Efficient reachability queries in multi-relation graph: An index-based approach

The reachability query is a fundamental problem in graph analysis. Most existing studies assume there is no or only one type of relation on the edges. To fill the gap, we propose two new reachability problems in multi-relation graphs. Specifically, given a multi-relation graph G, two vertices u,v, and a query relation set C, it aims to check whether u can reach v under the necessity/existence relation constraint. A baseline online reachability query framework firstly is proposed by extending the BFS method. To accelerate the processing, index-based methods are developed based on 2-hop cover framework. Moreover, several optimized strategies are developed to improve the efficiency of index construction. Extensive experiments are conducted to demonstrate the advantages of developed techniques. As shown, the optimized method can achieve up to 3X acceleration for index construction and the index-based strategy can achieve up to 36X speedup compared with the baseline for query processing. [Display omitted] •We propose two new reachability queries on multi-relation graphs.•Online searching method and index-based methods are developed for processing large graphs.•Priority strategies are introduced to accelerate the construction of index.•Comprehensive experiments are conducted on 6 networks to demonstrate the efficiency and effectiveness of our proposed techniques.