Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Detecting Frequently Recurring Structures in BPMN 2.0 Process Models (SummerSOC 2015)

500 views

Published on

Skouradaki, Marigianna; Leymann, Frank: Detecting Frequently Recurring Structures in BPMN 2.0 Process Models. In: Proceedings of the 9th Symposium and Summer School On Service-Oriented Computing: SummerSOC'15; Heraklion, Greece, June 28 - July 04, 2015.
==========================================
Abstract
Reusability of process models is frequently discussed in the
literature. Practices of reusability are expected to increase the performance
of the designers, because they do not need to start everything from scratch,
and the usage of best practices is reinforced. However, the detection of
reusable parts and best practices in collections of BPMN 2.0 process
models is currently only defined through the experience of experts in
this field. In this work we extend an algorithm that detects the recurring
structures in a collection of process models. The extended algorithm
counts the number of times that a recurring structure appears in a
collection of process models, and assigns the corresponding number to
its semantics. Moreover, the dublicate entries are eliminated from the
collection that contains the extracted recurring structures. In this way,
we assert that the resulting collection contains only unique entries. We
validate our methodology by applying it on a collection of BPMN 2.0
process models and analyze the results. As shown in the analysis the
methodology does not only detect applied practices, but also leads to
conclusions of our collection’s special characteristics.

Published in: Software
  • Be the first to comment

Detecting Frequently Recurring Structures in BPMN 2.0 Process Models (SummerSOC 2015)

