Conference PaperPDF Available

Web-Based Modelling and Collaborative Simulation of Declarative Processes

Authors:

Abstract and Figures

As a provider of Electronic Case Management solutions to knowledge-intensive businesses and organizations, the Danish company Exformatics has in recent years identified a need for flexible process support in the tools that we provide to our customers. We have addressed this need by adapting DCR Graphs, a formal declarative workflow notation developed at the IT University of Copenhagen. Through close collaboration with academia we first integrated execution support for the notation into our existing tools, by leveraging a cloud-based process engine implementing the DCR formalism. Over the last two years we have taken this adoption of DCR Graphs to the next level and decided to treat the notation as a product of its own by developing a stand-alone web-based collaborative portal for the modelling and simulation of declarative workflows. The purpose of the portal is to facilitate end-user discussions on how knowledge workers really work, by enabling collaborative simulation of processes. In earlier work we reported on the integration of DCR Graphs as a workflow execution formalism in the existing Exformatics ECM products. In this paper we report on the advances we have made over the last two years, we describe the new declarative process modelling portal, discuss its features, describe the process of its development, report on the findings of an initial evaluation of the usability of the tool, resulting from a tutorial on declarative modelling with DCR Graphs that we organized at last years BPM conference and present our plans for the future.
Case Management Process drawn as a box without a bar on top of it. These are essentially DCR Graphs inside the root process that need to be initialized through the new spawn relation (→ * ), creating a new copy of the sub-process for each time it is spawned. In the example the activity Create Meeting spawns a new copy of Meeting each time it is executed and Create Document and Create Minutes create a new Document Handling process each time they are executed. When one has a relation between an event of a sub-process and an event of its parent , the relation will apply to each instance of the event. For example a case can not be archived while there are pending meetings; this is modelled by adding a condition from Hold Meeting to Archive Case, meaning that while there is at least one Hold Meeting-activity that has not yet been executed or excluded (by cancel meeting), it is impossible to execute Archive Case. In a similar manner it should not be possible to archive the case while documents are checked out. We model this through the milestone relation (→), which blocks an event as long as some other event is pending. Because of the response relation from Checkout to Checkin, there will be a pending response on the latter whenever the file is checked out. The milestone from Checkin to Archive Case ensures that as long as there is at least one instance of Checkin that is pending, we can not archive the case. Finally, to archive the case, Archive Case excludes all instances of all events in Document Handling, except for Download as it should still be possible to download files. All instances of Upload Minutes are also excluded, it is not necessary to exclude the other events in Meeting as this will already have been done by holding or cancelling the meeting and unless one of these has been executed the case can not be archived.
… 
Content may be subject to copyright.
[INDUSTRY PAPER]Web-based Modelling and
Collaborative Simulation of Declarative Processes
Morten Marquard1, Muhammad Shahzad1,2, and Tijs Slaats1,3
1Exformatics A/S, Dag Hammarskj¨
olds All´
e 13, DK-2100 Copenhagen Ø, Denmark
{mmq, ts}@exformatics.com, http://www.exformatics.com
2TEO International
muhammad.shahzad@teo-intl.com
3IT University of Copenhagen
Abstract. As a provider of Electronic Case Management solutions to knowledge-
intensive businesses and organizations, the Danish company Exformatics has in
recent years identified a need for flexible process support in the tools that we pro-
vide to our customers. We have addressed this need by adapting DCR Graphs, a
formal declarative workflow notation developed at the IT University of Copen-
hagen. Through close collaboration with academia we first integrated execution
support for the notation into our existing tools, by leveraging a cloud-based pro-
cess engine implementing the DCR formalism. Over the last two years we have
taken this adoption of DCR Graphs to the next level and decided to treat the nota-
tion as a product of its own by developing a stand-alone web-based collaborative
portal for the modelling and simulation of declarative workflows. The purpose of
the portal is to facilitate end-user discussions on how knowledge workers really
work, by enabling collaborative simulation of processes. In earlier work we re-
ported on the integration of DCR Graphs as a workflow execution formalism in
the existing Exformatics ECM products. In this paper we report on the advances
we have made over the last two years, we describe the new declarative process
modelling portal, discuss its features, describe the process of its development, re-
port on the findings of an initial evaluation of the usability of the tool, resulting
from a tutorial on declarative modelling with DCR Graphs that we organized at
last years BPM conference and present our plans for the future.
Keywords: Declarative Modelling, DCR Graphs, Web-Based Process Modelling, Col-
laborative Process Simulation, Process Flexibility, Knowledge Work
1 Introduction
Former secretary of labor in the Clinton administration, Robert Reich, argued that the
competitiveness of nations depends on the education and skills of it’s people and on the
infrastructure connecting people with one another [21] He segmented the work force
into three types of employees (1) routine production services, (2) ”in-person” services
and (3) symbolic analysts. Today we often refers to symbolic analysts as knowledge
workers [3, 5], the employees which contribute to most of the economic growth in de-
veloped economies. Therefore, supporting these knowledge workers and ensuring easy
2 Morten Marquard, Muhammad Shahzad, and Tijs Slaats
and smooth collaboration is important to compete globally. Enabling knowledge work-
ers to work smarter, rather than just harder, involves various IT infrastructures to ensure
communication and collaboration.
Traditional process initiatives, typically seen in the automobile indutry, focus on
what Reich refers to as routine production services. While increasing productivity and
cutting costs in primary industries is important to compete globally, it is even more
important to provide similar support for knowledge workers. Comparing tools and pro-
cesses used in the automobile industry with knowledge intensive industries will proba-
bly reveal a paradox. The infrastructure supporting routine production services is much
more advanced what is found supporting knowledge workers. Often knowledge work-
ers use email as their primary communication and collaboration tool, and studies shows
that knowledge workers on average spend 28 percentage of the time reading and re-
sponding to emails [13]. This is hardly efficient and makes it hard to compete in a
global economy, especially as knowledge workers in the developing countries has cheap
and easy access to secretary and other in-person services, and therefore simply will be
more serviced than knowledge workers in the developed countries. The engineers, doc-
tors, financial analysts in the western world simply need to work harder as they cannot
leverage cheap primary services. Therefore, providing infrastructure and techniques for
knowledge workers is critical to compete globally. McKinsey Global Institute estimate a
productivity gain of 20-25 percentage on average knowledge workers by using modern
social technologies.
In the effort to make knowledge workers more productive ideas and techniques
found in the routine production services, like the automobile industry, has been adopted
in the hope that it can support knowledge workers.
Over the last decades business has tried to model their business processes in order to
increase productivity and increase quality. Many techniques and notations exists, but the
Business Process Management Notation (BPMN) [19,27] is at the center of most initia-
tives. Processes are visualized using BPMN, Swimlanes diagrams, flow charts etc. As
most notations cannot describe the whole process, many descriptions are supplemented
with textual descriptions typically found in a Word document. Engaging end-users in
the process dialogue is hard as the notation is hard to understand. Many initiatives
within the process space has been trying to simplify the notation and methods. Lack
of end-user engagement often leads to process implementations not supporting the real
business need. We’ve also witnessed that people interpret processes in different ways
and that a large degree of implicit logic is assumed from the process. This makes it very
hard for others to read it and ultimately leads to failure of many process initiatives.
