Process Quality and Architectures
Companies benefit from using concepts, methods, and techniques of Business Process Management by improving their operations. All of the company’s business processes follow a life cycle that comprises the phases of design & analysis, configuration, enactment, and evaluation. In each of these phases process quality and architectures comes into play.
Important Dates
8th February 2012 3 p.m. - Final Presentation
25.02.2012 - Final Paper and Prototype Submission
If not indicated, the submission deadline is usually 23:59:59 CET.
Slides
Seminar Structure
The seminar will contain two introductory lectures:
- How to write a research paper
- How to present a scientific talk
Topic Overview
Topic Proposals
Process Monitoring
Process Monitoring is a central discipline to find out what quality a process provides. This includes measuring of process instances and their activities, respectively extracting such data from various IT systems, e.g. workflow systems. Representation/transformation of this data in process models offers multifaceted research questions.
- Nico Herzberg
- How to model measurement/event-capturing points in BPMN models
- Context: One of today’s state-of-the-art modeling languages is BPMN. These models are used to run processes. During process execution log data (e.g. timestamps) are written in several systems. These data can be seen as measurement points that need to be displayed with the model for monitoring and analysis purposes to ensure and increase business process quality. But BPMN does not provide any suitable notation of such measurement points in the model.
- Task: Evaluate already existing approaches about modeling of measurement points in models. If there is no suitable approach come up with an own notation that allows measurement point modeling at least in a separate perspective to the model.
- Literature:
- Río-Ortega, Manuel Resinas Adela del: Towards modelling and tracing key performance indicators in business processes. Actas de los Talleres de las Jornadas de Ingenier ́ıa del Software y Bases de Datos, 3(3), 2009.
- Viara Popova, Jan Treur: A specification language for organisational performance indicators. 2005.
- Christof Momm, Robert Malec, Sebastian Abeck: Towards a model-driven development of monitored processes.
- Giese, Philipp: Introducing KPIs to BPMN. Hasso-Plattner-Institut 2011
- Pourshahid, Alireza. A URN-Based Methodology for Business Process Monitoring. Master thesis, University of Ottawa, Canada. 2008.
- Pourshahid, A., Chen, P., Amyot, D., Weiss, M. and Forster, A., (2007) Business Process Monitoring and Alignment: An Approach Based on the User Requirements Notation and Business Intelligence Tools. 10th Workshop of Requirement Engineering.(WER’07), pp. 80-91
- Techniques to learn about human actions during process execution
- Context: Many business processes are semi-automated, means that there is user interaction in between. To allow effective process monitoring, to ensure and increase business process quality, the events of user interaction are very valuable and needs to be captured as well.
- Task: Investigate on different types of user interaction in a business process (e.g. filling forms, pressing buttons, writing e-mails, calling somebody) and how they can be captured in a process log. Cluster those different types of user interaction in a suitable way. Select a relevant interaction type, apply it to a prototype and test it in a user study (e.g. measure awareness for capturing events in that way).
- Literature:
- M. Kunze, A. Grosskopf, H. Overdick, and M. Weidlich. Lightweight Collaboration Management. In Proceedings of the Conference on Object Oriented Programming Systems Languages and Applications, 2009
- J. Canny. The Future of Human-Computer Interaction. In Queue – HCI Volume 4 Issue 6, July-August 2006
- Andreas Rogge-Solti
- Exception Handling for Events
- Context: In medical care in hospitals, exceptions are the norm. Patients have very different indiviual traits and physical conditions that might lead to different outcomes of the same treatment process. Standardization efforts are being made to capture best practices for well known procedures and operations. However the standard "happy path" does usually only cover a rough majority of the cases.
- Task: Your task is to first get an understanding of what kind of exception handling mechanisms are out there in literature. As a second step exception handling mechanism based on rules should be investigated and compared with a taxonomy. Rule based systems Bonus question: (How) can these exception rules be integrated into an event monitoring system?
- Literature:
- [1] Luo, Z., Sheth, A., Kochut, K., & Miller, J. (2000). Exception Handling in Workflow Systems 1, 13(2), 125-147.
- [2] Russell, N., Aalst, W. V. D., & Hofstede, A. (2006). Workflow Exception Patterns. Advanced Information Systems Engineering, 288-302.
- [3] Adams, M., Hofstede, A.H.M., Edmond, D., & Aalst, W.M.P.v.d. (2005). Facilitating Flexibility and Dynamic Exception Handling in Workflows through Worklets. Knowledge Acquisition, 45-50.
