the condition that delay constraint users must be satisfied. Since the formed problem is mixed integer programming, the continuous relaxation is used and the dual problem is solved. Through analysis, this dissertation proposes a jointly network selection and resource allocation iterative algorithm for downlink.