Relevant Technologies For Process-Aware Information Systems Based On Petri-Net Process Modeling

Nowadays more and more business processes are controlled and managed by driven processmodel information systems, and numerous process engineering, process modeling and processapproach to achieve appeared accordingly. Many technologies in this field vividly illustrate theavailability and complexity for the theme of business process. Therefore, relevant supportingtechnologies and standards of the Process Aware Information System (PAIS) become aninternational research focus in recent years. Process mining is a key technology in PAIS, whichcan extract relevant information and reconstruct the process model from the running logs of allthe process instances. The instances of all running cases can meet the reconstructed processmodel. The research of process mining is of great significance for analyzing and improving theexisting business process as well as deploying new ones.Process mining mainly includes control flow, organization dimensions and case dimensions.This thesis focuses on the mining of the control flow which breaks the perspective of the twoopen mining problems of loops and duplicate activities. The main contribution of the thesis isthat an improved mining algorithm calledγ-algorithm is proposed based on theα-algorithm,which can find the loops and the duplicate activities from event logs of the business process.Firstly, the associations among activities are redefined, and the heuristic decision rules are putforward through analyzing the properties of duplicate activities. Then, all the loops and duplicateactivities in the event logs are identified and restored by the rules and extract the correspondingworkflow net usingαalgorithm to acquire the result of workflow model.Finally, the correctness and effectiveness ofγ-algorithm is verified by examples of the opensource process mining framework ProM5.