Estimation Of Critical Node In Opportunistic Sensor Network

Opportunistic sensor network(OSN)is a kind of self-organization networks,which does not need a complete path between source node and destination node.In OSN,if malfunction of a node leads to destroy the network connectivity and to crash the network,then the node is called critical node.The regions are not connected in multi-regions OSN,and the chance of communication between the region and Sink node comes from movement of Ferry nodes.Therefore,the critical node exists in the Ferry nodes.The program is supported by the National Natural Science Foundation.According to the characteristics of topology changing frequently in multi-regions OSN,this thesis studies on critical node estimation method in multi-regions OSN,so as to take corresponding measures in advance,optimize the network structure and to improve the robustness of the network.Furthermore,it can provides reference for multi-regions OSN topology control.This thesis analyzes the importance indexes of node,the existing estimation methods of critical node,and introduces the multiple attribute decision making theory.The thesis proposes a critical node estimation method based on game theory TOPSIS with combination weighting(GTCW_TOPSIS)for the multi-regions OSN.The multi-regions OSN is modeled by temporal reachability graphs to obtain the dynamic topology information.The average degree of Ferry nodes is defined to reflect the activity of Ferry nodes.The betweenness centrality of Ferry nodes is defined to reflect their impacts on the message path between region and Sink node.The network forwarding rate of Ferry nodes is defined to reflect their contribution to the message delivery in network.They are three attributes of Ferry nodes.Analytic hierarchy process(AHP)is employed to determine the subjective weight of each attributes,and entropy weight method is employed to determine the objective weight of each attributes.The final weight of each attributes is calculated by game theory with combinational weighting.The critical node is estimated by the technique for order preference by similarity to ideal solution(TOPSIS).The thesis designs three typical multi-regions OSN scenarios based on a simulation platform,Opportunistic Network Environment Simulator(ONE).The node removal method is employed to validate the estimation results correctness of GTCW_TOPSIS.Compared with the TOPSIS,the GTCW_TOPSIS has a better estimation effect.The thesis analyzes GTCW_TOPSIS and the correlation of attributes.According to the analysis,an improved estimation method of critical node is proposed,named Mahalanobis Distance GTCW_TOPSIS(MA_GTCW_TOPSIS).Compared with the TOPSIS and GTCW_TOPSIS,the MA_GTCW_TOPSIS has better accuracy.