What’s process discovery?
the act of gathering information about
an existing process and organizing it in
terms of an as-is process model.
What are the 4 main tasks of process discovery?
- Define the setting (assembling a team)
- Gather info (build an understanding of the process via discovery methods)
- Model task (actual creation with a modelling method)
- Assure process model quality (establish quality criteria and trust in process model)
2-4 are iterative
What are the 2 fundamental roles of process discovery?
Process analysts
- familiar with process modelling languages
- Responsible of gathering info
- hardly knowledgeable of all the process details
Domain experts
- intimate knowledge of process
- typically process participants, may also include external roles
- Nor efficient in process modelling
What are the 3 essential challenges of process discovery?
- Fragmented process knowledge: different domain experts, diverging assumptions, conflicts, iterative approach
- Thinking of process on a case level: problems responding to general questions
- Lack of familiarity with process modelling languages: can’t create or read process models
What are the 3 process discovery methods?
Evidence based:
- Document Analysis: existing documentations (might be outdated, lack info or not be process oriented)
- Observation: Follow in active or pasive way (not big picture or non natural behavior)
- Automated discovery: Event logs (not available, selection bias, hard to understand)
Interviews: 3 phases (interview, process modelling, model validation) 2 strategies (forward, natural process flow. Backwards, outcomes to triggers)
Workshops: 2 roles (facilitator, process modeler), usually 3 to 5 sessions
Create the Strengths and Weaknesses of the 3 process discovery methods
See notes
Which are the 5 steps for process modeling?
- Identify process boundaries (start and end events)
- Identify activities and (intermediate events) (main and intermediate events)
- Identify resources and their handoffs (pools and lanes, sequence flow)
- Identify control flow (order dependencies, AND, XOR/OR, loop structures)
- Identify additional elements (business objects)
How do we assure Process model quality?
Syntactic quality (verification):
- Conformance of a process model to the structural and behavioral rules
Semantic quality (validation):
- Adherence of a process model to its real-world process.
- Valid and complete
Pragmatic quality (certification):
- Usability of a process model
- Understandability: How easy is it to read a model?
- Maintainability: How easy is it to change a model?
- Learning: To what extent does a model reveal how its corresponding process works in reality?
What are the two syntactic rules that tell us if the model is structurally correct?
- Element-level
- Activities: at least 1 income & 1 outcome
- Events:
- Flows:
- Gateways: - Model-level
- All flow nodes must be on a path from a start to an end event
Which are the 4 behavioral activities anomalies should be avoided in syntactic quality?
- Deadlock: A running process instance is not able to progress any further once a given state is reached (i.e., a token gets stuck at that state);
- Livelock: A process instance keeps cycling in a loop (i.e., a token is trapped within a loop structure);
- Lack of synchronization: Two or more tokens are in the same sequence flow because they are not synchronized at some join gateway;
- Dead activity: An activity can never be executed in any instance of the process model.
What are the 3 behavioral rules to be satisfied that a process model has syntactic quality?
- Option to complete
- Any running process instance must eventually complete (i.e., there are no deadlocks and no livelocks); - Proper completion
- At the moment of completion, each token of the process instance must be in a different end event (i.e., no lack of synchronization); - No dead activities
- Any activity can be executed in at least one process instance (i.e., no dead activities).
What is a correct block structure so that a model can have syntactic quality?
If the two gateways match in type no behavioral anomaly can arise
What are some examples of non having pragmatic quality in labels?
- No decision activity
- No business object
- No common glossary
- Not following verb-object style
For pragmatic quality, what are the modeling conventions and guidelines?
Restrictions (either mandatory or suggested) to the following aspects of a process model:
- Vocabulary
- Structure
- Semantics
- Appearance
What are the 7PMG (7 Process Model Guidelines?
- Use as few model elements as possible (understandability and lower syntactic error rate);
- Minimize the routing paths per element (i.e., the number of incoming and outgoing arcs);
- Use one start event for each trigger and one end event for each outcome (understandability and lower error probability);
- Model as structured as possible (unstructured models more likely to show behavioral anomalies);
- Avoid OR gateways where possible
(complex semantics of both OR-join and OR-split); - Use verb-object activity labels (less ambiguous and more useful than action-noun labels);
- Decompose a model with 30+ elements
