Index Scale Weighted Clustering Algorithm For Ad Hoc Networks

Ad Hoc network is a dynamic network which is composed of a collection of wirelessmobile nodes, this network does not need any fixed infrastructure and distributed control, thecharanteristics of Ad Hoc made it widely used in the following cases: infrastructure does notexist or the infrastructure has been damaged. With limited energy of nodes and frequentlychanged network topology, therefore, energy-saving and maintaining the stability of Ad Hocnetwork are always a hot topic. Clustering algorithm is one of the key technologies of Ad Hocnetwork, a good clustering algorithm can effectively manage the resources in Ad Hoc networkand improve the stability of the network structure. The main work of the thesis is the Ad Hocweighted clustering algorithm in-depth research.The structure of thesis is organized as follows: Research background and significance ofthe thesis, the concepts of Ad Hoc network, the application fields of Ad Hoc network and theresearch status of clustering algorithms are introduced in chapter1. The network structure ofAd Hoc, the concepts and objectives of clustering algorithm, the evaluation of clusteringalgorithm, and the comparision of several classical clustering algorithms in Ad Hoc networkare described in chapter2. Chapter3presents an improved Weighted Clustering algorithm,and introduces the cluster head election process and maintenance process in detail. Theemulating platform is constructed by NS2in chapter4, the simulation of the proposedimproving algorithm is also introduced in chapter4. The conclusion points out theinadequacies and the prospects for future research of the thesis.Chapter3and4are the research priorities of the thesis. The energy consumption ofnodes in Ad Hoc network affects the performance of clustering and routing algorithmslargely，to this end, chapter3presents a new method to calculate the energy consumption ofnodes, obtaining the neighbor node changes through periodic Hello messages, replace themobile computing method in this method; further, according to the new node energyconsumption calculation method, the thesis proposes an AHP based index scale, to get thesubjective judgment matrix; using entropy method to get an objective judgment matrix basedon information entropy at the same time; by the combined judgement matrix of the subjectivejudgment matrix and the objective judgment matrix, getting the combined weight of eachnode and complete the clustering. Chapter4describes the simulation of NS2and gives thesimulation of the ISWCA algorithm, and the performance analysis compared with OWCAalogorithm is given.The simulation results show: in terms of the average number of cluster head, theperformance of ISWCA and OWCA have little difference. The number of cluster headupdateing per unit time and the average cluster attachment frequency change per unit time, the performance of ISWCA is better than OWCA, this reflects the Ad Hoc network clusterstructure of IEBWCA algorithm is more stable than the cluster structure of OWCA algorithm.The proposed clustering algorithm is superior in performance. However, this istheoretically, the superiority of the actual environment remains to be further examined; howto reduce the complexity of the algorithm; the proportion of the index scale AHP and Entropyin finding the combined weight; considering the network topology change in Clusteringprocess. All these problems are worth doing in-depth study.