| Wireless sensor networks,which are similar to small-scale internet,are composed of a large number of micro sensor nodes which have wireless communication capabilities.As a new research topic in in the field of information and data,there has been a lot of research work about wireless sensor networks.Because of the fact that the sensor node's power has small capacity and is non-rechargeable,improving energy utilization and extending lifetime of wireless sensor networks are the main research goals.LEACH is a classic distributed and self-organizing clustering protocol which has a broad influence.The main idea of LEACH protocol is to use a distributed clustering algorithm.That is,all nodes participate in the selection of the cluster heads.The mechanism is relatively simple to implement and the overall energy consumption of the network can be evenly distributed.The use of data digests and fusion technologies also reduces the energy consumption of network transmissions.However,it also has some defects.The formula ignores the impact of remaining energy and geographical location of nodes,which makes the node which are dying and in sparsely location can be selected as the cluster head.The value of p needs to be preset manually which in fact is very difficult work to get the idea value for good clustering.Hypergraph is a basic data structure and mathematical model which can effectively represent the relationship between nodes.Hypergraph has wide applications in image retrieval and video segmentation,however,few researches have been done in WSN field.The traditional hypergraph partitioning algorithms also have obvious drawbacks:the number of clusters needs to be preset manually and there is no reasonable standard for justifying the quality of the classification.Aiming at the defects mentioned above of LEACH,from the perspective of clustering this thesis combines wireless sensor networks with graph model and proposes a new clustering algorithm based on the hypergraph theory which models and clusters considering the density and the residual energy of the sensor nodes.Considering the drawbacks of traditional hypergraph partition,the thesis proposes a new conception which is called modularity.The formula of modularity is defined from the perspective of low coupling between clusters and high cohesion within a cluster.It's considered as the optimization criteria for hypergraph partition and one of the conditions for the hypergraph termination.In addition,the thesis designs and implements a hypergraph multi-level iterative clustering segmentation algorithm,which continuously iterates over the hypergraph modeled by the wireless sensor network,and finally forms a good network topology structure.Considering the computational complexity of the algorithm,the thesis optimizes the cutting point so that the computational complexity of the optimal cutting point is only linearly related to the hyperedge containing the cutting point,which greatly reduces the complexity.In the end,the simulation experiments are carried out for the algorithm and LEACH.The algorithm is compared with LEACH in three aspects.The experimental results show that the proposed algorithm is superior to LEACH in terms of energy consumption and network lifetime,which verifies the rationality and effectiveness of the proposed algorithm. |
