In this paper, we address clustering in ad hoc networks. Ad hoc networks are a wireless networking paradigm in which mobile hosts rely on each other to keep the network connected without the help of any pre-existing infrastructure or central administrator. Thus, additional features pertinent to this type of networks appeared. In fact, centralized solutions are generally inadaptable due to the need for cooperative network operations. To ensure efficient, tolerant and durable cooperative operations, nodes need to organize themselves. Clustering is an organization method which consists in grouping the nodes into clusters (groups) managed by nodes called clusterheads. In this paper, we present existing clustering algorithms and propose a new solution inspired from two of these algorithms (Lowest Id and WCA). This solution, called Lowest Weight, exploits their advantages and relieve to their drawbacks in terms of clusters stability and computational overhead. Simulation experiments were conducted to evaluate the performance of the algorithm proposed in terms of clusters numbers, clusterheads lifetime and the number of reaffiliations (node moving from a cluster to another). Results show that Lowest Weight ameliorate performs of existing algorithms especially regarding mobility leading to more suitable, adaptable, scalable and autonomous clustering.
展开▼
机译：在本文中，我们在ad hoc网络中解决了群集。 ad hoc网络是无线网络范例，其中移动主机依赖于彼此，以便在没有任何现有的基础架构或中央管理员的帮助下保持网络连接。因此，出现了与这种网络相关的附加功能。实际上，由于需要合作网络运营，集中解决方案通常是无可置的。为确保有效，宽容和耐用的合作操作，节点需要组织自己。群集是一个组织方法，它包括将节点分组为由名为群集头部管理的节点管理的群集（组）。在本文中，我们呈现了现有的聚类算法，并提出了一种从这两个算法（最低ID和WCA）的启发的新解决方案。该解决方案称为最低重量，利用它们的优点，并在集群稳定性和计算开销方面释放其缺点。模拟实验，以评估在簇号码，簇头的寿命和reaffiliations的数目（从群集节点移动到另一个）方面提出的算法的性能。结果表明，最大的重量改善了现有算法，特别是关于移动性，导致更合适，适应，可扩展和自主聚类。
展开▼
