Research On Topology Evolution And Routing Strategy For Mobile Ad Hoc Network

With the increasing demand of wireless communication, various wireless communication networks have been rapidly developed. However, there are some shortcomings in the existing networks, such as high costs to build and maintain the infrastructures, difficult achievement for the seamless coverage and inflexible deployment. Therefore, mobile Ad Hoc network (MANET) is proposed, which is more flexible deployment, self-organizing and without infrastructure. So it is a good solution to the communication problems of the mainstream wireless communication systems. MANET has a better development outlook because of its advantages. However, there are some limitations in MANET, such as complex network topology, inefficiency of routing protocols and lower link utilization. Based on the complex networks theory, the in-depth researches have been done on topology evolution, routing strategy and the use of unidirectional links of MANET in this dissertation.Through the combing and analysis of related references about MANET, the technical route of this research is structured and the subsequent research foundation is established.In this dissertation, the concepts and developments of the MANET are studied and the application scenarios of the networks are summarized. Then the main researches and network characteristics are declaimed in detail. The impacts on MANET from topology and routing are focused on. The complex networks theory was exposited. Especially the studies of the topology evolution and routing strategy are analyzed and discussed deeply.In order to describe the evolving features of MANET accurately and improve the performance of the network, an evolving topology model with local-area preference is proposed. The essential idea of the model analyzed by the mean field theory is to optimize network structure based on users’behaviors in MANET. The evolving model changes the previous research method which is based on the uniform topology when establishing network structures. And the network topologies with the actual characteristics can be set up according to the evolving model. In addition, the limited energy of MANET’s nodes was considered in this model. The connections among nodes are required in accordance with the principle of energy optimization, which can make the node using the energy more efficiently and improve the robustness of networks. The simulation results show that the evolving model which is proposed by the dissertation can reflect the characteristics of MANET better, and can prolong the lifetime of the networks. It is an available approach for establishing and analyzing the actual MANET.To deal with the routing efficiency of MANET, the Betweenness-optimized Energy-balanced Local Routing strategy is proposed in the dissertation. This strategy fully considers the non-uniform characteristics of networks to probabilistically select the most appropriate node as next hop according to the betweenness and the remaining energy of node. The simulation results show that this strategy can effectively improve the transmission capacity of the network. On this basis, the Energy-balanced Joint Routing mechanism is proposed. This mechanism further refines the structure of MANET to use the different routing strategy in different local structures for the improvement of the transmission capacity and throughput. The simulation shows that the mechanism can achieve the purposes of increasing network load, balancing of energy consumption and prolonging the network lifetime.In view of weighted network characteristics existed in MANET, in this study, a Dynamic Shortest Path Weighted Routing mechanism was proposed in the case of comparative analysis of different weighting methods. This mechanism, which can update network weights in real time, contributes to balance transmission load of nodes and improve transmission capacity of networks. The simulations show that the proposed mechanism can have good results in some aspects, such as balancing energy cost, reducing the failure rate of nodes and improving network throughput.To address unidirectional links which widely exist in MANET, two unidirectional link-state advertisement mechanisms are designed, i.e. unidirectional link-state advertisement based power adjustment (ULAPA) and Unidirectional-Link Advertisement based Dynamic Grading Power (ULADGP). Using the mechanisms presented by the dissertation, unidirectional links can be sensed by nodes at both ends of the link under the premise of reducing energy consumption of nodes as much as possible. According to the results of unidirectional link-state advertisement, the Local Routing based Unidirectional Link-state Advertisement (LRULA) proposed by this study can more efficiently transfer data. The simulation results show that the proposed advertisement mechanisms can reduce the end-to-end delay, decrease the average path length and maximize network utility.Finally, the summary of the thesis was given. The shortcomings of the research were pointed out and some suggestions to the future research were listed.