Research On Multi-objective Optimization Connectivity Recovery Technology In Wireless Mesh Networks

Wireless Mesh Network(WMN)has the characteristics of non-center,self-organization,strong invulnerability,convenient deployment and so on.It has great application value in emergency rescue,battlefield communication,environmental monitoring,industrial and agricultural production,public safety and so on.The connectivity between WMN nodes is an important foundation to ensure network performance.When some of the nodes in the WMN are damaged or failed,the connectivity restoration technology can rebuild the communication connections between the network nodes,so as to restore and guarantee the overall performance of the network.The existing connectivity restoration techniques are mainly achieved by moving the remaining nodes of the network or adding relay nodes.At present,the research on the connectivity restoration technology based on the movement of the remaining nodes in the network mainly focuses on the application scenarios of wireless sensor networks,focusing on the connectivity restoration cost.Research on connectivity restoration technology based on adding relay nodes mainly aims at minimizing the number of deployed relay nodes.There are still some shortcomings in the existing research work on multi-objective optimization problems in connectivity restoration technology,and it needs to be further deepened.This thesis focuses on the multi-objective optimization problem in the above two scenarios of WMN connectivity restoration.The main research contents and innovative achievements are as follows:(1)Aiming at the problem of restoring connectivity based on the movement of remaining nodes when a single node fails,In this paper,a Distributed connectivity recovery algorithm based on multi-objective weighted functions(MWF-DCR)is proposed,which jointly optimizes the overall moving distance of nodes,the number of mobile nodes and the coverage loss in the process of connectivity recovery.The algorithm divided the process of distributed connectivity recovery into three stages,namely,the construction stage of connected dominating set,the selection stage of fault handling nodes and the cascading movement stage of nodes.In the construction stage of connected dominating set,the construction method of connected dominating set based on maximal independent set was introduced into the distributed connectivity recovery scheme,and the construction metrics of connected dominating set based on the distance to edge,node degree and overlap coverage were designed for the construction of CDS to ensure that the nodes with more important to the network topology were preferentially selected as the connected dominating set.In the fault handling node selection stage,the fault handling node selection metrics were defined based on the distance to the dominated node,the node degree and the overlap coverage,which were used to select the mobile nodes to ensure that the mobile nodes involved had less impact on the network.In the stage of node cascading movement,the algorithm designed a judgment strategy for the importance of failed nodes,which judged the failed nodes twice to avoid unnecessary node movement.Simulation results show that the proposed algorithm has better performance in terms of mobility cost and coverage loss.(2)Aiming at the multi-partition connectivity recovery problem of large-scale Node failures,this paper proposes an Inner-Hull based Relay Node Placement algorithm(InnerHull)based on the convex hull skeleton.The connection restoration is carried out in the way of trunk node deployment,and the restoration cost(number of trunk nodes)and network connection quality(average node degree,average hops between zones,and coverage rate of unit trunk nodes)of the target balanced network.Firstly,a representative node is determined for each isolated partition,and then a normalization method for relay deployment problem is proposed.By constructing convex hull and minimum envelope rectangle,convex hull sets of irregular isolated partition representative nodes are unified into rectangles.Then,the algorithm designed a strategy to generate trunk backbone ring.Several tunable methods were used to generate trunk node candidate positions inside the normalized rectangle,and the trunk backbone ring was constructed using the trunk node candidate positions inside the convex hull of the representative node.Finally,the isolated partition represents nodes are iteratively connected to the internal backbone ring to complete connectivity restoration.In this paper,according to node degree,node degree mean square error,coverage rate,number of representative nodes in a zone,average hops between zones and other indicators,a network equalization index is proposed,according to which the best trunk ring generation scheme is selected.Finally,a method of constructing a two-connected network based on the node deployment after connectivity restoration is proposed.The simulation results show that the proposed algorithm has a better balance between network restoration cost and network quality than the comparison algorithm.Finally,the thesis summarizes the whole work and looks forward to the next step.