Rimer Based Workflow Network Model and Its Inference Algorithm

摘要

The transfer of activity information plays an important role in running process of workflow management system. It not only affects the self-adaption efficiency of workflow, but also can bring about resource consumption problem, which caused by the judgement of transfer of complex activity information. Aiming at this, based on belief rule-base inference methodology using the evidential reasoning approach (RIMER), it is proposed that a novel workflow network model to reduce complexity of the transfer process. After an activity node rule with belief degrees is defined, the activity node rule base is built by using activity history tables in workflow log data. Moreover, calculating the similarity between input activities and rules of activity node and judging the status of activity nodes are carried out to obtain activation rule set. The evidential reasoning (ER) approach is adopted for rule inference on the basis of calculating the activation degree of rule in activation rule set. Subsequently, comprehensive rules are constructed so as to achieve the transfer of activity information. And workflow network model is also successfully developed. Besides, this model can be further applied to improve the benefit and value of business process management.