Fast Recursive Data Processing in Graphs using Reduction

This paper presents an algorithm for recursive data processing in directed graphs. The proposed algorithm applies graph reduction in order to determine both starting points and a correct ordering of recursive operations, provided the directed graph is a-cyclic. Therefore it is essential that the algorithm is also able to detect cycles efficiently. The algorithm arose from the implementation of recursive, semantic query specifications and is implemented in a DBMS prototype. Experiments confirmed that the theoretically estimated time complexity is O(dN), where N is the number of arcs and d is the depth of the graph (d ≤ N). The worst-case performance is O(N2 ), also for cycle detection.

• Type: Conference
• Year: 2003
• Where: APPINF
• Authors: J. H. ter Bekke, J. A. Bakker
• Keywords: Algorithm, APPINF 2003, Directed Graph, Information Technology, Recursive Data Processing