In order to achieve the overall business goals, increased productivity, higher quality
and employee enthusiasm, we propose a new model, collaborative process simulation.
To ensure employee engagement we propose to play the processes like a computer game
among co-workers. The team of co-workers can define the process using a declarative
process-model and immediately start process simulation in order to verify whether the
modelled processes meet the real world need.
During simulations end-users often asks questions like what happens if...,can we
do ... etc. These questions can be simulated in the tool and the process model adjusted
to meet the increased understanding of the process. Rather than looking for the right
Web-based Modelling and Collaborative Simulation of Declarative Processes 3
process, we empower employees to modify processes as core-business processes are
adjusted to meet changing business needs and requirements.
The industry standard BPMN notation is founded in the concept of flow; the idea
that to describe the behaviour of a process one needs to describe how control passes
(flows) between its activities. It has been observed however that the flow-based paradigm
is not ideal for knowledge-centred processes: knowledge workers deal with very diverse
problems which rarely ”fit the mould”, instead of being given predefined sequences of
tasks they often need to decide themselves what actions they should take based on their
expert knowledge. The IT systems that support them therefore need to be able of of-
fering a large degree of flexibility. [17, 22, 28] Such flexible processes exhibit a large
degree of variability, exhibited in flow-based models by many different possible paths
and states, which leads to so-called spaghetti models which are no longer understand-
able by users.
As an alternative a new declarative or constraint-based paradigm has been proposed
[7,10,18, 20,24]. The declarative paradigm is grounded in the idea that one should only
model the constraints (or business rules) of a process and then derive the possible paths
from the constraints. Any execution allowed by the constraints is a valid execution of
the process model, i.e. the workers are given maximal flexibility within the rules.
Exformatics is a Danish software developer providing Electronic Case Management
(ECM) solutions to knowledge-intensive businesses and organizations such as LEGO,
ISS, ministries and government institutions. Already from their creation they realised
the need for flexible processes support in their tools and in initial versions this was
solved by very rough process definitions that only grouped tasks within specific phases,
but otherwise left maximal flexibility to the users. They realised that this approach
lacked the ability of adding more meaningful rules and constraints to their processes
and through participation in a Danish knowledge network Infinit [1], which supports
interaction and dissemination between academia and industry they came into contact
with the Process Models group at the IT University of Copenhagen (ITU) which was
working on related issues and in particular has developed the DCR Graphs notation
[4, 7, 8, 16]. DCR Graphs is a declarative notation for flexible processes that sets itself
apart from other declarative notations such as Declare by utilizing only a very small
set of constraints, yet yielding high formal expressive power. In addition it offers a
straightforward run-time semantics formalized as transformations of the graphs, which
means that its visual representation can be used both at design-time to represent process
definitions and at run-time to represent process instances, in a similar manner as Petri-
nets.
Exformatics became very interested in the work on DCR Graphs and how they
could employ the notation to leverage flexible processes. Therefore they initiated a
close collaboration with the researchers at ITU, facilitated through various Danish fund-
ing mechanisms supporting university-industry collaboration [?]. Most notable among
these was a 3 years industrial PhD project, where Exformatics employed a PhD student
to do research on flexible process notations while at the same time being enrolled at
ITU.
During this project DCR Graphs were first integrated into the existing ECM tools as
a formalism for process-control by implementing a cloud-based process engine based
4 Morten Marquard, Muhammad Shahzad, and Tijs Slaats
on the DCR Graphs semantics [23]. At the same time the student also developed a pro-
totype tool for the graphical modelling of DCR Graphs which was well received within
the company and opened the road to further adaptation of DCR Graphs not only as an
internalized notation for standardized processes, but also as a graphical notation for de-
signing processes as a part of business consultancy services, in essence making DCR
Graphs a product of their own. This was achieved by developing a stand-alone web-
based collaborative portal for the modelling and simulation of declarative workflows,
aimed at facilitating end-user discussion on how knowledge workers really work, by
enabling collaborative simulation and run-time adaptation of processes.
This paper focuses on this new declarative process portal. We start by giving a short
introduction to DCR Graphs. We then give a detailed overview of the portal and its
features, discuss the development processes that led to its creation and discuss initial
efforts at evaluating the usability of the portal. We finalize by discussing our plans for
the future, both in terms of new features to the portal and new avenues of research.
1.1 Related Work
Several web-based commercial tools exist for the modelling of business processes; such
as Signavio, IBM Blueworks Live and Oracles Business Process Management Suite.
However, to our knowledge DCRGraphs.net is the first web-based modelling tool aimed
in particular its constraint-based notations for flexible processes.
Currently there is ongoing work on developing the Case Management Model And
Notation (CMMN) [18], which is a new standard notation developed by the Object
Management Group aimed in particular at case management and adding support for
flexible processes to BPMN. The work on CMMN is strongly inspired on the research
on the Guard-Stage-Milestone (GSM) model [11] developed at IBM Research, which
in turn is based on earlier work on artifact-centric business processes [2]. While GSM
is foremost a data-centric model it has some declarative influences as well, the main
elements of the notation are stages containing tasks, which are either active or inactive
based on guards defined on the stage. The acceptance criteria of a stage are modelled
through milestones, which can in turn be part of the guards of other stages. Compared
to GSM, DCR Graphs put more focus on the behaviour of tasks and events then on the
data of the process.
Declarative process languages came to prominence in the BPM community through
the development of the Declare notation [20, 25, 26]. Declare consists of a relatively
large set of constraints typically found in business processes, which are traditionally
mapped to Linear Temporal Logic(LTL) formulae, although other formalizations also
exist [12, 14, 15]. DCR Graphs differ from Declare in the number of symbols used in
the notation: Declare uses a large number of constraint templates, each with their own
symbol where as DCR Graphs is limited to 5 elementary relations. Also, because the
runtime semantics of DCR Graphs are given in terms of transformations on their mark-
ing, it is straightforward to visualize and reason about the simulation of DCR Graphs.
Exforamtics adopted DCR Graphs in part because of their close research collaboration
with ITU, giving them direct access to the researchers behind the notation, because they
preferred a more concise notation and put particular importance into reasoning about
Web-based Modelling and Collaborative Simulation of Declarative Processes 5
the runtime of processes through simulation. We are not aware of any published work
reporting on industrial use of Declare.
2 Hierarchical DCR Graphs
In this section we exemplify DCR Graphs and their semantics using an abstracted ver-
sion of the main case management process of the Exformatics ECM system.
