The Modeling Research And Implement Based On Workflow Knowledge

Knowledge-rich workflow systems have shown powerful capabilities and promising extensive application in many fields such as Business Process Reengineering (BPR) and Process-Oriented Application. Compared with actualization and development pace of knowledge-based workflow, theoretical research on knowledge-based workflow modeling has lagged behind. No systematic theory has been developed on modeling methodology. Traditionally, the modeling techniques of workflow are mainly based on complete pre-definition of flows, which do not apply to the dynamic knowledge-rich workflow systems. Consequently, the Workflow Management Systems (WfMS) based on this approach lack flexibility. Hence there is an urgent need to establish a mechanism to generate a flexible, domain-independent knowledge-rich dynamic workflow.In view of the above problems, a knowledge-based workflow modeling method is proposed in this thesis, and a feasible extended workflow model consisting of organizations-role model, workflow process model, workflow resource model and workflow contextual analysis is flexibly built up on the basis of the existing theories and researches, taking into account enterprise application needs.In this paper, based on knowledge workflow technology refers to knowledge-rich workflow modeling technology.The originality and innovation of the work presented in this thesis are as following:1) It is pointed out after analyzing workflow knowledge-based modeling techniques that the critical work in modeling is"how to represent knowledge appropriately and efficiently", i.e., knowledge representing, The Workflow knowledge is classified into two categories: System Layer knowledge and Application Layer knowledge. The System layer knowledge includes the system resource entities and their relationship, and the formulation of resource performance and operating strategy etc; while the application layer knowledge contains business process experience, user preferences, operating strategy restrictions as well as other business-related knowledge.2) A framework is adopted in modeling workflow system-layer knowledge. The representation using a framework conforms to goal-driven disposal. In given scenarios, it enables the determining of the self-applicability or providing other frameworks. The knowledge organization of framework helps workflow illation, framework also being one of the ontology formalization approaches and the ideal modeling method based on system layer knowledge. The framework construction and the knowledge representation of the workflow modeling are provided in this thesis, the three types of framework being defined and thus modeling based on system layer knowledge being achieved.3) Workflow transition pattern, a cutting-edge technology, is applied in modeling application layer knowledge. In this thesis, process model is defined as a representation structure of business knowledge, denoting the best practice of completing particular goals and process definition in specific context. In this pattern, problems are defined as the business goals to be achieved; scenarios as the context of problems,while the solutions as the workflowlets of the business process which are capable of achieving the business goals. Then the model based on application layer workflow knowledge is presented, which uses goal-driven pattern instead of traditional direct calling pattern, and thereby improves the flexibility and adaptability of the workflow execution.4) The programming techniques from Artificial Intelligence is introduced and applied to the dynamic construction of workflow. Since the transition pattern is used on the knowledge representation, accordingly the pattern-oriented programming techniques are selected and the programming approach of artificial intelligence is further extended. The pattern-oriented programming approach with the extension of programming approach of artificial intelligence is the core of workflow construction.5) The modeling of system layer knowledge and application layer knowledge are integrated. As both the system layer knowledge and the application layer knowledge are represented by using ontology technology, the problems on representation, searching and reuse of design knowledge, can be appropriately resolved. A knowledge-based workflow model is ultimately constructed in the thesis, which has been applied to the project of gas station management system and turns out to be feasible.