- [4] Adams, M., Hofstede, A.H.M., Aalst, W.M.P.v.d., & Edmond, D. (2007). Dynamic, Extensible and Context-Aware Exception Handling for Workflows. Most, 95-112.
- Analytical Process Simulation Techniques (Literature review)
- Context: Business process simulation by discrete event simulation is quite straight forward. The quality of the result depends on how many tokens are run through the model, so these techniques depend on resources and duration of the simulation.
- Task: Get familiar with analytical methods for process simulation, provide an overview of what is possible and what is not. Where are the exact limits of analytical methods?
- Literature:
- [1] Zimmermann, A. (2007). Stochastic discrete event systems: Modeling, evaluation, applications, Springer-Verlag New York Inc.
- [2] Bhavan Pukhraj Shah (1993). Analytic solution of stochastic activity networks with exponential and deterministic activities. (master's thesis)
- Visualization techniques in monitoring
- Context: The interface of a business process monitoring system is driven by the questions, the system tries to answer to the stakeholders utilizing it.
- Task: Get an overview of published business process monitoring interfaces and provide a list of typical questions these systems answer. Learn also psychological rules and effects of visualization and map these to the questions. Design a paper prototype and get feedback from our research partner hospital in Jena.
- Literature:
- [1] Böhme, R. (Last semester's Master seminar work)
- [2] Moody, D. (2009). The “Physics” of Notations: Toward a Scientific Basis for Constructing Visual Notations in Software Engineering. IEEE Transactions on Software Engineering, 35(6), 756-779.
Process Similarity Search
As Business Process Management became one major driver in nowaday's companies, process model repositories emerged that contain hundreds or thousands of process models. They carry a central asset of the business. However, to use them effectively requires capabilities to efficiently manage these models, in particular means to find process models.
The area of process model similarity search focusses on this problem. Similarity of process models can be based on different aspects, i.e., structure and behavior. Search for similar process models in a repository requires a query model and a search radius that specifies how similar a model of the search result shall be to the query. As their is no natural or meaningful ordering among the structure or behavior of models, one can either resort to sequential search, i.e., compare the query with each and every model in the repository, or leverage efficient index structures and search algorithms in metric spaces.
The proposed topics in this field do not directly address such indexing approaches, but rather focus on foundations of similarity and how search results can be qualitatively evaluated.
- Matthias Kunze
- Quality Measures for Search Results
- Context: The result of similarity search may not always be very convincing, e.g., when every model of the result has a similar distance to the query.
- Task: Find means to evaluate the search result towards its quality, i.e., assess, whether it has a good quality, provide meaningful indicators for end users, and propose means to adjust the ranking of the search result. Some ideas are:
- evaluate the distance distribution between query and result set of models
- evaluate the agreement of different similarity notions towards the ranking
- evaluate the ratio between search radius and distance to best/worst result
- Literature:
- K. Beyer, J. Goldstein, R. Ramakrishnan, and U. Shaft. When is "nearest neighbor" meaningful? In In Int. Conf. on Database Theory, pages 217–235, 1999.
- R. Y. Wang and D. M. Strong. Beyond accuracy: what data quality means to data consumers.
J. Manage. Inf. Syst., 12:5–33, March 1996.
- P. Zezula, G. Amato, V. Dohnal, and M. Batko. Similarity Search: The Metric Space Approach. Springer-Verlag New York, Inc., Secaucus, NJ, USA, 2005.
- Existing Similarity Measures (Literature Review)
- Context: A large number of process model similarity measures have been proposed over the last two decades, and in particular recently. However, their relation is not clear.
- Task: Review literature for similarity measures. Collect, understand and categorize them in a catalog. Address questions such as: What to they have in common. Also, given a set of process models, evaluate in which cases they agree and how much, as well as how good do these measures simulate human assessment of process model similarity.
- Literature:
- M. Dumas, L. Garcìa-Bañuelos, and R. M. Dijkman. Similarity Search of Business Process Models. IEEE Data Eng. Bull., 32(3):23–28, 2009.
- Michael Becker, Ralf Laue: Analysing Differences Between Business Process Similarity Measures, International Workshop on Process Model Collections, Clermont-Ferrand, 2011
- ... many more papers on individual measures
- Process Model Alignment (together with Sergey Smirnov)
- Context: Any process similarity measure relies on a process alignment, that is, information, which nodes represent the same activity in two processes. Such an alignment is usually based on similarity of phrases, i.e., labels of business process models, but takes also into account the relation of activities to their ancestors. Alignments can be 1:1, n:1, and iteratively be optimized.
