Concurrent Model Synchronization with Conflict Resolution Based on Triple Graph Grammars - Extended Version

Triple graph grammars (TGGs) have been used successfully to analyse correctness of bidirectional model transformations. Most recently, also a corresponding formal approach to model synchronization has been presented, where a forward propagation operation updates a source model modification from source to target, and symmetrically, a backward propagation operation takes care of updates from target to source models. However, a corresponding formal approach of concurrent model synchronization, where a source and a target modification have to be synchronized simultaneously, has not yet been presented and analysed. This paper closes this gap taking into account that the given and propagated source or target model modifications are in conflict with each other. Our conflict resolution strategy is semi-automatic, where a formal resolution strategy – known from previous work – can be combined with a user-specific strategy. As first main result, we show correctness of concurrent model synchronization with respect to the TGG. This means that each result of our nondeterministic concurrent update leads to a consistent correspondence between source and target models, where consistency is defined by the TGG. As second main result, we show compatibility of concurrent with basic model synchronization. In other words, concurrent model synchronization can be realized either to coincide with forward or with backward propagation. The main results are illustrated by a running example on updating organizational models.