Case
Case Manager
Archive Case
Create Meeting
Post To Activity
Stream
Create Document
%
Fig. 1. Root Case Management Process
In figure 1 we start with the root pro-
cess. The main building blocks of a
DCR Graph are the events (or activities),
drawn as a box with a bar on top. The
box contains the name of the activity and
the bar contains the roles that are able
of executing it. Our process has a sin-
gle role: the case manager. Activities can
be grouped together by nesting them un-
der a super-activity, in which case only
the atomic activities are executable. Such
grouping are a graphical shorthand for
applying constraints or properties to mul-
tiple activities at once: in our process
the super-activity Case having the role
Case Manager means that the case
manager is able of executing every atomic activity nested under it.Constraints or busi-
ness rules can be added to the model by adding one of five relations, drawn as direc-
tional arrows between activities. The root process contains a single relation, the exclu-
sion relation (%) from Archive Case to Case. The exclusion relation is used to
remove activities from the process, for example to close tasks that should no longer
be executable, or to model an exclusive choice between two activities. Because the
super-activity Case acts as a grouping the exclusion relation applies to all five of the
underlying activities. This means that after archiving the case no further actions can be
taken as it removes all activities from the process. Following the declarative paradigm
unconstrained activities can be done at any time and any number of times, therefore the
process supports many different runs: one could for example upload two documents,
create a meeting, upload another document and finally archive the case. Note that while
Archive Case closes the process by removing all activities, it is not required to
happen and the previous example run would also have been valid if it had not ended by
archiving the case.
Figure 2 shows the process for organizing a meeting. Similarly to the previous ex-
ample we use the exclusion relation to remove activities from the process when they are
no longer relevant: the activities Invite Participants,Change Date,Hold
Meeting and Cancel Meeting are grouped together and removed by either hold-
ing or cancelling the meeting. Only after holding the meeting is it possible to upload the
minutes of the meeting, this is modelled by the condition relation (→•) which states that
before Upload Minutes can be done we first need to have done Hold Meeting.
6 Morten Marquard, Muhammad Shahzad, and Tijs Slaats
Case Manager
Upload Minutes
Case Manager
Invite Participants
Hold Meeting
!
Change Date
Cancel Meeting
%
%
%
Fig. 2. Meeting Sub-process
Cancel Meeting excludes this activity since it does not make sense to upload min-
utes for a meeting that was cancelled. Finally the goal of the meeting sub-process is
that we either eventually hold the meeting or cancel it. This is modelled by making
Hold Meeting apending response, drawn by adding a blue exclamation mark to the
activity box. A pending response denotes that an activity should either happen or be
removed from the process before we can finish or close the process; in our example
either Hold Meeting needs to be done or Cancel Meeting needs to exclude it.
Checkin
Case Manager
Checkout
Case Manager
Download
Case Manager
%
%
+
+
Fig. 3. Document Handling Sub-process
Figure 3 shows the pro-
cess for managing a
document in the ECM.
To edit the document
a user needs to check
it out, the file is then
locked until it is checked
in again. This is mod-
elled using first the ex-
clusion relation to exclude each activity when they happen (meaning they can only be
done once at a time) and the two new include relations (+) between the two events.
The include relation is used to add removed activities back into the process, so in this
case, when Checkout happens, it removes itself and add Checkin to the process and
vice versa. Finally when a file is checked out we always want it to be checked in again
before the process can finish. We model this through the response relation (•→) which
is a dynamic version of the pending response that we introduced earlier. It denotes that
after Checkout is executed we require Checkin to be executed (or excluded) at least
once before we can close the process. It is always possible to download the document
through the unconstrained Download activity.
Finally we would like to tie all these process together into a single process describ-
ing the handling of a case in the ECM system. For this we use an extension called hier-
archical DCR (Hi-DCR) Graphs, which adds a notion of spawnable multi-instance sub-
processes. Figure 4 shows the case management process as a Hi-DCR Graph. The main
new concept are the two new sub-processes Meeting and Document Handling,
Web-based Modelling and Collaborative Simulation of Declarative Processes 7
Document Handling
Meeting
Case
|||
|||
Case Manager
Archive Case
Create Meeting
Post To Activity
Stream
Create Document
Case Manager
Download
Case Manager
Case Manager
Upload Minutes
Case Manager
Cancel Meeting
Checkout
Invite Participants
Hold Meeting
!
Change Date
%
%
+
%
%
%
%
%
Checkin
%
+
Fig. 4. Case Management Process
drawn as a box without a bar on top of it. These are essentially DCR Graphs inside the
root process that need to be initialized through the new spawn relation (), creating
a new copy of the sub-process for each time it is spawned. In the example the activ-
ity Create Meeting spawns a new copy of Meeting each time it is executed and
Create Document and Create Minutes create a new Document Handling
process each time they are executed.
When one has a relation between an event of a sub-process and an event of its par-
ent, the relation will apply to each instance of the event. For example a case can not be
archived while there are pending meetings; this is modelled by adding a condition from
Hold Meeting to Archive Case, meaning that while there is at least one Hold
Meeting-activity that has not yet been executed or excluded (by cancel meeting), it
is impossible to execute Archive Case. In a similar manner it should not be pos-
sible to archive the case while documents are checked out. We model this through the
milestone relation (), which blocks an event as long as some other event is pending.
Because of the response relation from Checkout to Checkin, there will be a pending
response on the latter whenever the file is checked out. The milestone from Checkin to
Archive Case ensures that as long as there is at least one instance of Checkin that
is pending, we can not archive the case. Finally, to archive the case, Archive Case
excludes all instances of all events in Document Handling, except for Download
as it should still be possible to download files. All instances of Upload Minutes are
also excluded, it is not necessary to exclude the other events in Meeting as this will
already have been done by holding or cancelling the meeting and unless one of these
has been executed the case can not be archived.
8 Morten Marquard, Muhammad Shahzad, and Tijs Slaats
The example DCR Graphs from figures 1, 2 and 3 are all available on DCRGraphs.net,
the Hi-DCR Graph in figure 4 was drawn using the development version of the portal
which is not yet ready for release.
3 The DCR Graphs Process Portal
dcrgraphs.net provide a solution for DCR Graphs management with collaboration for
registered users. It provides online web based tool for creating and editing DCR graphs,
which is also integrated to DCR processes engine services to simulate the graphs even
with friends or automated users invited to simulation. The simulation takes place in a
collaborative environment where each participant can contribute/interact.
dcrgraphs.net provides easy management for DCR graphs as well as sharing graphs
with friends and co-workers. Each graph has it’s own activity stream where discussions
about the graph can take place among people having access to the graph. A global
activity stream provides an overall discussion forum for people using dcrgraphs.net.
dcrgraphs.net supports several organisations, with the public organisation as the de-
fault organisation for users. Users can invite other users to use the portal simply be en-
tering their email addresses. Users can easily create new graphs here and manage their
existing graphs, they can also visit graphs shared to them or graphs which are public for
all. Activity stream provides a general mechanism to communicate and discuss/com-
ment either about a specific graph or in organization, well known from social medias
like LinkedIN and Facebook. Activity stream also support common social features like
most of sites have such as like, comment, picture upload.
