Non-Local Correction of Process Models Using Event Logs
Information and software systems almost never stabilize after the release. They change, among other reasons, due to difficulties of implementation, bug repairs, and environmental shifts. Thus, their behavior differs from models, which were used at the design stage. Process owner usually wants relevant, up-to-date models. Fortunately, most modern information systems write detailed event logs of their functioning. In this paper, we describe an algorithm of non-local process correction, that employs the two fundamental algorithmic paradigms: “divide and conquer” and “greedy processing”. It decomposes the process model and repairs sub-models in a greedy way using event logs with actual behavior. Using this procedure, it is possible to correct both local and non-local inconsistencies. The paper considers the algorithm and the results of its experimental evaluation using several artificial examples.