BPM – Business process management

Lecture 1
introduction
 The business process: each product or service that a company provide to the market
is the outcome of a number of different activities performed
 In order to create, maintain or improve a company’s business, one has to
understand what the company does
 Pillars of company: people, processes, information, technology
o BP(M) is used to acquire an understanding of the business and be able to
organize business activities in a meaningful way
o BI is used to do things better based on an understanding of the business
Business processes
 A business process:
o a set of activities performed in coordination to jointly realize a business goal
o May involve one or multiple departments or different organizations
o May interact with business processes of other organizations
o May be part of industry value chain which creates value for customers
o We restrict a business process to be a part of a single business organization
 Once business processes are defined, they can be subject to analysis, improvement,
and enactment
 BPM supports the business process lifecycle
o Design cycle: iterative cycle consisting of modeling and analysis for improving
the business process definition
o Engineering cycle: iterative cycle for improving the business process
implementation in the organization
o Design/analysis -> configuration -> enactment -> evaluation
BPM systems
 Business processes can be enacted manually
 Automation of business process by software systems can bring considerable benefits
 Business process management system: software system that coordinates enactment
of business processes driven by explicit process schema
o BPM model vs executable models (workflow models)
o Separation of process logic and application logic
To sum up
 Business process describes how a company operates
 Business intelligence acquires information on the operation of a company to support
decision making
 BPM supports the BP lifecycle
o From design/analysis -> configuration -> enactment -> evaluation
 BPM system automates business processes
 It coordinates execution of business process instances
Identification and representation
 Abstract from real world details and technology, but be concrete and precise where
it matters
 Identify activities, input, output, relationships
To sum up
 There are different languages to model processes
 Control flow: coordinated activates
o Gateways to allow splitting and joining control flows
o Transition conditions can be evaluated at runtime to determine with which
path to continue execution
 Data flow: links data objects to activities
o Can be referenced by transition conditions
o Data from outside environment can be passed to the process via start event,
and from the process to the environment via end event

Representation of ordered message flow while abstracting from connected business

processes
 In order to realize correct interactions, participants need to agree on the flow of
messages
o If a process wants to send a message, the other process should be able to
receive it
o If a process wants to receive a message, the other process should be able to
send it
 Sometimes it is useful to define the flow of messages separated, e.g., to facilitate
cooperation in a business sector
o Abstract from internal details of participants
 This can be done with a process choreography (inter-organizations) or
orchestration (intra-organizations)
 Choreographies and Orchestrations define minimal but sufficient constraints on
participants to collaborate at a process level
analysis
 Once the business process is modelled, we can try to assess its efficiency and find
weaknesses and limitations in the current situation
o Process performance analysis: cost, quality, flexibility, speed, ... •
 Before that, we must be sure that the model is correct
o Process correctness analysis: syntactically correct, structurally sound,
satisfying soundness properties
Process correctness analysis: deadlock and other problems
 Deadlock: possibility that a process cannot be completed
 No proper completion: possibility that some activities are left in enabled or running
state after termination of process
 Dead activity: an activity that is never completed after termination of the process
 Missing data: possibility that required data for an activity is not produced
 Lost update: possibility that a data update is overwritten by another
activity before the activity that requires the update can read it
Process Performance Analysis Measures
 Process capacity: maximum output rate measured in units produced per unit of
time; depends on capacity of activities in sequence flows and parallel flows
 Capacity utilization: percentage of the process capacity that is being used
 Throughput (or flow) rate: average rate at which units flow pass a specific point in
the process (max. throughput = process capacity)
 Flow (or throughput) time: the average time a unit requires to flow through the
process
 Cycle time: time between successive units output by the process (cycle time =
1/throughput rate)
 Idle time: time that no work is done; time that activity is waiting for work to arrive
from previous activity
 Inventory: maximum number of units in progress within the process (Little’s Law:
inventory = throughput rate x flow time
To sum up
 Correctness of business process models should be checked after each design cycle
o Deadlocks, no proper completion etc.
 o Formal methods available for this (not treated in this lecture)
 Annotated business process models can be used to gain some insights on
performance
o Capacity, throughput etc.