Fig. 5. dcrgraphs.net frontpage
Users have their profiles and friends whom with they can simulate or share graphs,
and they can also invite anyone on the site by email. Portal also supports external logins
like LinkedIN and Facebook which can automatically retrieve the basic information of
Web-based Modelling and Collaborative Simulation of Declarative Processes 9
user’s account ready to be registered. The portal provides social aspects for users to
easily workaround there graphs in collaboration by providing such functionalities.
dcrgraphs.net has an online web based DCR Graph editor for users to easily create
and edit their graphs. It provide the GUI for the users to easily create graphs like any
other graphical editing tools. Activities and sub-processes can easily be added and var-
ious properties edited. The DCR Editor supports import of standard DCR XML [23]
and can export the Graph in XML, SVG, PNG formats to be easily used in other ap-
plications. Easy to use interface for creating and managing graph resources plus the
collaborative simulation sits within this editor.
Fig. 6. dcrgraphs.net editor
The Editor also supports revisions management and major versions for changes
made in normal state i.e. when not in simulation, which can be used to roll back the
changes to any previous version or preview changes. Besides the drag and drop creation
of the Elements in Graph a wizard is also provided for creating graph content much
rapidly.
As DCR graphs are often complex features to filter events based on role and level
has been added. Thereby the users can select to view specific aspects of the graph de-
pending on their need.
In order to facilitate discussion and collaboration among co-workers, easy access to
the activity stream of the graph has been provided from within the editor.
dcrgraphs.net support online collaborative simulations of processes which can be
shared among a group of people. A process-coordinator can invite friends to participate
in the simulation playing different roles in a particular process. The friends can choose
to join the simulation and participate on their own computer independently of phys-
ical location. Once the simulation starts the users plays different roles in the process
simulation.
Users can invite their friends or choose system automated users in the simulation
and assign roles to the invitees for them to take part in simulation. Simulations can be
paused/resumed, edited and restarted at any point during the runtime and a record of
10 Morten Marquard, Muhammad Shahzad, and Tijs Slaats
each successful completed simulation is kept in the system which can be later viewed.
Its called simulation replay in the Editor. During simulation available activities with
visual cue is also available for the contributes also with filtration of the activities and
with collaboration everyone has updates what’s happening in simulation.
During the simulation the activities executed are logged in the Execution Log, which
can be used later to replay the simulation. An activity (task) list displays activities which
are either required to complete the process or enabled to be executed. Whenever a activ-
ity is executed it is marked and the process continues. For every activity execution the
simulation invokes the process engine to modify and register the state of the process,
which typically results in new and/or changes to the existing activities.
Fig. 7. dcrgraphs.net simulation
The task list displays the current state of the process, i.e. whether the process can
be considered completed or more activities most be executed to complete the process.
The participants in the process are listed during simulation with the roles they play
in the current simulation.
The simulation can be paused by the process-coordinator and edited to support
changing requirements found during simulation. Later the simulation can be resumed
from the last point.
3.1 Key features
During development we’ve addressed set of key requirements which we’ve found im-
portant to implement in order to serve the overall business purpose, ensuring knowl-
edge workers participate in the description and further development of their business
processes.
Auto-user execution
In order to enable users to simulate processes alone, or give various roles to ”au-
tomatic users”, we’ve created two roles, a lazy and aggressive users. The lazy user
Web-based Modelling and Collaborative Simulation of Declarative Processes 11
executed only pending activities it is responsible for. The aggressive user executes
any enabled event it is responsible for, whether pending or not. The idea is to enable
simulation where the process editor only plays a one or few roles while assigning
other roles to automatic users. Ultimately, all roles could be assigned to automatic
users which would result in a ”video” playing the process.
Organisation
In order to support different organisations and universities, we’ve extended dcr-
graphs.net, which originally was intended as a public portal only, to support private
areas of users, graphs etc. This enables organisations and universities to setup their
own area, inviting their own employees and students, and define processes within a
closed community. As business processes are often sensitive we hope their model
will encourage organisations to setup private space within dcrgrahs.net. Users can
belong to several organisations, including a ”public” organisation, and can interact
with others.
Levels
We’ve added functionality to zoom-in and -out of the processes similar to zooming
in and out of Google Maps, to see more or less details of the process. Zooming out
enabled users to focus initially of the overall goal and purpose of the process, and
while zooming in more details are added which increase the details of the process.
Plugin framework
In order to enable 3rd party development, initially intended for student projects,
we’ve extended the portal and provide a plugin framework. We foresee a series
of development projects such as model checking (dead- and live-lock checking),
process visualisation (drawing ”happy paths”) etc. The details of the framework is
described on http://wiki.dcrgraphs.net.
Integrated to LinkedIN and Facebook
In order to smoothen user adoption we decided to leverage external portals like
Facebook and LinkedIN for providing user credentials and user profile informa-
tion. User registration can be accomplished simply by clicking on a LinkedIN or
Facebook icon. Then the user profile is retrieved from the 3rd party portal, typi-
cally after a login/approval process. When users return to dcrgraphs.net they can
re-login using their LinkedIN or Facebook credentials, therefore avoiding having
to remember several userid/passwords.
4 Development of the DCR Portal
Development of the portal has been carried out using the Scrum methodology, through
close co-operation between the development team at TEO International in Pakistan and
the design team at Exformatics in Denmark. We used short 2 to 3 week sprints with
frequent updates of the portal. Different technologies have been used to achieve the
solution so far, including: JavaScript with Rapha¨
el, jQuery, Microsoft .NET with MVC
4, REST services and the existing DCR Process Engine which was developed in a mix
of C# and F#. We are currently working on the 10th major version of the portal, which
will include support for sub-processes.
12 Morten Marquard, Muhammad Shahzad, and Tijs Slaats
The DCR Editor is purely a JavaScript application which utilizes REST services
developed in .NET to communicate with the web-server. The editor uses the Rapha¨
el
Library to visualize the graphs in SVG-format. Simulation of the graphs uses DCR pro-
cess engine services which have been upgraded over the time to support newly added
features. The wrapper of the Editor, which provides listing of the graphs, sharing, ac-
tivity stream and friendship functionalities is developed in MVC 4 .NET. Simulation in
Editor uses realtime notifications, updates with the help of Signal R which uses Web-
Sockets where possible. The editor utilizes caching techniques to minimize the requests
to server and improve the performance of the product.
5 Evaluation
At last years BPM conference we organized a tutorial on flexible business process mod-
elling using DCR Graphs where we first gave an introduction to DCR Graphs, exempli-
fying the notation through a demo of the process portal, then asked the audience to try
out the portal for themselves by doing a number of exercises and finally requested their
feedback in the form of a questionnaire.
