Simple Business Process Mining Based On Role Complexity

With the development of market development, enterprises and organizations should make adaptation to the changes of market requirements, adjust their business as well as sharpen their reaction capability. Therefore, the business process is in need of reengineering and optimizing. Activity based process mining focuses on the implementation of a process, and the optimizing of the process means to analyze activity value and adjust logical relationship between activities. Compared with that, role based process mining stresses the status of roles in a workflow and their interactive relationship, studies that have nowadays received increasingly high attention. However, before there is a problem with role based process mining that the roles identified are too complex, which makes it hard to implement and understand the process model. As the process actually becomes more and more complex, simple process has been gradually attached great importance to, leading to new requirements in role based process mining.In order to mine a role based simple process model with efficiency, cohesion and coupling degree was defined firstly, and the measure for cohesion and coupling degree based complexity which lay a foundation for the mining of simplified business process model was also designed. Then, based on the complexity of roles, the complex fitness function was proposed, so as to realize the role based simplified business process model. Furthermore, the thesis gave a more overall measure of the variability and sensitiveness of information in roles for the complexity of process. Through the measurement of information processing capability and reliance, the capability that a role uses information and the reliance of a role upon information were described. Then, an expanded measure index for process complexity was created, and the dependency relationships between information in the DSM-described process were established. Then process mining genetic algorithm based was replaced with based on particle swarm optimization algorithm, further improving the efficiency of process mining algorithm. Meanwhile, through experiment, the effectiveness and advantages of this algorithm was proved.