Ad hoc分层结构中的神经元聚类分簇算法

A Clustering Algorithm of Ad Hoc Hierarchical Architecture Based on Nerve Cell and Clustering-Analyzing

目前移动Ad hoc的分簇算法大都自适应能力不强,不能够满足条件受限的野外环境。该文通过利用神经元和移动结点的相似性,结合神经网络的竞争学习和模糊数学中的聚类分析,设计了一种基于神经网络和模糊数学的新型分簇算法。这是一种高效的、自适应的路由算法,具有一定的创新性,对于探讨Ad hoc网络有着重要的意义。该文首先介绍了Ad hoc网络概念和特点,然后分析了几种现存的分簇算法。接着深入研究了这种基于自组织神经网络中的竞争学习和聚类分析的混合算法。最后指出了需要改进的方面。

Most existing clustering algorithms only consider some single factors, whose adaptability is not good, cant meet the requirement of field environment. So, by using the similarity between nerve cell and networks node, and combining the competitive learning in self- organizing neural networks with clustering analyzing in misty mathematics, a novel clustering algorithm based on neural networks and misty mathematics is designed. The paper firstly introduces some distinct characteristics and conceptions of Ad hoc network. Then several existing clustering algorithms are explained. Afterwards, a clustering commix algorithm with good performance based on competitive learning in self - organizing neural networks and clustering analyzing is analyzed in detail.