More and more events are being recorded by computer applications. These events correspond to a given process with some associated workflow. Companies usually have some (more or less accurate) Business Process Models, but these can also be inferred from the event logs using process mining techniques. Either one way or another, a process must always be analysed to find its flaws and be able to improve it.

Firstly, we could analyse the conformance of the data in the log (corresponding to real executions) to the process model at hand (corresponding to the thoretical workflow). Moreover, the quality of the output of the process will clearly depend on the steps (and workflow branches) taken to obtain that output. Thus, the goal of this topic will be to find the relevant criteria to analyse the quality of a given process based on the associated process model and log.

In a service-oriented context, analysing quality is of obvious importance both to users (since it guides their decisions) but also to providers (because it affects their reputation and costs). This topic aims at producing the needed infrastructure (in the form of models, techniques, and tool support) for discovering which are the key measures and performance indicators to decide whether services satisfy a particular request from an end-user (either a human or another software service), as well as the dimensions that affect the values of that measures. Process mining (whose complexity suggests the use of the power of a cloud or private cluster of parallel machies) and statistical techniques should be used to extract patterns from the existing data generated by a service, as well as the feedback from the user (stating the perception of quality in the service). Then, we should be able to correlate both and discover analysis dimensions for each measure.

Main Advisor at Universitat Politècnica de Catalunya (UPC)
Co-advisor at Technische Universität Dresden (TUD)