Research On Resource Management Of Device-To-Device Uniderlaying Network

Recently, according to the increasing requirement of communication data rate and high energy efficiency, wireless communication system is evolving towards more efficient and new wireless architecture. However, due to the spectrum limitation, traditional wireless network is not competent to these challenges. Therefore, with the development of heterogeneous network,3GPP Long Term Evolution (LTE) desires to transfer core network based wireless architecture to user equipment based pattern. Under this circumstance, Device-to-Device communication (D2D) has drawn widely consideration in both academic research and commercialization development. D2D communication can increase resource efficiency and system throughput. Besides, it has huge potential in reducing transmit power, achieving load balance and improving energy efficiency. In this thesis, we mainly focus on resource allocation scheme and system performance, like interference mitigation, system capacity, energy efficiency, etc. The main research contents of this thesis are shown as following:To begin with, this thesis focuses on the resource allocation problem between cellular communication and D2D communication. Then a Self-adaptive Genetic Algorithm (GA) algorithm based D2D uplink resource allocation is proposed to increase resource utilization as well as reduce interference. Specifically, innovations like multi-dimensional binary individual coding method, self-adaptive mutation and elite strategy can significantly enhance the wide genetic diversity of GA. Simulation results and analyses show that the proposed GA based resource allocation scheme can achieve30Mbps system throughput gain comparing with the classic method. With lower computation complexity and higher convergence rate, GA is demonstrated as better scheme to deal with resource allocation problem.Moreover, considering the more normal scenario in practical wireless network, where there are massive D2D users. In this situation, more potential of resource utilization should be tapped, such as resource-multiplexing by more users. Firstly, we propose the k-means algorithm based D2D cluster establishment scheme. The cluster scheme can divide user equipments into different clusters, and different clusters can reuse the same resource orthogonally with little interference. Besides, we proposed the enhanced Genetic Algorithm based resource multiplexing scheme to allocate resource into different clusters. Finally, through theoretical derivation and simulation results, the proposed scheme is demonstrated as the most feasible and efficient approach to improve the device energy efficiency than Greedy based scheme.At last, we consider the joint optimization of uplink and downlink time slot and frequency resource in radio-beam transmitting D2D network. Since current research only focuses on frequency resource optimization, considering resource utilization in other dimensions, such as time and space, will provide more benefits to increase system performance. In this thesis, we present detailed steps and scheme to optimize resource utilization in both time slot and frequency. At first, Base Station obtains the information about the location, antenna configuration and the time slot allocation situation of different D2D users. Secondly, the vertex coloring scheme is used to allocate frequent resource for different D2D users. And then, we utilize genetic algorithm based optimization scheme to find the optimal time slot. Finally, simulation results demonstrate that combining with different optimization algorithms, the NP-hard problem in the two-dimensionality resource allocation scenario can be addressed with better performance.