Automatic generation of process models
AssigneeInternational Business Machines Corporation
InventorsMati Golani, Shlomit Pinter
Life of this patent
Estimating the term…
Abstract
A method for process modeling includes reading a record of executions of a process including at least first and second activities, and identifying in one of the executions in the record respective first and second lifespans of the first and second activities, defined by respective initiating and finish events, such that the initiating event of the second lifespan occurs during the first lifespan. A graphic model of the process is generated, so as to reflect a concurrency of the first and second activities, responsively to occurrence of the initiating event of the second lifespan during the first lifespan.
Claims10
A method for process modeling, comprising:
- using a log processor, reading from a computer memory respective records of a plurality of executions of a process comprising activities;
- generating in the log processor a plurality of execution graphs corresponding respectively to the executions of the process and having nodes corresponding to the activities;
- and
- merging the execution graphs using the log processor to create a process model graph that comprises the nodes and models the process, including start and end activities of the process,
- wherein the execution graphs represent legal flows over the process model, which comprise both the start and end activities, while different ones of the execution graphs comprise different, respective subsets of the nodes in the process model graph.
A method for process modeling, comprising:
- using a log processor, reading from a computer memory respective records of a plurality of executions of a process comprising activities;
- generating in the log processor a plurality of execution graphs corresponding respectively to the executions of the process;
- and
- merging the execution graphs using the log processor to create a process model graph that models the process and comprises a plurality of nodes that correspond to the activities in the process and are connected by directed edges;
- identifying, using the log processor, a strongly-connected component comprising three or more of the nodes in the process model graph;
- partitioning the nodes in the strongly-connected component, using the log processor, into multiple sets, depending on the directed edges connecting the nodes in the strongly-connected component to the nodes outside the strongly-connected component;
- and
- based on the partitioning, removing from the process model graph, using the log processor, at least one of the directed edges between the nodes in the strongly-connected component.
Dependencies are inferred from the claim text.