Lecture 2: BPM Methodology

 Supply chain: a process produces results for the next link in the chain
 As a management discipline the focus of business process management (BPM) is
twofold,
o (i) on integrating the strategy of an organization with its operations.
o (ii) on delivering operational excellence

 Primary
 Secondary
 Third (management of the processes)
 Lean Principles (they change between definitions):
o Focus on the customer.
o Identify and understand how the work gets done (the value stream) .
o Manage, improve, and smooth the process flow.
o Remove Non-Value-Added steps and waste.
o Manage by fact and reduce variation.
o Involve and equip the people in the process.
o Undertake improvement activity in a systematic way
 Agile, we need to change
o VUCA world; Volatility, uncertainty, complexity, ambiguity

 BPM quick start

o Phase 1: process identification
o Phase 2: process optimization
o Phase 3: Process realization
 BPR cycle: identify processes -> review update -> design to be -> test and
implement
 Deming cycle (PDCA cycle): plan, do, act, check
 Mendix modeling methodology
BPM meta-models

 M0: the physical world

 M1: the models
 M2: meta model
 M3: meta meta model
 7PMG
o G1 Use as few elements in the model as possible
o G2 Minimize the routing paths per element
o G3 Use one start and one end event
o G4 Model as structured as possible
o G5 Avoid OR routing elements
o G6 Use verb-object activity labels
o G7 Decompose a model with more than 50 elements
Lecture 3: Business process automation
Business process automation
 Efficiency of business processes can be improved by software systems that partially
automate the processes
o Automation opportunities
o Orchestrating activities in compliance with process model
o Creating ‘work items’ for human activities and assigning these to qualified
actors
 Invoking application services for system activities and exchanging parameter data
with these services
 Combined in process-aware information systems (PAIS)

Automation islands
 Evolved from uncoordinated automation projects of different units in an
organization to improve their own performance
 independent decisions with respect to automation solutions: technology, systems,
data formats, communication protocols, ...
 Automation islands (software systems that automate parts of a process) often
cannot directly exchange or share information with other systems

Consequences of ‘bottom-up’ automation

• Automation is local and ad-hoc
o Limited impact at process level if bottlenecks are not addressed
o Automation islands
 Inability to share or exchange information
 Inconsistent storage of related information
• •Difficulty in connecting the diverse automation islands

Connecting automation islands

• Manual mediation: tedious and error-prone
• Automated translations: middleware solutions addressing
o Interoperability: systems A and B can exchange data, and they can interpret
and use the received data
o Integration: combining systems A and B into new system such that users of
the new system can use functions and data of A and B in a uniform way
• Interoperability is a shared multi-level problem which cannot be solved with ad-hoc
and proprietary solutions
o Standards
o Adoption of and compliance to (same) standards
Workflow management systems
Workflow management systems
• Executable process model (workflow)
• Coordinates tasks of software applications and humans
• Pro
o Explicit modeled business process
o No hard-coded application integration
o Separation of process logic from application code

Workflow concepts
• Workflow: executable process, realizes automation of business processes (hence
implementation of business process models)
• Workflow instance: execution of workflow
• System workflow: execution does not require human user involvement (all
activities are implemented by software applications)
• Human interaction workflow: execution requires human user involvement (some
activities are implemented with work performed by human users)
o Work item: task that needs to be assigned to a qualified human user
o Work (item) list: GUI-based asynchronous interaction mechanism for
presenting collection of currently available tasks to human user
o Work item allocation: mechanism for assigning tasks to humans, possibly
based on an actor expression associated with the task and used to query an
organizational model

Workflow instance and its lifecycle

• Once an executable workflow model is deployed to the workflow engine, workflow
instances can be created and executed
• The workflow engine can have multiple executable workflow models deployed to it,
and multiple instances of each model may exist at any time
• Each instance exhibits an internal state
o Progress towards completion
o Status of activities and data objects
• Workflow engine uses state model to manage concurrent execution of instances
• Monitoring of progress of running instances can be based on recording relevant
events during execution in event/execution log or audit trail

