Research Of Trust-Based Recommendation Service Method In Migrating Workflow

The Migrating Workflow is a fresh techonology which applies mobile agent to workflow management. Migrating workflow consists of one or more Migrating Instances migrate between different Work Places to find suitable work location services, in which migrate instance is the main body of the workflow activities, and work place is to provide services for migrating instance node. Migration workflow greatly improves environmental adaptability and flexibility of the workflow system, so migrating workflow model has drawn the extensive attention by researchers and industry.Along with the rapid development and expansion of the network, the internet provides people with a vivid changeable and mutual collected informational environment. However, due to the complication of the information environment, it becomes a difficulty to predict the internet environment precisely. Migrating workflow model has overcome the disadvantages of the traditional workflow model and distributed environment, in which migrating instances are the executives and also interpreted as the workflow instruction of combining static explanation, the local rules and the effect of the user’s decision on work places; while the work place is the abstract manifestation of resources and participants. The migrating instances are able to finish one or more tasks by using resources in the current work place. If the current work place cannot meet the requirements of the task, the migrating instances will automatically ask for migration to the next more suitable workplace with the task and try to finish the task in the new work place. It has become an important issue for the migrating instances to figure out the right recommended strategy, in order to meet the requirements of the current task, which greatly influences the executive efficiency of workflow system and its adaptability to the environment. This thesis is to put forward a workplace networking model upon trust, and to propose the recommendation service in workflow.This thesis firstly proposes the workplace networking model based on trust in the work places. In interpersonal networks, the trust relationship is the core of human relationships, and trust is the evaluation of the credibility behavior of a particular individual. The individual’s trust often depends on the individual embodied in cooperation with other individuals. Work location networks and interpersonal networks have a lot of similarities. The recommended service between the work places in the network will leave the information to reflect the behavioral characteristics of the work place and any obligation to the other location in the network to provide recommended information. Meanwhile, any work place for collaboration objects has the full right to choose.According to the network model based on trust in the work places, recommendation service in work places is used by this thesis and referral service based on trust mechanism:when the migrating instance in the current work place after completing the current task, migrating instance target has not yet completed, current work place will adopt the recommended service in the workflow to complete the next task. To achieve better optimal selection, recommended in migrating workflow service requires a plurality of work place to cooperate for the completion of the task, and the recommended work place on the feasible region, and finally recommends the most optimal work place from the feasible domain migrating instances. Degree of trust between the operating positions,time, and other factors needed to be taken into consideration in the development of different selection recommended by the strategy. Here the time issue is composed of recommendation time and service time. According to the different environmental requirements, this thesis puts forward two different recommendation service methods to optimize both of the time factors.In this thesis, the trust-based migration workflow recommendation mechanism is the key issue in the Migrating Workflow. By simulation experiments, experimental results show that the recommended service method recommendations effectively enhance the efficiency of the service, and can preferably prevent malicious recommendation.