Path queries on massive graphs based on multi-granular graph partitioning

Abstract

The quotient space theory can represent the world at different granularities and deal with complicated problems hierarchically. In this paper, we introduce a method for finding shortest path on massive graphs based on multi-granular graph partitioning. Path queries into two parts: preprocessing and query step. In preprocessing, we introduce a heuristic local community structure discovery algorithm to decompose massive graphs into some local communities and construct a sequence of hierarchical quotient spaces to describe hierarchy structure on massive graphs. In query, we improve evaluation function of heuristic search method in path query. The implementation works on massive networks. From experimental results, it can be stated that proposed algorithm is effective and efficient in transportation road networks of US. © 2011 Springer-Verlag.