부산대학교 도서관

Process modeling and discovery techniques aim to construct sound and valid process models for different types of processes, i.e., process orchestrations and collaboration processes. Orchestrations represent behavior of cases within one process. Collaboration processes represent behavior of collaborating cases within multiple process orchestrations that interact via collaboration concepts such as organizations, agents, objects, and services. The heterogeneity of collaboration concepts and types such as message exchange and resource sharing has led to different representations and discovery techniques for collaboration process models, but a standard model class is lacking. We propose collaboration Petri nets (cPN) to achieve comparability between techniques, to enable approach and property transfer, and to build a standardized collaboration mining pipeline similar to process mining. For cPN, we require desirable modeling power, decision power, modeling convenience, and relations to existing model classes. We show the representation of collaboration types, structural characterization as workflow nets, automatic verification of soundness, bisimulation equivalence to existing model classes, and application in a general discovery framework. As empirical evidence to discover cPN, we conduct a comparative evaluation between three discovery techniques on a set of existing collaboration event logs.