To sum up
• Workflow Management Systems (WfMS) area prominent example of PAIS, featuring
o Automation based on process-thinking
o Separation of process logic from application logic
o Build-time environment for creating executable process models (workflows)
o Run-time environment with process engine that creates, executes and
manages workflow instances
Lecture 4: Business process analysis with petri nets
Business Process Analysis
 Business process implemented in a company should have required properties
 Special languages tailored to investigation of (structural) properties
o Activity net: graph-based, explicit representation of data dependencies,
precise but informal semantics
o Petri net: graphical representation with formal semantics
 Typically, an analysis model (using an analysis language such as PN) is made after the
(BPMN) business process model has been agreed upon
Petri Nets
 Most widely used for specifying and analyzing business processes
o Abstract: independent of execution environment
o Formal: well-defined and precise semantics
 Different classes of Petri nets
o Common definition of static structure
o Dynamic behavior definitions for different trade-offs between model
simplicity and ability to capture real-world properties of business processes
 Places denote conditions and are represented as circles
 Transitions denote activities and are represented as rectangles or vertical lines
 Arcs are directed arrows which connect places with transitions and
 vice versa
o A place p is an input place for a transition t if there is a
directed arc from p to t
o A place p is an output place for a transition t if there is
a directed arc from t to p
  Dynamic behavior is represented in Petri nets with tokens, represented as black dots
within places
 Subject to the rules defined by the used type of Petri net
o The presence of tokens in input places and output places of a transition
determines whether a transition is enabled
o An enabled transition can take place, or ‘fire’
o The firing of the transition results in tokens being removed from input places
and tokens being added to output places
 The state of a Petri net is indicated by the distribution of tokens among its places,
called marking
o Represented by an array of integer elements, with the indices corresponding
to places and the integer value of an element indicating the number of
tokens in the corresponding place
Event condition (E/C) Petri Net
 EC: Simplest type of Petri net, where places express simple conditions for activities
to happen
 Each place may carry one token at maximum
o If a place carries a token its condition is satisfied
o Places without token correspond to dissatisfied conditions

Place Transition (P/T) Petri nets

 PT: Extension of E/C-Nets in order to express more complex relationship between
conditions and activities
o Place has capacity (default ) for holding tokens
o Arc has weight ≥ 1 (default 1) for removing/adding tokens if transition fires
 A transition t is enabled if each of its input places contains at least the number of
tokens defined by the weight of the connecting arc
 When a transition t fires, the number of tokens withdrawn from its input places and
the number of tokens added to its output places (Constrained by their capacity) are
determined by the weight of the respective arcs
 Reachable state: A state reachable from the current state by firing a sequence of
enabled transitions
 Dead state: A state where no transition is enabled
More about Petri Nets
 E/C-Nets and P/T-Nets have important limitations
o Not very useful for distinguishing business process instances
o Elaborate representations are needed for certain common patterns
o Not possible to handle time and data
 Extentions were proposed to address these limitations
o Coloured Petri nets: tokens have colours, i.e. have typed values (e.g. process
instance identifiers) – Read section 4.2.3 of Weske’s book
o Time Petri nets: tokens have time stamps
o Hierarchical Petri nets: transitions can be decomposed into subnets (also
Petri nets)
Workflow Net
 Enhance traditional Petri nets to ease the representation of business processes
o Notation (syntactic sugaring) to represent AND-split, AND-join, XOR-split,
XOR-join
o Annotations to transitions (called triggers) that provide information on who
or what is responsible for an enabled transition to fire
 Introduce structural restrictions on Petri nets that prove useful for investigating
soundness of business processes
o Unique initial place (no incoming edge)
o Unique final place (no outgoing edge)
o Every place and every transition is located on a path from the initial place to
the final place
Conclusion
 Petri nets are most widely used to model and analyse business processes
o Dynamic behavior is captured by ‘token play’, which shows the succession of
different ‘markings’ (states) that are possible for the execution of a process
instance
o Operationalizations (e.g., reachability analysis) to check soundness, e.g.
reachability of states and presence of dead states
o More powerful expresssion of real-world business processes and analysis of
soundness properties is possible with richer Petri net variants
Lecture 5 BPM languages
 Phase 1: process identification
o Basic objectives for business process reengineering
o Identifying relevant Processes, Organizational Issues and Resource Allocation
 Phase 2: process optimalisation
o Process analysis (As-is)
o Process redesign (To-be)
 Phase 3: process realization
o Pilot process
o Process implementation
Key business process concepts
 Trigger →event that initiates a process
 Activity →task, activities have attributes such as: start time, duration, etc.
 Actor →party capable of accomplishing an activity/task; can be human, or
organization
 Control flow relation →relation between two activities; defines an order
(sequencing) between two activities
 Process →sequence of activities
 Flow Swim Lane →sequence of activities carried out by a particular actor
 Parallel flows →sequences of activities that are executed in parallel
 Alternative flow →succession of activities that is executed when a guard condition
takes the value true or false
 Loop →succession of activities that is executed multiple times, as long as a guard
condition is takes the value true or false
 Sub-process →part of a process
 Process Decomposition →dividing a process in sub-processes (based on the nesting
relation)
 Artifact →information object that is created, updated, accessed or deleted during an
activity (i.e., it is input or/and output of an activity)