基于蚁群算法的树环结构应用层组播树构建算法

Building algorithm for tree-ring application layer multicast based on ant colony algorithm

作为对IP组播技术的改进,应用层组播(ALM)技术具有不受网络架构限制、资源丰富、数据传输速率高等优势。综合考虑节点性能和端到端延迟,提出一种快速且高效的建立应用层组播树的方法:基于蚁群算法的树环结构ALM模型(ACOTRM)。目前对ALM结构的研究大部分都是仅仅粗略地给出拓扑结构,而对整个构建过程却缺乏完整且清晰的描述。鉴于此,完整地给出了一种树环结构ALM树的具体构建过程,主要包括网络节点分簇、簇内成环连接、数据转发数的建立等关键步骤。另外,为每个节点设定特有的优先级,使得整个ALM树达到最优化状态。仿真结果表明,该模型得到了较低的平均时延以及较高的平均数据传输率,在增加系统稳定性的同时也提高了其转发效率。

Abstract： As an improvement of IP multicast technology, Application Layer Multicast （ALM） has many advantages such as unlimited network architecture, rich resource and high data transfer rate. Considering node performance and end to end delay, a fast and efficient method was proposed to establish application layer multicast tree, it was named Ant Colony Algorithm based Tree-ring Application Layer Muhicast Model （ACOTRM）. The available studies only gave a topology cursory and had no complete and clear description of the concrete construction process. In view of this, a complete ALM hierarchical tree-ring concrete construction process was put forward including several key steps, such as clustering division, connection in cluster ring, generation of feasible solution and maintenance of the model in survival time. In addition, in order to optimize the ALM state tree, each node was set with a specific priority. The simulation results show that the proposed model provides lower average delay and higher average data delivery ratio, which increases the system stability and forwarding efficiency at the same time.