一种新的基于递归分解的图可达性查询算法

New way based on RDD to evaluate graph reachability queries

针对现实中许多超大规模图可达性查询的问题,提出了一种新的基于递归分解的算法,即将原图递归分解成一系列生成树和剩余图两类子图,并通过分别查询这两类子图来减少查询开销。相比于区间标记、链分解、2-hop标签和路径树等传统算法,该算法不仅空间开销更小,且时间复杂度更低。仿真实验表明,该算法对处理大规模有向图可达性问题上存储规模更小且查询效率更高。

For many large scale graph reachability query problems in reality,this paper proposed a new algorithm based on RDD which decomposed a graph into a series of spanning trees and a residue graph. In this way,it transformed any reachability query into these two queries to reduce the cost. Compared with the traditional algorithms,like interval labeling,chain decomposition,2-hop labeling,and path-trees etc,this algorithm could lead to smaller space overhead,and lesser time complexity. The simulation illustrates that the algorithm has a smaller storage and more effectively to deal with the large scale groph reachability query problem.