Very Easy Easy Neutral Hard Very hard
Events (Activities) 4 5 2 1
Roles 5 5 1 1
Condition Relation 2 7 2
Response Relation 2 5 3 2
Inclusion Relation 1 2 4 5
Exclusion Relation 1 4 5 2
Pending response 1 5 4 1
Nesting 6 2 3
Table 1. Understandability of Concepts
Twelve people filled
out the questionnaire,
eleven identified as re-
searchers and one as
a practitioner. On aver-
age the attendants had
been active in the BPM
field for 5 years, rang-
ing from 6 months to
14 years. Five peo-
ple had previous experi-
ence with DCR Graphs,
whereas 9 people had previous experience with declarative modelling notations in gen-
eral.
Very Easy Easy Neutral Hard Very hard
Events(Activities) 3 7 2
Roles 3 4 4 1
Condition Relation 6 3 3
Response Relation 8 2 2
Inclusion Relation 5 5 2
Exclusion Relation 6 5 1
Pending response 7 3 2
Nesting 1 6 2 3
Table 2. Understandability of Notation
In the first set of
questions we asked the
audience if they found
the underlying concepts
of DCR Graphs hard
or easy to understand.
We specifically asked
them to only comment
on the understandabil-
ity of the concept and
not the graphical nota-
tion used. The results
are shown in table 1, perhaps not surprisingly most found events (9) and roles (10)
easy or very easy to understand. Of the relations the participants found the condition
Web-based Modelling and Collaborative Simulation of Declarative Processes 13
the easiest to understand, 9 people scored it easy or very easy, followed by the response
(7), exclusion (5) and finally inclusion (3). It is noteworthy that despite being closely
related the audience found the inclusion relation significantly harder to understand then
the exclusion relation, we conjecture that the exclusion more closely matches familiar
concepts such as mutual exclusion, whereas the inclusion relation was more novel to
the audience.
In the next set of questions we asked the audience to rate the understandability of
the graphical notation. Table 2 shows the results, overall events and roles were found
to be easy or very easy to understand, whereas the users were more neutral about the
notation of the relations. The condition and response scored higher then the inclusion
and exclusion, we conjecture that this may be because many attendants were already
familiar with the precedence and response constraints in Declare.
Very Easy Easy Neutral Hard Very hard
Modelling Screen 11 1
Adding Friends 1 6 2 1
Individual simulation 11 1
Collaborative simulation 6 2 1
Table 3. Usability of the Tool
In our final set
of questions, whose
results are shown in
table 3, we focussed
on the tool itself
and asked the par-
ticipants to rate the
usability of its vari-
ous components. Both the modelling and simulation functionality scored high, with 11
people finding them easy to use and the final participant being neutral on their usability.
Adding connections scored a little less well, with 7 people finding this part of the tool
easy or very easy to use. Collaborative simulation scored only slightly lower, with 6
participants finding it easy to use, out of 9 people answering this particular question.
The portal has been used for teaching a process modelling course at ITU, with 75
users signed up. While we have not organized a similar questionnaire for the students,
overall the experience has been positive and no major issues were encountered in using
the portal. More recently we also initiated a collaboration with the Federal University
of the State of Rio de Janeiro (UNIRIO) where they use the portal for teaching a similar
course.
6 Future Work
We recognize that it is too early to draw strong conclusions from the questionnaire
based on a small number of academic participants. In the near future we plan to run
multiple industrial workshops following the same structure as the tutorial, both as a
method for raising awareness of the portal and attracting potential new customers and
as a way to receive additional feedback from actual practitioners which will allow us to
make stronger claims regarding its usability.
Further initiatives focus on:
Sub-processes
The ability to split processes into sub-processes which can be instantiated and exe-
cuted separately from the parent process. Experience from existing examples [6],
shows that sub-processes are important to model processes in an easy to describe
14 Morten Marquard, Muhammad Shahzad, and Tijs Slaats
and understandable model. Work on this item is already ongoing and we expect
to have sub-processes included in the live version of the portal by the summer of
2015.
Verification
Various mechanism has been developed to analyze DCR graphs for dead- and live-
lock. Adding such analysis to the tool is important, both when designing graphs
but also when adjusting DCR graphs running within a case management tool, like
Exformatics ECM. When the user adds new activities and/or constraints it might
lead to live- and dead-locks.
Gamification
We introduced Keith Swensson, author of various books on Adaptive Case Manage-
ment, to the DCR graphs at the Adaptive Case Management conference in Ulm in
august 2014, as well as at the Business Process Management (BPM) conference in
Eindhoven in September 2014. Keith suggested that we added more gamification
features to the portal and allow participants to earn points and credits in various
ways.
Collaborative Editing
Collaborative editing, a brilliant feature of Google Docs, has been suggested by
several users of dcrgraphs.net. As this goes well along with collaborative simulation
this is a feature we’re likely to add in the near future.
Time
Models for adding time to DCR graphs has been investigated [9] and will be added
to the semantics of the process engine and to the semantics of DCR graphs. When
simulating time we need a way to scale time, so, say, a day take 5 minutes, so
simulation can capture such events. We probably also need to have a global clock
which can be stopped by the process facilitator.
Probability and Cost of Events
DCR graphs allow many path from source to destination just like a GPS can provide
many different paths. The GPS knows which roads are highways and which are
small roads, and tries to find an optimal way with the given conditions. Modern
GPS system link up with live feed about average speed on various roads and include
such information into the calculation. An accident on the highway will guide drivers
onto secondary roads as this is faster in the given situation. We plan to mark DCR
graphs with similar information, initially statistics added to the graph like the type
of road added to the GPS. Long term we plan to capture statistics from process
mining to provide improved accuracy.
Process Mining
Process mining can be used to analyse logs obtained from various processes. While
analysis techniques will be fairly straightforward to integrate into the tools, new
research will need to be done on the discovery of DCR Graph based models.
Hybrid Techniques
Despite the simplicity of DCR graphs we often meet the argument that business
people are more familiar and accustomed with BPMN diagrams, swim-lanes and
flow-charts. As we argued in the introduction this familiarity is often not really the
case, as business process drawings are interpreted very different by business people.
Web-based Modelling and Collaborative Simulation of Declarative Processes 15
Nevertheless we feel that adding well known graphical models like swim-lanes will
ease adoption of DCR graphs within the business community.
7 Conclusion
Over the last decade, social technologies has experienced significant adoption and growth,
most siginificantly Facebook. Traditional social technologies enables people to commu-
nicate, but often lack ability to coordinate the various activities people participate in.
We believe DCR graphs can play a role in social technologies for such coordination.
Whether the coordination is developed manually by drawing DCR graphs or extracted
from various logs using process mining techniques we don’t know, and maybe the two
models can be combined.
Exformatics will continue to invest in declarative process technologies, and has en-
tered into a new collaboration initiative with IT University of Copenhagen and is partly
funding a Post.Doc. over the next two years. Further, Exformatics has hired Tijs Slaats,
PhD. student for three years, in a part time position also jointly with ITU. Further,
Exformatics participate in various industry initiatives, currently with a major Danish
financial institution which has worked with process modelling for many years but are
looking into declarative models to provide a more flexible process model.
