A Framework for Managing Process Variability Through Process Mining and Semantic Reasoning: An Application in Healthcare

Abstract : The efficiency of organizations relies on its ability to adapt their business processes according to changes that may occur in the dynamic environment in which they operate. These adaptations result in new versions of the process model, known as process variants. Thus, several process variants can exist, which aim to represent all the related contexts that may differ in activities, resources, control flow, and data. Thus, has emerged the concept of customizable process model. It aims to adapt the process model according to changes in the business context. A process model can be customized by representing the process family in one single model enabling to derive a process variant through transformations in this single model. As benefits, this approach enables to avoid redundancies, promotes the model reuse and comparison, among others. However, the process variant customization is not a trivial-task. It must be ensured that the variant is correct in a structural and behavioural way (e.g. avoiding disconnected activities or deadlocks), and respecting all the requirements of the application context. Besides, the resulting process variant must respect all requirements related to the application context, internal and external regulations, among others. In addition, recommendations and guidance should be provided during the process customization. Guidance help the user to customize correct process variants, i.e., without behavioural problems. Recommendations about the process context help the user in customizing process variants according specific requirements. Recommendations about the business context refers to providing information about the best practices that can improve the quality of the process. In this context, this research aims to propose a framework for customizing process variants according to the user’s requirements. The customization is achieved by reasoning on ontologies based on the rules for selecting a process variant and in the internal/external regulations and expert knowledge. The framework is composed by three steps. The first step proposes to identify the process variants from an event log through process mining techniques, which enable to discover the variation points, i.e., the parts of the model that are subject to variation, the alternatives for the variation points and the rules to select the alternatives. By identifying the process variants and their characteristics from an event log, the process model can be correctly individualized by meeting the requirements of the context of application. Based on these aspects, the second step can be developed. This step refers to the development of the questionnaire-model approach. In the questionnaire approach each variation point is related to a question, and the alternatives for each question corresponds to the selection of the process variants. The third step corresponds to apply two ontologies for process model customization. One ontology formalizes the knowledge related with the internal and/or external regulations and expert knowledge. The other refers to the variation points, the alternatives for them and the rules for choosing each path. The ontologies then are merged into one new ontology, which contain the necessary knowledge for customize the process variants. Thus, by answering the questionnaire and by reasoning on the ontology, the alternatives related with the business process and the recommendations about the business context are provided for the user. The framework is evaluated through a case study related to the treatment of patients diagnosed with acute ischemic stroke. As result, the proposed framework provides a support decision-making during the process model customization.
Document type :
Theses
Complete list of metadatas

Cited literature [111 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/tel-01725605
Contributor : Hervé Panetto <>
Submitted on : Friday, March 9, 2018 - 2:27:51 PM
Last modification on : Friday, May 17, 2019 - 4:27:24 PM
Long-term archiving on : Sunday, June 10, 2018 - 2:26:44 PM

File

Detro_SP_A Framework for Manag...
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-01725605, version 2

Collections

Citation

Silvana Pereira Detro. A Framework for Managing Process Variability Through Process Mining and Semantic Reasoning: An Application in Healthcare. Computer Aided Engineering. Université de Lorraine; Pontifical Catholic University of Parana (PUC-PR), 2017. English. ⟨NNT : 2017LORR0310⟩. ⟨tel-01725605v2⟩

Share

Metrics

Record views

95

Files downloads

264