基于并行化蚁群算法的网络测量节点选取算法

Network Measurement Node Selection Algorithm Based on Parallel ACO Algorithm

提出了一种基于并行化蚁群算法,用于网络测量中测量节点自动选取的算法(MNS)。首先对传统蚁群算法进行改进,使其适用于网络测量中测量节点的选取场景。然后分析蚁群算法的并行化方案,设计并实现并行化框架。最后通过多元函数求解极值分析和在模拟网络中运行选点任务两种方法,对并行化选点算法(P-MNS)和非并行化选点算法进行对比。通过实验验证,并行化的蚁群算法不仅能满足网络测量节点选取的要求,同时相比非并行化算法具有更快的收敛速度,更适用于大规模网络测量。

This paper presents an algorithm based on parallelized ant colony algorithm（ ACO） for selection of measurement nodes in network measurement（ MNS）. First of all,the traditional ant colony algorithm is adapted to the scenario of selecting nodes for the selection of measurement nodes in network measurement. Then,the parallelization scheme of ant colony algorithm is analyzed and the parallel framework is designed and implemented. Finally,compare the parallel measurement node selection algorithm（ P-MNS） and the non-parallelized ant colony algorithm by analyzing the function optimization problem in continuous domain and simulating the task of node selection in the network. As the experiments shows,the parallelized ant colony algorithm can not only meet the requirements of the network measurement node selection,but also has faster convergence speed than the non-parallelization algorithm and is more suitable for large-scale network measurement.