Abstract. The current period of IT development is characterized by an explosive growth of diverse information representation languages. Applying integration and composition of heterogeneous information components it is required to develop the canonical information model serving for adequate expression of semantics of various information models used in the environment encompassing required heterogeneous components. Basic principles of the canonical model synthesis include fixing of its kernel, constructing the kernel extensions for each specific information model of the environment so that this extension together with the kernel could be refined by this information model, and forming the canonical model as a union of all such extensions. Previously these principles have been successfully applied to the synthesis of structural and object canonical models. This paper 1 applies this technique to synthesis of the process canonical model. The method proposed is based on interpretation of process model semantics in logics, and specifically, in the Abstract Machine Notation that made possible to construct provable refinements of process specifications. This method has been applied to the environment of process models defined by workflow patterns classified by W.M.P. van der Aalst. Thus the canonical process model synthesized possesses a property of completeness with respect to broad class of process models used in various Workflow Management Systems as well as the languages used for process composition of Web services.

No abstract