An overlapped execution technique for graph rewriting rules

Graph transformation is a well-known technique to process arbitrary object structures represented as attributed graphs. Recent graph transformation engines usually focus on the optimization of the individual rules, however, the similarity of the rules could be harnessed in many practical cases. A typical transformation engine performs the matching phase of the rule, then executes the rewriting phase, and starts the matching phase of another rule anew. However, in the matching phase, the (partial) results of previous pattern matches are not harnessed. In order to optimize the matching phase, an overlapped matching technique is presented in this paper: based on the definition of the rewriting rules, a compiled matcher is generated, that overlaps the matching phase of several rules at their common parts. The overlapped matching idea is generalized to incorporate structural hierarchy: overlapped rules may use parts of the common matches to a different degree.