The single-pair all-shortest-path problem is to find all possible shortest paths, given a single source-destination pair in a graph. Due to the lack of efficient algorithms for single-pair all-shortest-path problem, many applications used diverse types of modifications to the existing shortest-path algorithms such as Dijkstra’s algorithm. Such approaches can facilitate the analysis of medium-sized static networks, but the heavy computational cost impedes their use for massive and dynamic real-world networks. In this paper, we present a novel single-pair all-shortest-path algorithm, which performs well on massive networks as well as dynamic networks. The efficiency of our algorithm stems from novel 2-hop label-based query processing on large-size networks. For dynamic networks, we also demonstrate how to incrementally maintain all shortest paths in 2-hop labels, which allows our algorithm to handle the topological changes of dynamic networks such as insertion or deletion of edges. We carried out experiments on real-world large datasets, and the results confirms the effectiveness of our algorithms for the single-pair all-shortest-path computation and the incremental maintenance of 2-hop labels.

单对全最短路径问题是在给定图中单个源－目的地对的情况下，找到所有可能的最短路径。由于缺少用于单对全最短路径问题的有效算法，因此许多应用程序对现有的最短路径算法（例如Dijkstra算法）进行了多种修改。这样的方法可以促进对中型静态网络的分析，但是沉重的计算成本阻碍了它们在大规模动态的现实世界网络中的使用。在本文中，我们提出了一种新颖的单对全最短路径算法，该算法在大规模网络和动态网络上均能很好地运行。我们算法的效率源于大型网络上新颖的基于2跳标签的查询处理。对于动态网络，我们还演示了如何在2跳标签中增量维护所有最短路径，这使我们的算法能够处理动态网络的拓扑变化，例如边的插入或删除。我们在现实世界的大型数据集上进行了实验，结果证实了我们的算法对单对全最短路径计算和2跳标签增量维护的有效性。