Research On Ontology-based Key Technologies Of Task Management In Manufacturing Grid

Under the tendency of economic globalization and global information, manufacturing grid (MG) has become the research focus in networked manufacturing field. The final goal of MG is to realize manufacturing resource sharing and collaborative working, provide the better quality of service (QoS), so as to accomplish tasks submitted by consumers (service requestors). The task management (TM) is the fundamental and critical research in MG. During the task implementation process, TM system must finish the service discovery and matching according to service requirements and the service status of MG resources. Meanwhile, the creating and description, decomposition and planning, scheduling and execution of the task must also be solved effectively. TM is an important bridge and tie related with MG and applications, if its function is imperfect, it will restrict the applications and development. Thus, TM system concerns service efficiency, success or failure, determines the degree of consumers'participation and acceptance.The dissertation analyses and studies deeply the issues of TM in MG, centers on several critical technologies of TM from the theoretical research and applications point of view. The research and implementation of TM key technologies solved the bottleneck problem of affecting MG applications.In order to study and implement ontology-based MG task and service descriptions, MG task decomposition based on Modularization, MG task planning based on activity period and service constraint, service discovery and matching based on ontology and service capability, the dissertation analyses the grid and semantic web technology. Combining with the properties of networked manufacturing, the dissertation constructs a MG system framework supporting ontology service and the TM system structure, analyses properties of the local service layer, core service layer, application service layer and domain service layer. TM is an application of MG system. The dissertation analyses the functions of the main modules in TM: creating and modelling, decomposition and planning, service discovery and matching, executing and monitoring. It established the theoretical base for the research of TM key technologies.In MG system, the description of service requestors for service requirements is different from the service description of service providers for service because of different enterprises. In order to find the advanced theory and method for guiding the task and service description, to realize the service discovery and matching rapidly and effectively, the method that is ontology-based MG task and service modeling is put forward. Combining with ontology and networked manufacturing, by analyzing the basic ontology classes, an OWL-S based web service ontology model, i.e. MG ontology is constructed. Based on the MG ontology, the MG task and service description model are set up, and MG task and service semantic description are implemented. It is the basis for researching MG service discovery and matching based on ontology.A MG task needs to be decomposed to a series of subtasks and activities, so as to be implemented collaboratively by several service nodes. According to the task characteristics in manufacturing, the dissertation proposes task decomposition strategies based on modularization and sets up the decomposition model. According to the requirements of"weak coupling and strong cohesion", a judgment approach for correlation degree between activities and a measurement way for the task granularity are presented. The dissertation also establishes QoS-based task decomposition rules, proposes fundamental steps and method of MG task decomposition.For effective management and control of MG tasks, MG task managers or experts need to plan the activities after the task is decomposed, to determine critical activities and time schedule according to the constraint relation among activities, so as to achieve the objective of shortening service time and improving QoS. The dissertation puts forwards a method that sets up firstly a task planning based on activity periods then modifys the task planning based on service constraints, and the activity period-cost optimizing methods, achieves MG task planning goal that gets time from the critical path and gets resources (services) from non-critical path.In MG environment, the task service capabilitiy requirements matching with service capability is the key for service discovery and matching, and concerns the goal of MG resource sharing. Therefore, on the basis of analyzing and summarizing related research, the OWL-S based MG service publishing, discovery and matching process is presented. The task and service matching factors (IOPE), matching degree's definitions, and matching rules based on service capability are analyzed. The MG service discovery and matching Algorithm based on ontology and service constraints is proposed.Finally, the development and application validation for main research contents are carried out. The function of MG task submission, description, decomposition and planning are realized. A prototype system for MG task and service description, service discovery and matching based on semantic web service is designed and implemented preliminarily. The rationality and feasibility of TM system have been validated by a task which belongs to the manufacturing machining task type, and many valuable experiments and data are gotten. The solid foundation has been laid for further improve and apply MG and TM system.