Research On Partition Connectivity Optimization In Wireless Sensor Networks

With the rapid development of electronic technology, wireless sensor network(WSN) has been widely used. Wireless sensor network is a network of self-organized formation by a large number of micro sensor nodes, network connectivity, which is the basic requirement of network communication. Due to the consumption of energy as well as the impact of other external factors, nodes in wireless sensor networks can easily encounter the problem of large-scale failure, thus getting the networks divided into several independent partitions which can’t effectively communicate with each other, affects the service function of the network. The thesis mainly solves the partition problems in wireless sensor networks from the perspective of optimization, namely in a given wireless sensor networks, how to use fewer relay nodes to complete partition connectivity, while the restored network has a better connectivity and fault tolerance. Here are main contents completed:(1) Cobweb algorithm is studied, this paper analyzes the though of the algorithm in the process of partition connectivity, and through the example analysis, from which can see the algorithm implementation partition connectivity by deploying a large number of relay node. On the other hand, the study found that existing a large number of cut points in the restored network, arbitrary cut point of failure will cause partitioned again. In order to reduce the number of relay nodes and improve fault tolerance of the restored network, the paper gives new algorithm to resolve the problem for single connected and bi-connected respectively on the basis of the work.(2) Connectivity algorithm in partition based on triangle Steiner tree which is called CAPTST is presented. The algorithm constructs triangle Steiner tree by introducing Steiner point. At the same time, the algorithm conducted twice optimization of the heuristic and Steiner deployed in the process of deployment of relay nodes, to complete the network connectivity with maximize reduce the number of relay nodes. Experimental results show that CAPTST can reduce the required number of relay nodes of the partition connectivity, reduce the communication distance and energy consumption between partitions, improve the network performance compared to existing 1C-Spider Web algorithms under the same experimental conditions from the number of relay node, average node degree and communication hop count.(3) Double connectivity algorithm in partition, namely DCAP, is proposed. The algorithm is aimed to build the backbone polygon in the center area of the network by introducing optimal node deployed, then partitions are connected to the polygon by two disjoint paths, so as to realize double connected between partitions. The experimental results are found by comparing with existing connectivity restoration with assured fault tolerance algorithm and 2C-Spider Web algorithm, DCAP not only reduces the number of relay nodes deployed, but also can more quickly determine the location of the relay nodes deployed, so as to improve the timeliness of partition connectivity.