References
1. Innovations netværk for it. http://www.infinit.dk/.
2. Kamal Bhattacharya, Cagdas Gerede, Richard Hull, Rong Liu, and Jianwen Su. Towards
formal analysis of artifact-centric business process models. In In preparation, pages 288–
304, 2007.
3. Thomas H Davenport, Sirkka L Jarvenpaa, and Michael C Beers. Improving knowledge
work processes. Sloan management review, 1996.
4. Søren Debois, Thomas Hildebrandt, and Tijs Slaats. Hierarchical declarative modelling with
refinement and sub-processes. In Business Process Management, volume 8659 of Lecture
Notes in Computer Science, pages 18–33. Springer International Publishing, 2014.
5. Peter F. Drucker. Management Challenges for the 21st Century. HarperBusiness, 2001.
6. Søren Dubois, Thomas Hildebrandt, Morten Marquard, and Tijs and Slaats. A case for
declarative process modelling: Agile development of a grant application system. 2014.
7. Thomas Hildebrandt and Raghava Rao Mukkamala. Declarative event-based workflow as
distributed dynamic condition response graphs. In Post-proceedings of PLACES 2010, 2010.
8. Thomas Hildebrandt, Raghava Rao Mukkamala, and Tijs Slaats. Nested dynamic condition
response graphs. In Proceedings of Fundamentals of Software Engineering (FSEN), April
2011.
9. Thomas Hildebrandt, Raghava Rao Mukkamala, Tijs Slaats, and Francesco Zanitti. Contracts
for cross-organizational workflows as timed dynamic condition response graphs. Journal of
Logic and Algebraic Programming (JLAP), may 2013.
10. Richard Hull, Elio Damaggio, Riccardo De Masellis, Fabiana Fournier, Manmohan Gupta,
Fenno Terry Heath, III, Stacy Hobson, Mark Linehan, Sridhar Maradugu, Anil Nigam, Pi-
wadee Noi Sukaviriya, and Roman Vaculin. Business artifacts with guard-stage-milestone
lifecycles: managing artifact interactions with conditions and events. In Proc. of DEBS ’11,
pages 51–62, New York, NY, USA, 2011. ACM.
16 Morten Marquard, Muhammad Shahzad, and Tijs Slaats
11. Richard Hull, Elio Damaggio, Fabiana Fournier, Manmohan Gupta, Fenno Terry Heath, III,
Stacy Hobson, Mark Linehan, Sridhar Maradugu, Anil Nigam, Piyawadee Sukaviriya, and
Roman Vaculin. Introducing the guard-stage-milestone approach for specifying business
entity lifecycles. In Proc. of WS-FM’10, pages 1–24, Berlin, Heidelberg, 2011. Springer-
Verlag.
12. Fabrizio Maria Maggi, Marco Montali, Michael Westergaard, and Wil M. P. van der Aalst.
Monitoring business constraints with linear temporal logic: An approach based on colored
automata. In Business Process Management (BPM) 2011, volume 6896 of Lecture Notes in
Computer Science, pages 32–147, 2011.
13. Jacques Bughin Richard Dobbs Charles Roxburgh Hugo Sarrazin Geoffrey Sands
Michael Chui, James Manyika and Magdalena Westergren. The social economy: Unlock-
ing value and productivity through social technologies. 2012.
14. Marco Montali. Specification and Verification of Declarative Open Interaction Models: a
Logic-Based Approach, volume 56 of Lecture Notes in Business Information Processing.
Springer, 2010.
15. Marco Montali, Maja Pesic, Wil MP van der Aalst, Federico Chesani, Paola Mello, and
Sergio Storari. Declarative specification and verification of service choreographiess. ACM
Transactions on the Web (TWEB), 4(1):3, 2010.
16. Raghava Rao Mukkamala. A Formal Model For Declarative Workflows - Dynamic Condition
Response Graphs. PhD thesis, IT University of Copenhagen, March 2012.
17. N. A. Mulyar, M. H. Schonenberg, Mans, and van der Aalst. Towards a Taxonomy of Process
Flexibility (Extended Version). 2007.
18. Object Management Group. Case Management Model and Notation, version 1.0. Webpage,
may 2014. http://www.omg.org/spec/CMMN/1.0/PDF.
19. Object Management Group BPMN Technical Committee. Business Process Model and No-
tation, version 2.0. http://www.omg.org/spec/BPMN/2.0/PDF.
20. M. Pesic, H. Schonenberg, and W.M.P. van der Aalst. DECLARE: Full Support for Loosely-
Structured Processes. In Proceedings of the 11th IEEE International Enterprise Distributed
Object Computing Conference, pages 287–. IEEE Computer Society, 2007.
21. Robert B. Reich. The Work of Nations: Preparing Ourselves for 21st Century Capitalism.
Vintage Books, 1992.
22. Manfred Reichert and Barbara Weber. Enabling Flexibility in Process-Aware Information
Systems: Challenges, Methods, Technologies. Springer, Berlin-Heidelberg, 2012.
23. Tijs Slaats, Raghava Rao Mukkamala, Thomas Hildebrandt, and Morten Marquard. Exfor-
matics declarative case management workflows as dcr graphs. In Business Process Manage-
ment, volume 8094 of Lecture Notes in Computer Science, pages 339–354. Springer Berlin
Heidelberg, 2013.
24. W. M. P. van der Aalst, M. Pesic, and H. Schonenberg. Declarative workflows: Balancing
between flexibility and support. Computer Science - R&D, 23(2):99–113, 2009.
25. Wil van der Aalst, Maja Pesic, Helen Schonenberg, Michael Westergaard, and Fabrizio M.
Maggi. Declare. Webpage, 2010. http://www.win.tue.nl/declare/.
26. Wil M.P van der Aalst and Maja Pesic. DecSerFlow: Towards a truly declarative service
flow language. In Proceedings of Web Services and Formal Methods (WS-FM 2006), volume
4184 of LNCS, pages 1–23. Springer Verlag, 2006.
27. Hagen V ¨
olzer. An overview of bpmn 2.0 and its potential use. In Business Process Modeling
Notation, volume 67 of Lecture Notes in Business Information Processing, pages 14–15.
Springer Berlin Heidelberg, 2010.
28. Barbara Weber, Manfred Reichert, and Stefanie Rinderle-Ma. Change patterns and change
support features - enhancing flexibility in process-aware information systems. Data and
Knowledge Engineering, 66(3):438–466, September 2008.
... To combat this challenge, it has been proposed to use declarative notations [42,55,68] for formalizing the rules governing knowledge work processes. Indeed, the use of declarative models can be seen as an alternative line of work in the area of computer supported work processes that can be traced back to the early rule-based expert systems [35]. ...
... DCR Graphs is a declarative notation developed during the last decade with the aim of providing flexible process support. The notation is supported by a commercial design tool [55] and a stand-alone rule engine, which has can be embedded in third-party case management tools. In particular, the engine has successfully been embedded in two commercial case management solutions developed by the company KMD, which is part of the multi-national company NEC. ...
