Research On Routing Algorithms For Opportunistic Networks

Opportunistic networks are an evaluation of mobile ad hoc network and a variation of disruption tolerant network, which utilize the encounter opportunities arising from node mobility to complete communications, and resolve the message communication problem in frequent interruption networks. Because of nodal mobility, sparse network, limited storage and energy etc., message routing in opportunistic networks is confronted with great challenge, including the source and destination node in opportunistic networks lack an unbroken end-to-end path at the same time and greater latency. As a result, the traditional routing algorithms for mobile self-organizing network can not work well in opportunity networks, it is valuable and significant to design routing algorithms according to the characteristics of opportunity network, and this is also the main work in the paper.In this paper, we firstly classify and compare existing typical routing algorithms in opportunistic networks. From the routing algorithms based on flooding, to based on fixed flooding and utility, researchers at home and abroad have done a lot of work about the routing problem in opportunistic networks, mainly focusing their efforts on the performance of routing algorithms, the complexity of programming, the validity and availability of algorithms. The routing algorithms based on flooding increase the success delivery probability, and reduce the transmission delay of message at the expense of large amounts of network resources consumption. Then researchers come up with the routing algorithms based on fixed flooding to limit message replications, and reduce the cost of network resources. Routing algorithms based on utility, give certain forwarding utility value to each node exploiting different types of network knowledge. By limiting the number of copies of message, these kind of routing algorithms reduce the waste of the network resources while ensuring higher success delivery probability.It has been proved that, in opportunities networks, forwarding messages with the similarity between people can reduce the number of copies (network resources consumption), and increase the possibility of a message been forwarded to the destination node. This paper propose a routing algorithm based on SimRank model, for those classic routing algorithms did not fully consider the similarity between nodes. It adequately considers the structural similarity between nodes, chooses the nodes who have higher similarity with current node carried message as the next hop nodes. The simulation results on three real mobility traces of nodes show that, RBSR can get better performance than three other classic routing algorithms.Communication in opportunities networks entirely depends on the encounter opportunities arising from nodal movement. The history contact information between nodes is of great significance for link prediction and message routing. Then a class of routing algorithms based on node meet history information is put forward in Chapter4. They introduce social relations weights extracted from the history information to the weighted similarity index (CN, RA, AA), compute encounter probability value between nodes as the utility value of forwarding messages, compare the utility value between destination node and their neighbor nodes, select the nodes with higher utility value as the next-hop forwarding nodes. Results on ONE platform show that compared with the Prophet algorithm, this kind of algorithms have a good performance on delivery probability, network overhead and average delay.