A Parallel Subgraph Isomorphism Algorithm on Multi-Core Platform

Subgraph isomorphism is an elemental issue in graph theory. Being the NP-hard problem overall, it is suitable for developing parallel algorithm to reduce the cost time. This paper presented an efficient isomorphism algorithm based on breadth first strategy and a scheme to decompose the matching task over multi-core platforms. The algorithm sorts the vertices of the two graphs by the the degree of outedge and inedge, then adds all the vertices to the feasible pair according to the connection relations of the current vertex. All the tasks distributed among the multi-cores share the same memory. The experiment shows that it has the better performance than current algorithm as the edges increase.