... To support the local development and maintenance of the declarative DCR models, several modeling tools have been developed [19,23,55], supported by formal understandability studies [5][6][7]. Along with the tools, a methodology for modeling with DCR has been developed, advocating an iterative and incremental, scenario-driven approach with three main tasks. ...
Preprint
Declarative process modeling formalisms - which capture high-level process constraints - have seen growing interest, especially for modeling flexible processes. This paper presents DisCoveR, an extremely efficient and accurate declarative miner for learning Dynamic Condition Response (DCR) Graphs from event logs. We precisely formalize the algorithm, describe a highly efficient bit vector implementation and rigorously evaluate performance against two other declarative miners, representing the state-of-the-art in Declare and DCR Graphs mining. DisCoveR outperforms each of these w.r.t. a binary classification task, achieving an average accuracy of 96.2% in the Process Discovery Contest 2019. Due to its linear time complexity, DisCoveR also achieves run-times 1-2 orders of magnitude below its declarative counterparts. Finally, we show how the miner has been integrated in a state-of-the-art declarative process modeling framework as a model recommendation tool, discuss how discovery can play an integral part of the modeling task and report on how the integration has improved the modeling experience of end-users.
... With the increased need to accommodate flexible, knowledge-intensive processes, notations focusing on essential rules, rather than detailed procedures have seen increased attention from researchers [28,55,65,70,75]. This approach is often characterized as declarative and juxtaposed with imperative notations like BPMN [39,52,66]. ...
... To support the local development and maintenance of the declarative DCR models, several modeling tools have been developed [16,20,52], supported by formal understandability studies [1][2][3]. Along with the tools, a methodology for modeling with DCR has been developed, advocating an iterative and incremental, scenario-driven approach with three main tasks. ...
... This allows modelers to straightforwardly reason about the execution semantics of a model by simulating it and observing the changes to the markings as events are executed. [52] Since their inception, DCR Graphs have been extended with nesting [37], time [38], data [16,56,78], and hierarchy [22]. ...
Article
Full-text available
Declarative process modeling formalisms—which capture high-level process constraints—have seen growing interest, especially for modeling flexible processes. This paper presents DisCoveR, an efficient and accurate declarative miner for learning Dynamic Condition Response (DCR) Graphs from event logs. We present a precise formalization of the algorithm, describe a highly efficient bit vector implementation and present a preliminary evaluation against five other miners, representing the state-of-the-art in declarative and imperative mining. DisCoveR performs competitively with each of these w.r.t. a fully automated binary classification task, achieving an average accuracy of 96.1% in the Process Discovery Contest 2019 (Results are available at https://icpmconference.org/2019/process-discovery-contest). We appeal to computational learning theory to gain insight into its performance as a classifier. Due to its linear time complexity, DisCoveR also achieves much faster run times than other declarative miners. Finally, we show how the miner has been integrated in a state-of-the-art declarative process modeling framework as a model recommendation tool and discuss how discovery can play an integral part of the modeling task and report on how the integration has improved the modeling experience of end-users.
... While initially an academic oddity, declarative notations have recently seen more commercial success. In particular in Denmark, the Dynamic Condition Response (DCR) Graphs [14,18,19] notation has been adopted by industry [20], culminating in their integration in a case management solution that is used by 70% of Danish central government institutions, including the police, military, tax authorities and largest public universities. 1 Not all processes are either completely strict or flexible, for example, a patient process in a hospital may consist of both a flexible treatment part and a much more structured and strict data-processing part. ...
... This allows modellers to straightforwardly reason about the execution semantics of a model by simulating it and observ- ing the changes to the markings as events are executed. [20] On the other hand, this does have the disadvantage that the orthogonality of the Declare notation, where one is always certain that adding new constraints will never allow new behaviour in the model (because at the top-level, it is a conjunction of all constraints), does not apply to DCR Graphs. In particular, it can be quite counter-intuitive for new users that the introduction of exclusion relations can allow for new behaviour through their interactions with condition relations. ...
... Even less work has been done towards comparing the relative understandability of different declarative notations in comparison with each other. The closest come two studies reporting on workshops with industrial practitioners [12] and academics [20], each including a somewhat informal questionnaire. The workshop with industrial practitioners involved modelling exercises using both Declare and DCR Graphs and showed a comparable response to each, with the practitioners finding the notations challenging to use, but nonetheless saw value for their use in their own work practices, especially if combined with existing imperative notations. ...
Article
Full-text available
Knowledge-intensive processes, such as those encountered in health care, finance and government, tend to allow a large degree of flexibility: there are many possible solutions towards a goal, and it is left to the expertise of knowledge workers to find the one most suitable for the particular case at hand. As a result, such processes usually exhibit more varied behaviour than traditional production processes. This poses a challenge for process discovery algorithms that return imperative, flow-based, models. The models tend to become highly complex when representing many alternative paths, and therefore, the miners need to either sacrifice on simplicity, fitness, or precision. It has been proposed that one should discover the constraints of the process instead, based on the assumption that such a constraint-based, declarative process model can describe highly varied behaviour more concisely. More recently, it has been observed that many processes do not neatly fall in one category or the other; instead, they contain both flexible and rigid parts. In such cases, it may be helpful to identify these parts and mine constraints for some and flow for others, resulting in a hybrid model. In this paper, we provide an overview of recent advances in both declarative and hybrid process discovery, discuss a number of open challenges that still remain, and propose directions for future research.
... Enactment of declarative process models (Carvalho et al., 2013;Goedertier et al., 2008;Schönig et al., 2018;Marquard et al., 2015;Ackermann et al., 2018;Montali et al., 2013a;Pesic, 2008;Westergaard and Maggi, 2012;Mertens et al., 2019) The proposed approach allows for an efficient replanning during process enactment while guiding users based on the optimization of a given objective function. ...
... Similarly, DPIL Navigator enables the enactment of textual DPIL models, including the support of certain workflow resource and data patterns. A more human-understandable approach is presented by Marquard et al. (2015), which proposes a process engine for DCR Graphs. This engine supports the enactment of declarative models in a collaborative environment. ...
Article
This work has been motivated by the needs we discovered when analyzing real-world processes from the healthcare domain that have revealed high flexibility demands and complex temporal constraints. When trying to model these processes with existing languages, we learned that none of the latter was able to fully address these needs. This motivated us to design TConDec-R, a declarative process modeling language enabling the specification of complex temporal constraints. Enacting business processes based on declarative process models, however, introduces a high complexity due to the required optimization of objective functions, the handling of various temporal constraints, the concurrent execution of multiple process instances, the management of cross-instance constraints, and complex resource allocations. Consequently, advanced user support through optimized schedules is required when executing the instances of such models. In previous work, we suggested a method for generating an optimized enactment plan for a given set of process instances created from a TConDec-R model. However, this approach was not applicable to scenarios with uncertain demands in which the enactment of newly created process instances starts continuously over time, as in the considered healthcare scenarios. Here, the process instances to be planned within a specific timeframe cannot be considered in isolation from the ones planned for future timeframes. To be able to support such scenarios, this article significantly extends our previous work by generating optimized enactment plans under a rolling planning horizon. We evaluate the approach by applying it to a particularly challenging healthcare process scenario, i.e., the diagnostic procedures required for treating patients with ovarian carcinoma in a Woman Hospital. The application of the approach to this sophisticated scenario allows avoiding constraint violations and effectively managing shared resources, which contributes to reduce the length of patient stays in the hospital.
