Wdm-based Network Planning For Cloud-Radio Access Network Deployment

During the last decades, the tremendous growth in mobile communication services has exerted hot competition for cellular network operators. The subsequent challenges, such as energy consumption, establishment and maintenance cost, intensive spectrum resources, ever increasing traffic demands and profitability concern have exhausted the cellular network operators to achieve better network performance. Based on the existing networks and advances in information and communication industries, the thesis aims at dealing with the above challenges and ensuring the sustainable growth in profitability. Besides, cloud computing which has emerged as promising technique is also introduced to the existing Radio Access Network (RAN). Therefore, the new architecture and deployment of Cloud-Radio Access Network (C-RAN) which is considered as a green cellular networks evolution will be investigated in this thesis.In the thesis, we focus on the problem of Infrastructure Deployment and Layout Planning (IDLP) in C-RAN. Our research work can be categorized into two main parts. First, a mathematic model of IDLP in C-RAN is presented, and Gurobi optimizer is employed to solve the IDLP. To validate the proposed model, various simulations have been carried out and proved to be corrected. The overall deployment cost of the conventional cellular networks solution and our mathematic model solution are also compared. However, the time complexity of Gurobi will be an exponential growth as the network scale increases. Then, Genetic Algorithms and our new heuristic algorithms on IDLP are proposed to the layout results.The above solutions of the IDLP are verified by various simulation scenarios, and feasible solutions can be reached. Among the three solutions, the Gurobi optimizer can achieve an optimal solution. If the network scale is large enough, the Gurobi optimizer could not obtain a feasible solution. But the GA and our proposed algorithm can obtain a near-optimal solution. By comparing the two algorithms, we find GA performance even better than our proposed algorithm. Since that our new algorithm can reach a near-optimal solution of the IDLP very fast. It can get a feasible solution in a relatively short time. Furthermore, the sensitivity tests of the key parameters such as the loss exponent of radio channels and transmission power of RRU are validated by extensive simulations.Finally, we can conclude the research work of this thesis. Our work is expected to be benefitial to the development of the next generation radio access networks with high performance and low investment cost. The optimal framework can provide a guideline to the deployment of C-RAN in practice.