A data flow perspective for business process integration

Abstract

Business process integration has become prevalent as business increasingly crosses organizational boundaries. To address the issue of protecting organizations’ competitive knowledge and private information while also enabling business-to-business (B2B) collaboration, past research has mainly focused on customized public and private process design and structural correctness of the integrated workflow. However, a data flow perspective is important for business process integration because data is always sensitive when inter-organizational business is conducted. Data flow is also a key aspect to enabling the correct and efficient execution of integrated business processes in today’s highly agile and competitive business environments in the global market. This dissertation presents a data flow perspective using workflow management and mathematical techniques to address data exchange problem in independent multi-stakeholder business process integration in dynamic circumstances. The research is conducted following a design science paradigm. We build artifacts including inter-organizational workflow concepts, a workflow model and a public data set calculation method. The objective is to provide a formal approach for data flow specification and analysis in inter-organizational workflows, which assist practitioners’ data exchange decisions to support process collaboration. The research first formally defines concepts in inter-organizational workflows and describes workflow integration based on these concepts, e.g., using adjacency matrix to store the control flow structure of integrated workflow. It then provides a model (including inter-organizational workflow concepts, lemmas and propositions) for data flow specification and analysis of inter-organizational workflow. Based on this model, a method is provided to calculate a public data set to be exchanged with partners for each involved party in both static and dynamic circumstances. In addition, it uses probability theory to make a prediction of integrated workflow instance to be executed. The use of the proposed artifacts is illustrated by applying them to a real world case in the Shenzhen (Chiwan) port. The utility of the artifacts is evaluated through semi-structured interviews with practitioners in industry. We conclude that this research complements the control flow perspective in the inter-organizational workflow management area and also contributes to B2B information sharing literature, and the data flow formalism can help practitioners to formally provide the right data at the right time in dynamic circumstances.