Automated Reasoning Support for Process Models using Action Languages

Master's thesis by Itzel Vázquez Sandoval

Supervisor Steffen Hölldobler
Wissensverarbeitung
28 September 2014 – 28 September 2014
Download

By cause of the human-interactive nature of business processes a diversity of problems emerge from their automation, extensively effected nowadays. Such automation allows to monitor process and service executions, and consequently to analyze and reason about the observed data. These activities are weakened nevertheless by the detection of incomplete information concerning the process-level tasks.

The purpose of the present thesis is to solve the particular problem of reconstructing incomplete observed executions of processes, in order to support the performance of reliable business analysis over monitored behavior. Such analysis permits among others, the enhancement of the process models. The proposed solution consists in the characterization of the model of the process, along with the recorded trace, as a logical formulation in terms of a planning problem; by means of automated reasoning then, plans conforming to possible sequences of activities that correspond to complete instances of process executions can be generated.