  1. 1. University of Stuttgart Universitätsstr. 38 70569 Stuttgart Germany Phone +49-711-685 88477 Fax +49-711-685 88472 Research Marigianna Skouradaki, Frank Leymann Institute of Architecture of Application Systems {firstname.lastname}@iaas.uni-stuttgart.de “Detecting Frequently Recurring Structures in BPMN 2.0 Process Models” SummerSOC 2015
  2. 2. 22 Research © Marigianna Skouradaki BPMN 2.0 Process Models Collection Fragments Repository Workload Mix Graph Matching Selection Criteria Composition Criteria 80% 20% Motivation: Generation of Realistic Workload
  3. 3. 33 Research © Marigianna Skouradaki Agenda  Process Model Matching  Basic Concepts  Algorithms  Validation & Discussion  Conclusions & Outlook
  4. 4. 4 Process Model Matching
  5. 5. 55 Research © Marigianna Skouradaki BPMN 2.0 Collection Characteristics  Detect the reoccurring structures on BPMN 2.0 process models which:  Might be anonymized (no text information available)  Might be mock-up models (non-executable)
  6. 6. 66 Research © Marigianna Skouradaki Text Matching  Cannot be applied to anonymized models
  7. 7. 77 Research © Marigianna Skouradaki Behavioral Matching log  Cannot be applied on mock-up models log 2
  8. 8. 88 Research © Marigianna Skouradaki Structural Matching
  9. 9. 99 Research © Marigianna Skouradaki The Challenge of Graph Isomorphism NonDeterministic Polynomial Time (NP – Complete) The time required to solve the problem using any currently known algorithm increases very quickly as the size of the problem grows.
  10. 10. 1010 Research © Marigianna Skouradaki Subgraph Isomorphism on BPMN 2.0 Process Models BPMN 2.0 Process Models are special types of graphs Subgraph isomorphism can be applied in lower complexity1 1 R. M Verma.; and S. W. Reyner; “An analysis of a good algorithm for the subtree problem, correlated,” SIAM J. Comput., vol. 18, no. 5, pp. 906–908, Oct. 1989.
  11. 11. 11 Basic Concepts
  12. 12. 1212 Research © Marigianna Skouradaki Exiting Attributes: Nested
  13. 13. 1313 Research © Marigianna Skouradaki Exiting Attributes: Different Positions
  14. 14. 1414 Research © Marigianna Skouradaki Exiting Attributes: Partially Similar
  15. 15. 1515 Research © Marigianna Skouradaki Process Fragment A Process Fragment is a piece of process model with loose completeness and consistency. The existence of process graph elements (start, end, activities, context etc.) is possible but not imperative in a process fragment. However, a process fragment must have at least one activity and there must be a way to convert it to an executable process graph.2 1. It is not necessarily related with a complete process model 2. A starting point is not defined 3. Existence of split, merge node or event is optional 2D. Schumm, F. Leymann, Z. Ma, T. Scheibler, and S. Strauch, “Integrating Compliance into Business Processes: Process Fragments as Reusable Compliance Controls” in MKWI’10, Göttingen, Germany, February 23-25, 2010, Ed., Conference Paper, pp. 2125–2137.
  16. 16. 1616 Research © Marigianna Skouradaki Checkpoints & Relevant Process Fragments (RPFs)  Checkpoint (the starting points)  A pre-configured type of node that is used as start point of the “extended” process fragments  Relevant Process Fragment  Exists in at least K business processes  Starts with a checkpoint  Has at least N nodes including the checkpoint  Contains at least one activity
  17. 17. 1717 Research © Marigianna Skouradaki Checkpoints & Relevant Process Fragments K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  18. 18. 1818 Research © Marigianna Skouradaki Checkpoints & Relevant Process Fragments K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  19. 19. 1919 Research © Marigianna Skouradaki Checkpoints & Relevant Process Fragments K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes RPF RPF
  20. 20. 2020 Research © Marigianna Skouradaki Checkpoints & Relevant Process Fragments K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  21. 21. 2121 Research © Marigianna Skouradaki Checkpoints & Relevant Process Fragments K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes RPF RPF
  22. 22. 2222 Research © Marigianna Skouradaki Checkpoints & Relevant Process Fragments K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  23. 23. 2323 Research © Marigianna Skouradaki Checkpoints & Relevant Process Fragments K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes RPF
  24. 24. 24 Algorithms
  25. 25. 2525 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  26. 26. 2626 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  27. 27. 2727 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  28. 28. 2828 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  29. 29. 2929 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  30. 30. 3030 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  31. 31. 3131 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  32. 32. 3232 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  33. 33. 3333 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  34. 34. 3434 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  35. 35. 3535 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  36. 36. 3636 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  37. 37. 3737 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  38. 38. 3838 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  39. 39. 3939 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  40. 40. 4040 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  41. 41. 4141 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  42. 42. 4242 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  43. 43. 4343 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  44. 44. 4444 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  45. 45. 4545 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  46. 46. 4646 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  47. 47. 4747 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  48. 48. 4848 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  49. 49. 4949 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  50. 50. 5050 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  51. 51. 5151 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  52. 52. 5252 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes
  53. 53. 5353 Research © Marigianna Skouradaki Algorithm: Discovery of RPFs K = 2 Process ModelsCheckpoints: Events, Gateways N = 3 Nodes …continue likewise..  Will not work for cycles
  54. 54. 5454 Research © Marigianna Skouradaki Discover Duplicates and Count Appearance RPF Collection Newly Discovered RPF
  55. 55. 5555 Research © Marigianna Skouradaki Discover Duplicates and Count Appearance RPF Collection Newly Discovered RPF
  56. 56. 5656 Research © Marigianna Skouradaki Discover Duplicates and Count Appearance RPF Collection Newly Discovered RPF For each RPF in Collection: If RPF have the same size: COMPARE
  57. 57. 5757 Research © Marigianna Skouradaki Discover Duplicates and Count Appearance RPF Collection Newly Discovered RPF
  58. 58. 5858 Research © Marigianna Skouradaki Discover Duplicates and Count Appearance RPF Collection Newly Discovered RPF 2
  59. 59. 5959 Research © Marigianna Skouradaki Discover Duplicates and Count Appearance RPF Collection Newly Discovered RPF
  60. 60. 6060 Research © Marigianna Skouradaki Discover Duplicates and Count Appearance RPF Collection Newly Discovered RPF 1
  61. 61. 61 Validation and Discussion
  62. 62. 6262 Research © Marigianna Skouradaki Validation  43 BPMN 2.0 Process Models  BPMN 2.0 Standard Example Processes  Models used in Pietsch and Wenzel, 2012  903 Comparisons  1544 non-filtered RPFs  83.22% decrease of results when filtering duplicates (259 RPFs)  54 RPF appear > 1 time Median = Threshold = 14  27 RPFs with re-appearance rate above the threshold
  63. 63. 6363 Research © Marigianna Skouradaki Some Representative RPF and Number of Appearance ID Fragment Count 1 178 2 169 3 117 4 101 5 62 6 60 7 44 8 42 9 42
  64. 64. 6464 Research © Marigianna Skouradaki Conclusions & Outlook  Extension of RPF Discovery algorithm  Automatic count of the RPF appearance in collection  We have evaluated the approach on 43 BPMN 2.0 process models  Conclusions on frequently used structures (best practices)  Conclusions for collection’s special characteristics  Extend the algorithms for the complete set of BPMN 2.0  Apply to thousands real world BPMN 2.0 process models and execute thorough analysis  Implement the prototype for process synthesizing methodology Thank You!!! Marigianna Skouradaki: skourama@iaas.uni-stuttgart.de 

×