Important Node Mining In Wireless Sensor Networks Based On Complex Network Theory

Wireless sensor networks are distributed networks composed of a large number of small,low-power,and distributed sensor nodes.They are widely used in military,environmental monitoring,aerospace,traffic monitoring,smart healthcare,and other fields.In wireless sensor networks,some nodes play more important roles than others,such as those that handle a large amount of information in the network.These nodes are considered as important nodes in wireless sensor networks.In the research of wireless sensor networks,accurately identifying important nodes is a very important research direction.Discovering these nodes is essential for optimizing network structure,protocols,improving information processing capabilities,enhancing network resilience,reducing the possibility of data loss and attacks,extending network life cycle,etc.Based on the method of identifying important nodes in complex networks,this project mainly focuses on the deficiencies of two evaluation indicators,local structural entropy and clustering coefficient.It proposes two new wireless sensor network node importance evaluation indicators based on complex network theory.The main research work and innovation of the thesis are as follows:(1)The existing methods for evaluating node importance in complex networks are summarized.In response to the defect of the clustering coefficient in the node importance evaluation based on social network analysis,which ignores the size of the node’s neighbors,and the inaccuracy of important node evaluation in high-aggregation networks based on local structural entropy in the information theory method,a new node importance evaluation index,EC,which considers both the clustering coefficient and the local structural entropy,is proposed.This index comprehensively considers the neighbor information of the node and the tightness between the node and the neighboring nodes.This makes the evaluation of node importance more comprehensive,avoids the one-sidedness of a single method and the inaccuracy of evaluation,and finally identifies important nodes more accurately.(2)The topology of the complex network is used as the research object,and it is found that the local structural entropy ignores the situation where nodes and their neighbors form a "loop" structure,that is,a triad closure structure.The local structural entropy repeatedly counts the edges on the "loop," which leads to an overestimation of the calculation result and inaccurate measurement results.Therefore,a new improved local structural entropy wireless sensor important node evaluation index,PLEA,is proposed.This index introduces a punishment term based on the local structural entropy when calculating the node importance on the "loop," to ensure that the local structural entropy can accurately evaluate node importance in high-aggregation networks.(3)Robustness and anti-destructive experiments were carried out on the classic WS small-world network model in complex networks and the real power system network topology structure similar to the topology and network function of wireless sensor networks.Five node importance evaluation indicators were selected for comparison,including degree centrality,k-Shell,local structural entropy,and two other recently published node importance evaluation indicators in the Science journal.The experimental results show that the two new indicators proposed in this thesis can more accurately measure node importance in wireless sensor networks compared to the five indicators.The two new indicators proposed in this thesis have a low time complexity,and only require the local information of the node to accurately measure node importance.In addition,the two new indicators ensure the integrity of the network,that is,they will not destroy the network or affect its normal operation.Therefore,the two new indicators are suitable for node importance evaluation in large-scale and high-aggregation networks,providing a new method for optimizing and managing wireless sensor networks.