Knowledge-guided Evolutionary Optimization Strategy For Connective Coverage Control In Wireless Sensor Networks

The coverage optimization control plays an important role in using wirelesssensor networks (1) It will reduce the redundancy of perception data and improve thecommunication efficiency.(2) It will decrease the energy consumption and extendthe life of network.How to control the state of wireless sensor node and choose therouter is the core issues of the coverage optimal for control strategy of WSNs. In thethesis we focus on the selection method of the active nodes and the restore algorithmfor the invalid network. To select active nodes, the detection algorithm for redundantnode and the construction strategy for connected coverage sets were discussed.Themain contents of this thesis include the following three aspects:(1)To improve the accuracy and the efficient of the detection algorithm forredundant nodes, a novel detection algorithm for redundant nodes based on coveragetemplate was proposed.The coverage template can judge whether a grid in themonitor area is covered by nodes. Thus the computational complexity of algorithmwas lower.Besides,the strategy can effectively identify redundant nodes near themonitor area’s border.Further more,the redundancy of node was defined by a numberbetween0to1.(2)An optimization model considering energy-consumption and connectivity-coverage was constructed.And then, a knowledge-inducing evolutionary optimizationmethod for constructing connected cover sets with minimum energy-consumption wasproposed. The active node’s importance was defined as a knowledge based on thelocation and the distance from sink node, which instructs individual coding andmutation operation.The experiments results show that the knowledge can beeffectively guide evolutionary direction, and accelerate the convergence rate.(3)A node scheduling approach based on the pretreatment for invalid wasproposed in thesis.In order to reduce the number of neighborhood nodes, theneighborhood radius was gotten in terms of the importance of active nodes.Thecandidate node-sets was constructed according to the neighborhood radius so as toreplace the invalid node.A local replacement strategy is presented based on the activesnode’s importance.When network is invalid caused by failed nodes, the dynamicrecovery strategy will select a node from the candidate node-sets to restore thenetwork so as to ensure the full connectivity and satisfy the coverage requirement.Researches achievements not only enrich the theories of coverage optimization of wireless sensor network,but also provides the foundation for the application ofwireless sensor networks.