A Knowledge-Enriched Computational Model to Support Lifecycle Activities of Computational Models in Smart Manufacturing

Abstract

Due to the needs in supporting lifecycle activities of computational models in Smart Manufacturing (SM), a Knowledge Enriched Computational Model (KECM) is proposed in this dissertation to capture and integrate domain knowledge with standardized computational models. The KECM captures domain knowledge into information model(s), physics-based model(s), and rationales. To support model development in a distributed environment, the KECM can be used as the medium for formal information sharing between model developers. A case study has been developed to demonstrate the utilization of the KECM in supporting the construction of a Bayesian Network model. To support the deployment of computational models in SM systems, the KECM can be used for data integration between computational models and SM systems. A case study has been developed to show the deployment of a Constraint Programming optimization model into a Business To Manufacturing Markup Language (B2MML) -based system. In another situation where multiple computational models need to be deployed, the KECM can be used to support the combination of computational models. A case study has been developed to show the combination of an Agent-based model and a Decision Tree model using the KECM. To support model retrieval, a semantics-based method is suggested in this dissertation. As an example, a dispatching rule model retrieval problem has been addressed with a semantics-based approach. The semantics-based approach has been verified and it demonstrates good capability in using the KECM to retrieve computational models. A KNOWLEDGE ENRICHED COMPUTATIONAL MODEL TO SUPPORT LIFECYCLE ACTIVITIES OF COMPUTATIONAL MODELS IN SMART MANUFACTURING