Run-Time Composition of Partly Structured Business Processes Using Heuristic Planning

Context: Organizations act in a highly competitive markets, which forces them to be able to quickly adapt and optimize their processes. Case Management Model and Notation (CMMN) supports modelling run-time adaptation of partially structured business process models, but does neither allow selection nor sequence of tasks. Objective: Our goal is to develop an approach that computes sequences of tasks such that a context-specific goal is satisfied. We apply the SMA algorithm to support run-time planning of partly structured parts of a business process in order to find an appropriate sequence of tasks. Method: We perform a literature review in order to derive the most suitable heuristic search algorithm. To support the selection and specification of tasks, we extend the CMMN metamodel. A running example is presented to show practical applicability of the approach. Results: The extension of the CMMN metamodel reflects the requirements for serializing tasks at run-time. Furthermore, we modified the SMA algorithm and defined a heuristic function. Conclusion: Our approach allows to automatically derive execution plans for partly structured business processes depending on the context.