Study Of Dynamic Coverage Problem In Wireless Sensor Networks Based On Multi-Objective Genetic Algorithms

Wireless Sensor Networks (WSNs) have a prospect of wide application, so WSNs attract a growing concern day by day. WSNs generally work in the unmanned areas and their energy is very limited. The limitation of energy constraints the network coverage and network lifetime, so there are two evaluation criteria: network coverage and network lifetime. So a core problem of WSNs is how to place the sensors of the network to meet the need of our network coverage and network lifetime in different situations. This is the"Dynamic Coverage Problem in WSNs"(DCP-WSN).This dissertation presents a model of the DCP-WSN and studies the Multi-objective Genetic Algorithms for this problem. Furthermore, the genetic coding and decoding of Multi-objective Genetic Algorithms is mainly studied.The model of the DCP-WSN includes two components: the definition of network coverage and the definition of network lifetime. Because of some studies consider prolonging the network lifetime as saving energy, we first present that prolonging the network lifetime and saving energy is not the same problem. Moreover, taking into account the character of the energy of sensors, we give the definition of network lifetime. Secondly, considering the definition of the network lifetime and the character of the network placement, the definition of network coverage is given. Ultimately the model of the DCP-WSN is given.The network lifetime in the model of the DCP-WSN is uncertain, which brings much difficulty to the coding in the algorithms. This paper gives three algorithms based on the Multi-objective Genetic Algorithms for the model, which are Prearranging the opening sequence of each time segment, Pre-layout the time segments, Global energy distribution method.The order of open or close the sensors affects the opening sensors'placement and the level of Energy Consumption, so it also affects the quality of the network detection. Based on this we give the Prearranging the opening sequence of each time segment.The goal of the DCP-WSN is to determine which sensors should open in each time segment, which can be directly reflected in the code. Based on this we give the Pre-layout the time segments.As the network is given, the whole network energy is given. Programming all energy of the network in the code, and then distribute the energy in each time segment. Based on this we give the Global energy distribution method.In the instance, we compare the three algorithms based on the model and make some improvements.