4 main tasks of process discovery
- Defining the setting: assembling a team
- Gathering information: building an understand of the process through discovery
methods - Conducting the modeling task: actual creation of the process model
- Assuring process model quality: consideration of quality criteria and establishing trust
2 fundamental roles of process discovery
- Process analysts:
responsible for gathering information with process modeling languages - Domain experts:
intimate knowledge of how a process is performed, typically process participants, not proficient in process modeling
challenges
- Fragmented process knowledge:
- info from different domain experts
- diverging assumptions likely to surface
- iterative approach - Thinking of a processes on a case level
- Domain experts have problems responding to general questions
- questions need to reverse-engineer the conditions that govern the routing decisions of a process - Lack of familiarity with process modeling languages
Process discovery methods
- evidence-based discovery:
1. document analysis
2. observation
3. automated process discovery
-interview-based discovery
-workshop-based discovery
document analysis
- documents point to existing roles, activities, and business objects
-issues:
○ not process-oriented
○ too abstract or detailed
○ outdated or idealistic
observation
-follow the execution of individual process instances and then abstract active and passive roles
-issues:
○ no big picture (active)
○ non-natural behavior (passive)
automated process discovery
-use of event logs and process execution data
-issues:
○ unavailability
○ selection bias
○ hard to understand
Interview-based discovery
- based on interviewing multiple domain experts to inquire about how a process is executed
- phases:
1. interview
2. process modeling
3. model validation - basic strategies:
1. forward –> follow the natural process flow
2. backward –> start with process outcomes - issues:
○ rests on the assumption that process analysts and domain
experts use the same technology
○ discrepancies between domain experts
○ shows tendency towards neglecting exceptional behaviour
Workshop-based discovery
- 2 roles:
1. facilitator → coordination of verbal contributions
2. process modeler → creation of the process model - allows the perspective of multiple domain experts at the same time
- issues:
○ simultaneous availability of domain experts
○ speech time of talkative participants
process modeling steps
- Identify process boundaries (definition of start and end events)
- Identify activities and (intermediate) events
- Identify resources and their handoffs (pools, lanes and sequence flows)
- Identify control flow
- Identify additional elements
Quality assurance
- syntactic quality
- semantic quality
- pragmatic quality
syntactic quality
- Conformance of a process model to the structural rules and behavioral rules of the modeling language used
-structural correctness
-behavioral anomalies
-soundness
structural correctness
- Element-level rules:
activities, events, flows and gateways - Model-level rules:
all flow nodes must be on a path from a start to an end event
Behavioral anomalies
-Deadlock: a running process instance is not able to progress any further
-Livelock: a process instance keeps cycling in a loop
-Lack of synchronization: 2 or more tokens are in the same sequence flow
-Dead activity: an activity can never be executed in any instance of the
process
soundness
-Option to complete
-Proper completion –> at the moment of completion, each token of the process instance must be in a different end event
-No dead activities
semantic quality
-adherence of a process model to its real-world process
-visual and logical structures
-meaningful labels
2 criteria of semantic quality
- validity: all statements that one can make from the model are correct and relevant
- completeness: model contains all relevant statements about the corresponding process
Pragmatic quality
-usability of a process model:
1. understandability
2. maintainability
3. learning
-modeling conventions:
○ Vocabulary → avoiding certain elements
○ Structure → limiting the model structure
○ Semantics → avoiding certain element meanings
○ Appearance → restricting the model appearance in terms of labels, layout, and notation
7 Process Modeling Guidelines (7PMG)
- Use as few model elements as possible
- Minimize the routing paths per element
- Use one start event for each trigger and one end event for each outcome
- Model as structured as possible
- Avoid OR gateways where possible
- Use verb-object activity labels
- Decompose a model with 30+ elements
