Abstract
To process shortest path and distance queries on road networks, various preprocessing techniques have been proposed. State-of-the-art methods for distance queries offer superior query time but do not provide any efficient retrieval mechanism for the shortest path. In contrast, state-of-the-art methods for shortest path queries show relatively poor performance for distance queries. In this paper, we propose a Partition-based framework for Distance and Shortest Path queries (ParDiSP), which efficiently supports both types of queries. After partitioning a road network into components, ParDiSP precomputes the distances between any node in a component and the border nodes of the same component, the pairwise distances between all border nodes, and the union of the shortest paths between all border nodes. For query processing, ParDiSP runs the ALT algorithm for both distance and shortest path queries if source and target are located in the same component. Otherwise, ParDiSP answers distance queries by combining distances from exactly three precomputed distance tables. For shortest path queries, the information in the distance tables allows to identify two border nodes that are traversed by the shortest path, thereby decomposing the path into three segments which can be computed in parallel. A detailed experimental evaluation shows that ParDiSP outperforms two state-of-the-art solutions for shortest path queries and is comparable to the state-of-the-art for distance queries. For mixed query loads containing both distance and shortest path queries, ParDiSP outperforms a combination of the best methods for each query type, while its space requirements are significantly smaller.