ABSTRACT

This article considers scheduling in tree data gathering networks. The worker nodes of a network acquire datasets of known sizes at possibly different times. The datasets are sent to intermediate nodes which process them and then pass the results to the base station. Each dataset is assigned a due date by which it should arrive at the base station. The scheduling problem is to minimize the maximum dataset lateness. It is proved that this problem is strongly NP-hard, and heuristic algorithms based on priority rules are proposed. The algorithms are tested in a series of computational experiments, for a wide range of changing network parameters. The results show that, although most algorithms perform well only for certain instances, one of the proposed priority rules achieves very good results for all settings analysed.

摘要

本文考虑了树数据收集网络中的调度。网络的工作节点在可能不同的时间获取已知大小的数据集。数据集被发送到中间节点进行处理，然后将结果传递给基站。每个数据集被分配一个到期日期，该日期应到达基站。调度问题是使最大数据集延迟最小化。实践证明，该问题具有较强的NP难性，提出了基于优先级规则的启发式算法。在一系列不断变化的网络参数中，对算法进行了一系列的计算实验。结果表明，尽管大多数算法仅在某些情况下性能良好，但是所提出的优先级规则之一对于所分析的所有设置都取得了很好的结果。