Testing equivalences for event structures

A flexible abstraction mechanism for models of concurrency, which allows systems which "look the same" to be considered equivalent, is proposed. Using three classes of atomic observations (sequences of actions, sequences of multisets of actions and partial orderings of actions) different information on the causal and temporal structure of Event Structures, a basic model of parallelism, is captured. As a result, three different semantic models for concurrent systems are obtained. These models can be used as the basis for defining interleaving, multisets or partial ordering semantics of concurrent systems. The common framework used to define the models allows us to study the relationship between these three traditional approaches to the semantics of concurrent communicating systems.