Week June 30th, 2008

Process Mining

Beyond of process models

By Ricardo Seguel P.

 

We are very pleased by having this week to Dr. Ana Karla Alves de Medeiros.  She has been one of precursors and leader around the world in research and development of Process Mining area.  Actually, she is doing a Post-doc at Eindhoven University of Technology  TU/e, as continuation of her successful Ph.D. research project which was done under supervision of famous Prof.Dr.ir. W.M.P. van der Aalst



Dr. Ana Karla A. de Medeiros
Eindhoven University of Technology
The Netherlands

 

 

 

Could you give us your point of view about relevance of Process monitoring and Process Mining for industry today?  How could we sell a Process Mining Project to a CEO?


Both Process Mining and Process Monitoring are relevant to industry because they provide for objective analysis of processes based on their actual executions. In this sense, process mining techniques focus more on the analysis of executed process instances while process monitoring ones target more real-time analysis of running instances.
For me, the selling point for the CEO is that the analysis is based on what is actually happening, i.e., the registered behavior in event logs. Furthermore, the process mining techniques provide various feedbacks involving the three main aspects of business processes: control-flow models, organizational models and data models.
Actually, the flyers at http://tabu.tm.tue.nl/wiki/promflyers contain more information about process mining and some of the case studies in this area.
 

What is Process Intelligence and what is the relation with Process Mining?


As described in the BPM-BPI 2008 workshop website (http://is.tm.tue.nl/bpi08/):

"Business Process Intelligence (BPI) refers to the application of various measurement and analysis techniques in the area of business process management. In practice, BPI is embodied in tools for managing process execution quality by offering several features such as analysis, prediction, monitoring, control, and optimization."

Process mining is a BPI tool! The process mining techniques are "intelligent" because the analysis they provide for is based on event logs and models. They do not require (much) human intervention to perform the analysis.
 

What are the challenges by considering a Process Mining project in a organization today?  Does an organization need a certain maturity level in their processes and systems architecture?


It is very simple to apply process mining techniques in an organization. The only requirements are that :

  1. the organization has event logs showing what has been executed for given processes,
  2. it is possible to identify which events (i.e., tasks or audit trail entries) belong to a same process instance (or case), and 
  3. there is a order of execution for the events in a same process instance. 
Actually, from all these points, the notion of process instances is the most important one. If this notion exists in the company, than the event log can be converted to the Mining XML (MXML) log format that is used by the ProM tool. To perform such conversion, one can use our free tool ProMimport. Both tools are open-source and can be downloaded at www.processmining.org.

 

What do you think about Process Mining and Process Optimization relation? Is there a relation with KPI's defined only or are there other aspects also?


The analysis provided by process mining techniques can be used to detect points of optimization for existing business processes. Some algorithms provide feedback based on KPI's, like the Performance Analysis Based on Petri Nets analysis ProM plug-in, but the great majority goes beyond KPI's. For instance, the Conformance Checker analysis ProM plug-in can be used to quantify how much a reference model matches the actual observed behavior. The paper " ProM 4.0: Comprehensive Support for Real Process Analysis" (at http://tabu.tm.tue.nl/wiki/publications/pn07-prom40) provides a good overview of the possibilities offered by current process mining techniques.

 

What is your point of view about Process Simulation and its relation with Process mining and Process Optimization? Is it really useful for industry today?


Simulation is a useful technique to assess the quality of designs of systems before their actual deployment. In the Business Process Management (BPM) area, simulations are often used to test (re-)designs of various aspects of process models (such as different placements of tasks in control flows, different resource allocations, predictions, what-if analysis, etc.). Recently, there has been a trend to try to automatically generate simulation models for business processes that are already in production (be it supported by a BPM system or otherwise). The claim is that simulation models that are built based on actual executions of business processes better reflect reality and, therefore, are more reliable when assessing the impacts of process re-designs or when making (short-term) predictions. In this sense, process mining techniques can be used to create such simulation models.

For the interested reader, the paper "Workflow Simulation for Operational Decision Support Using Design, Historic and State Information" (at http://tabu.tm.tue.nl/wiki/publications/workflowsimulation) is a good starting point.
 

Could you give us an invitation to know more about Process Mining?


Of course! If you want to know more about process mining, visit our website www.processmining.org. There you can find our free process mining tools (ProM and ProMimport), courses, papers, presentations, flyers, people, demo videos, etc.


 Santiago, June 30th 2008