Automatic Synthesis of Linear Process

We present PM-MetaFrame, a tool for automatic synthesis (i.e. prior to any enaction) of linear process models from speciications written as global constraints in linear time temporal logic. Key to the automatic synthesis is its level of abstraction: process model components are described logically by a taxonomic classiication, expressing semantic properties, and the ow of data is only considered on the level of types, which is suucient to guarantee executability of synthesized process models. Thus our approach can be regarded as a high-level extension of existing process model approaches, where, if at all, global constraints are described on the data level. The result of the synthesis is the set of all executable linear process models satisfying the, typically very loose, temporal constraints. Incremental reenement of global constraints by strengthening/loosening via conjunction/disjunction helps minimizing the search space.