On Utility-based Resource Allocation Strategies And Algorithms In OFDMA Networks

Recently, the wireless communication technology keeps fast development. With the increase in the capacity of the wirless communication systems and the popularity of the mobile intelligent terminals, besides the traditional voice service, the user demand for data communication and multimedia service is becoming higher. Therefore, a reasonable resource allocation mechanism which involves a variety of different service in the resource allocation problem is essential. Orthogonal frequency division multiple access (OFDMA) network can allocate the available bandwidth to users according to their requirement dynamically, thus utilizing the system resource efficiently. Therefore, the resource allocation problem in OFDMA networks is hot in recent years. The object of OFDMA resource allocation is to allocate resource including subcarrier and power to multiple services efficiently, and to guarantee the QoS (Quality of Service) of each user.In our thesis, we first investigate the ergodic resource allocation problem for various services (heterogeneous services) scene and propose an optimal resource algorithm based on utility function. The algorithm which satisfy the QoS requirements of services makes the sum of average utility maximized. The simulation results show that as the algorithm considers the rate constraints and utility function of heterogeneous services, it can get better average rate and utility than the other algorithm and be statistically optimal.The resource allocation in practical OFDMA network should consider many factors comprehensively, including the fast convergence of algorithm. Therefore, in our thesis we propose a practical cross-layer resource allocation algorithm. The algorithm combines the instantaneous channel state in physical layer, the queue state in data link layer, and the user’s QoS parameters, it aims to get the near optimal results to maximize the instantaneous user utility. The simulation results show that the algorithm can get better throughput and lower package loss rate than other algorithms, which performs well.