International Business Machines Corporation

Jussi H Vanhatalo

Jochen M Kuester

Jana Koehler

Michael S Wahler

Hagen Voelzer

Ksenia Ryndina

Simon D Moser

Christian Gerth

Olaf W Zimmermann

Rainer F Hauser

Felix Feger

A system and associated method for hierarchically decomposing a workflow graph G into a process structure tree PST. The workflow graph G is a two-terminal graph parsed into a tree T having triconnected components. Boundary pairs of all triconnected components in T are computed and fragments are discovered with boundary pairs. T is restructured into PST pursuant to categories of triconnected components in T. PST is deterministic and modular. PST represents a block-based process model that has fine blocks of execution units. PST is computed in time linear to the number of edges in G.

System and method for hierarchically decomposing process model

Methods and apparatus are provided for production of a difference log in a data processing system. The difference log defines differences between process models defined in system memory. For each of the process models, model structure data provided in memory defines a hierarchy of SESE regions representing the structure of that model. Also provided in memory are model comparison data defining correspondences between elements of the models, and region comparison data defining correspondences between regions of the SESE region hierarchies for the models. The model comparison and region comparison data are analyzed to identify differences between the SESE region hierarchies, and a difference log defining said differences is produced. In preferred systems, the model structure data and the region comparison data are computed for the models, and the difference log has a hierarchical structure corresponding to the structure of the process models.

Difference log production for model merging

Techniques for composition of model transformations from a predetermined set of model transformations. A state machine is provided in memory. The states are defined in the state machine in terms of predetermined model attributes. In response to specification of a target state for an input model to be transformed, an execution sequence in the state machine, between a start state corresponding to the input model and an end state corresponding to the specified target state, is selected. Each transformation in the selected sequence is then successively executed on the input model. After executing each transformation in the selected sequence, the transformed input model state is compared to the model state defined in the state machine to determine if the selected sequence is inoperable for the input model. If so, an alternative execution sequence in the state machine, between the input model state and the specified target state, is selected.

Automatic composition of model transformations

A method for modeling business transformation includes steps of: constructing a first analysis model based on at least one legacy-business-process-model; and analyzing the first analysis model in conjunction with at least one business-process-reference-model. The method further includes deriving a second analysis model in response to analyzing the first analysis model in conjunction with the at least one business-process-reference-model and refining the second analysis model to obtain a design model.

Method for modeling business transformation

Growing community of inventors

Zurich, Switzerland

Jussi H Vanhatalo