Incremental Pattern Matching in the VIATRA Model Transformation System

摘要

Incremental pattern matching is a key challenge for many tool integration, model synchronization and (discrete-event) model simulation tasks. An incremental pattern matching engine explicitly stores existing matches, while these matches are maintained incrementally with respect to the changes of the underlying model. In the current paper, we present an adaptation of RETE networks [6] in order to provide incremental support for the transformation language of the Viatra2 framework. We evaluate the performance of the incremental engine on a benchmark problem assessing the speedup of incremental processing in the case of as-long-as-possible type of rule applications.