Research On Traffic Offload Strategy In Wireless Heterogeneous Networks

In recent years, the explosive growth of various mobile devices and wireless data traffic in the mobile cellular network generates serious traffic congestion problems within this network. The traditional homogeneous architecture has been incapable of coping with the continued growth in data volumes and user requirements for communi-cation quality. Therefore, heterogeneous networks (HetNet), as a more flexible network architecture model, have gradually become a hot topic among academic and industry fields. In HetNet, there are a lot of low power nodes (LPNs) under the coverage of a macro base station. Most of them are deployed in hot spots or blind spots in macro-cellular network, and they can provide better services to users with much lower en-ergy consumption than the macro base station. Thus to offload the data traffic of the macro users to the low-power base stations is considered to be an effective technologi-cal means to reduce macrocell network’s load, increase the system capacity as well as reduce the energy consumption in the whole network. However, how to economically and effectively offload traffic involves multiple aspects such as interference coordina-tion, mobility management, energy efficiency as well as the offloading costs. In order to implement a proper allocation of network resources and improve the network perfor-mance on the background of NetNet, this thesis carries out the traffic offload strategy research under different application targets, mainly solving the user association issues in the strategy.When the macro and small cells are in cooperation mode, we study the joint user association and interference coordination problems in HetNet traffic offload stragety. For the purpose of tradeoff between system throughput and fairness among users, we aim at maximizing the minimum weighted user rate in the network-wide scale and try to solve the optimal user association and almost blank subframe (ABS) proportion value in HetNet. This is a mixed integer nonlinear programming problem which is difficult to tackle, so we decompose the original problem into two sub-problems. First of all, solve the optimal ABS ratio problem under the given user association results and derive the closed formulas of optimal ABS proportion corresponding to user association. Based on this expression, an iterative algorithm of solving the user association with low com-putational complexity is further proposed. Simulation results show that the proposed algorithm is fast convergent, and helps improve the resource utilization efficiency.When the macro and small cells are in the non-cooperative mode, the offloaded macro users will occupy the limited bandwidth resources and influence the user capac-ity in small cells, so the small cell owner is unwilling to provide free service to macro users. Under this circumstance, we solve the user association and price paid policy in HetNet traffic offload strategy based on an auction mechanism, which incentivses small cells to rent their resource for the offloading purpose. Firstly, the utility functions of macrocell bid and small cell bid are given. Then, solve the best user association in traffic offload strategy using a linear programming model and calculate macrocell corre-sponding payment rule based on VCG (Vickrey Clarke Grove) mechanism, which guar-antees the market truthfulness and individual rationality as well as improves network performance. Further, a greedy algorithm is proposed to reduce computation complex-ity problem while keeping good economic properties, making the auction more practical for implementation.