Development and validation of customized process models

Abstract

Configurable reference process models encompass common and variable processes of organizations from different business domains. These reference process models are designed and reused to guide and derive customized business processes according to the requirements of stakeholders. The customization process is generally initiated by a configuration step, selecting a subset of the reference process model. Configuration is followed by a customization step, which assumes adapting or extending the configured business process based on the specific or unforeseen requirements. Hence, it is crucial to validate the correctness and compliance of the final customized business process with respect to the patterns and business constraints that are specified in the reference model. In this paper, we firstly introduce a technique to develop a customized process model and then present a set of identified inconsistency patterns that may happen during the configuration of a reference model and the customization of configured process models. Furthermore, we describe our proposed approach including formal representations and algorithms that provide logical reasoning and enable automatic inconsistency detection by leveraging description logic. In order to explore the scalability of the approach, we designed the experiments with various process models sizes and inconsistency distributions. The results of the experiments revealed the scalability of our approach with large size process models -500 activities-.