- Task: Research existing string similarity and activity alignment algorithms and integrate them into a general framework.
Experimentally compare and evaluate these algorithms.
- Literature:
- Felix Naumann, Melanie Herschel: An Introduction to Duplicate Detection. Morgan & Claypool Publishers 2010
- Jonathan Gumpp: Die Berechnung der Ähnlichkeit von Geschäftsprozessen. Eine Implementierung. Seminararbeit 2010.
- Matthias Kunze, Matthias Weidlich, Mathias Weske: Efficient Retrieval of Similar Business Processes (Draft).
- existing algorithms and documentation
Process Quality and Architectures in General
- Rami-Habib Eid-Sabbagh
- Generating process models from forms, data and form annotation
- Context: In the public sector many processes derive from the same regulations and laws which purport the rough structure of the process. These regulations are mirrored in the application forms of the according public service. Hence the hypothesis is that one can derive processes or parts of processes from application forms. Forms bear information on data, roles, and activities or workflow semantics. Extracting business models from text and from text to business models has been researched already. Based on these ideas analyze application forms, data and annotations to forms as well as according regulations and apply and adapt given methods or design new ones.
- Task:
- Research on text based derivation of process models
- Analyze application forms, e.g. Construction permit
- Define relationships between parts of the application forms and process parts – roles, activities, data, events and control flow elements
- Extract process models from application forms
- Implement a tool that creates process models or parts of process models from application forms
- Define compliance relations between forms and given processes
- Literature:
- Manuel Blechschmidt – Master Seminar
- Fabian Friedrich Master Thesis – Automated Generation of Business Process Models from Natural Language Input *“From textual scenarios to a conceptual schema” Günther Fliedl, Christian Kop, Heinrich C. Mayr 2004 data and knowledge engineering
- Generating Natural Language specifications from UML class diagrams Farid Meziane, Nikos Athanasakis ,Sophia Ananiadou Requirements Engineering 2006
- IT-Services to Business Process Alignment
- Context: More and more processes are elicited in the German administration bearing reoccurring process parts. Parallely large administration start to introduce service oriented infrastructures to support the automization of their processes. Many of the processes in German administration may contain the same process blocks. Identification of these process blocks from which all processes in German administration can be composed is challenging but a necessity for aligning services and porcesses.
- Task:
- Research on IT-Infrastructures in the Public Sector
- Analyze Process models and identify reoccuring process parts
- Identify services
- Match IT-Systems to process models
- Link services to process parts and process models
- Design an information systems that manages and gives insights on these relations, e.g which services realizes a particular process
- Modeling Guidelines for BPMN 2.0 (together with Gero Decker)
- Context:
- Guidelines help to create readable and consistent models (especially if many modelers are involved)
- How do I create suitable modeling guidelines?
- Task:
- Collect and categorize existing guidelines (from books, consultants, end users)
- Example rule categories: BPMN subset, labeling, model structure, layouting, (anti-)patterns
- Which guidelines can be applied / checked automatically?
- Evaluate existing / conceive novel algorithms for this
- References:
- „7 Modeling Principles“
- Camunda Guidelines
- „BPMN Method and Style“
- Interviews
- Artem Polyvyanyy
- Towards Canonical Process Models
- Context: A process can be seen as a partial order of tasks. Process models describe processes as compositions of tasks by means of gateways and control flow arcs. An interesting theoretical problem to investigate is to discover the minimal set of tasks that can be composed in a process model which describes the same process as a given process model. A process model with the minimal set of tasks can be seen as the canonical version of the given process model. Tasks in canonical process models can be treated as such that have unique names (even if two tasks share the same name, e.g., two tasks with name A can be treated as such that are labeled A1 and A2, where A1 and A2 are semantically same as A).
- Task: Propose an initial definition (describe properties) of a canonical process model. Motivate canonical process models and collect use cases, e.g., it is faster to check isomorphism of clones.
- Literature:
- Artem Polyvyanyy, Luciano García-Bañuelos, Marlon Dumas: Structuring Acyclic Process Models. BPM 2010:276-293
- Artem Polyvyanyy, Luciano García-Bañuelos, Dirk Fahland, Mathias Weske: Maximal Structuring of Acyclic Process Models CoRR abs/1108.2384 (2011)
- Reina Uba, Marlon Dumas, Luciano García-Bañuelos, Marcello La Rosa: Clone Detection in Repositories of Business Process Models. BPM 2011:248-264
Additional Information
- All submissions have to be send via e-mail to nico..herzberg (at) hpi..uni-potsdam..de
