Research On Channel Assignment Strategy In Multi-channel Wireless Mesh Network

Wireless Mesh Networks(WMN) is a multi-hop wireless networks that consist of Mesh routers and clients. It has emerged as one of the promising solutions for next generation wireless networks because it could provide high-speed date rate, enlarge service coverage area and reduce network installation cost. Utilizing multiple channels can increase the network throughput, but a proper channel assignment strategy is much more important. In this thesis, the algorithm of channel assignment in multi-radio multi-channel WMN was studied. The achievements are summarized as follow.Firstly, an channel assignment(CA) algorithm based on game theory is proposed to solve the problem of channel assignment in WMN. The algorithm models the nodes as the players, whose strategies are the assignment of channels. The goal of the payoff is minimize the interference of the network. To ensure the fairness of nodes, the algorithm is improved by power adjustment, which could make network resource used sufficiently. Simulation results based on NS2 show that two algorithms have a quite good performance on convergence, throughput and channel access delay.Secondly, a link grading system based on load is adopted for interference estimating, as that interference should be considered in the process of CA. With the optimization aim of minimizing the interference, a CA model based on integer programming is proposed. The Lingo software is used to solve the CA models, and by setting different test scenarios, the validity and reliability of the model are verified, in the form of numerical analysis.Finally, in order to solve the joint optimization of channel assignment(CA) and multichannel routing(MCR), a scheme called ILPGA(Integer linear-programming and genetic algorithms) for WMNs is proposed and studied. Genetic algorithms(GA) is applied to CA problem and integer linear-programming(ILP) is used to tackle MCR. For the interaction of CA and MCR, the fitness value of a chromosome(certain CA configuration) in our genetic algorithms by computing the linear objective function. C# and Lingo are selected for GA and ILP separately. The simulation results proves that ILPGA successfully solves the joint optimization of CA and MCR and effectively increases the network throughput.