... DCR Graphs are a well-known declarative process modeling language based on directed graphs whose nodes represent events and whose edges capture the relationship between them [17]. DCR Graphs benefit from the support of the DCR Graphs Portal [18], a researchbased commercial tool supporting the design, enactment and analysis of DCR Graphs. The DCR Graphs Portal features a graphical web-based editor, a textual process specification that can be visually linked to parts of the process model and an interactive simulator that can be directly enabled from the editor and visualizes the process execution directly on the DCR Graph. ...
... The comprehension tasks were designed and displayed on the DCR Graphs Portal [18] either in English or in Danish depending on the individual preference of the participants. Visually, the content of the web-based user interface was organized into the four main areas outlined in Fig. 1 ...
Article
Full-text available
Process design artifacts have been increasingly used to guide the modeling of business processes. To support users in designing and understanding process models, different process artifacts have been combined in several ways leading to the emergence of the so-called “hybrid process artifacts”. While many hybrid artifacts have been proposed in the literature, little is known about how they can actually support users in practice. To address this gap, this work investigates the way users engage with hybrid process artifacts during comprehension tasks. In particular, we focus on a hybrid representation of DCR Graphs (DCR-HR) combining a process model, textual annotations and an interactive simulation. Following a qualitative approach, we conduct a multi-granular analysis exploiting process mining, eye-tracking techniques, and verbal data analysis to scrutinize the reading patterns and the strategies adopted by users when being confronted with DCR-HR. The findings of the coarse-grained analysis provide important insights about the behavior of domain experts and IT specialists and show how user’s background and task type change the use of hybrid process artifacts. As for the fine-grained analysis, user’s behavior was classified into goal-directed and exploratory and different strategies of using the interactive simulation were identified. In addition, a progressive switch from an exploratory behavior to a goal-directed behavior was observed. These insights pave the way for an improved development of hybrid process artifacts and delineate several directions for future work.
... The work on the DCR notation and associated tool support was followed up in the PhD thesis of Slaats [21]. The notation is by now supported by mature modelling and simulation tools [2,13] and has been embedded in commercial case management systems widely used in Denmark by both central government institutions such as universities and local government institutions such as municipalities [3]. ...
Chapter
We provide a formal semantics for timed DCR graphs with decision events and data, and show how the model supports merging declarative decision modelling and process modelling in one notation. We exemplify the model by an expense report process and discuss how decisions may be presented using DMN jointly with the DCR process notation. If data domains are bounded, models are finite state and allow formal verification of both safety and liveness properties. The formal model is fully implemented in commercial process design and workflow management systems provided by DCR Solutions and also freely available for academic use.
Chapter
This paper addresses the open technical problems of evolving executable, event-based process models by refinement, that is, by iteratively expanding a model until it has the required level of detail. Such iterative development is helpful because of the expectation that the next-step model is semantically compatible with the previous one, only with more detail. We provide in this paper a formal notion of refinement of single atomic actions (events) into entire subprocesses, and a theoretical framework for providing guarantees that such a next-step model is formally a refinement of the previous one. Our work is set within the declarative, event-based process modelling language of timed Dynamic Condition Response (DCR) graphs, which can express timed constraints (conditions with delay and obligations with deadlines) between events, liveness, safety, and concurrency. Concretely, we extend DCR graph syntax and semantics with a notion of subprocess, provide examples of its use, and give sound approximations of situations where replacing an event with a subprocess formally is a refinement of the original process.
Article
In order to manage evolving organisational practice and maintain compliance with changes in policies and regulations, businesses must be capable of dynamically reconfiguring their business processes. However, such dynamic reconfiguration is a complex, human-intensive and error prone task. Not only must new business process rules be devised but also, crucially, the transition between the old and new rules must be managed. In this paper we present a fully automated technique based on formal specifications and discrete event controller synthesis to produce correct-by-construction reconfiguration strategies. These strategies satisfy user-specified transition requirements, be they domain independent - such as delayed and immediate change - or domain specific. To achieve this, we provide a discrete-event control theoretic approach to operationalise declarative business process specifications, and show how this can be extended to resolve reconfiguration problems. In this way, given the old and the new business process rules described as Dynamic Condition Response Graphs, and given the transition requirements described with linear temporal logic, the technique produces a control strategy that guides the organisation through a business process reconfiguration ensuring that all transition requirements and process rules are satisfied. The technique outputs a reconfiguration DCR whose traces reproduce the controller’s reconfiguration strategy. We illustrate and validate the approach using realistic cases and examples from the BPM Academic Initiative.
Chapter
Process models provide a blueprint for process execution and an indispensable tool for process management. Bearing in mind their trending use for requirement elicitation, communication and improvement of business processes, the need for understandable process models becomes a must. In this paper, we propose a research model to investigate the impact of modularization on the understandability of declarative process models. We design a controlled experiment supported by eye-tracking, electroencephalography (EEG) and galvanic skin response (GSR) to appraise the understandability of hierarchical process models through measures such as comprehension accuracy, response time, attention, cognitive load and cognitive integration.
Chapter
Process modelling notations fall in two broad categories: declarative notations, which specify the rules governing a process; and imperative notations, which specify the flows admitted by a process. We outline an empirical approach to addressing the question of whether certain process logs are better suited for mining to imperative than declarative notations. We plan to attack this question by applying a flagship imperative and declarative miner to a standard collection of process logs, then evaluate the quality of the output models w.r.t. the standard model metrics of precision and generalisation. This approach requires perfect fitness of the output model, which substantially narrows the field of available miners; possible candidates include Inductive Miner and MINERful. With the metrics in hand, we propose to statistically evaluate the hypotheses that (1) one miner consistently outperforms the other on one of the metrics, and (2) there exist subsets of logs more suitable for imperative respectively declarative mining.
Conference Paper
Full-text available
We report on a recent industrial project carried out by Exformatics A/S in which the company used the declarative DCR Graphs notation to model and implement the grant application process of a Danish foundation. We present the process and discuss the advantages of the approach and challenges faced both while modelling and implementing the process. Finally, we discuss current work on extensions to the DCR Graphs notation aiming to address the challenges raised by the case study and to support the declarative, agile approach.
Conference Paper
Full-text available
We present a new declarative model with composition and hierarchical definition of processes, featuring (a) incremental refinement, (b) adaptation of processes, and (c) dynamic creation of sub-processes. The approach is motivated and exemplified by a recent case manage-ment solution delivered by our industry partner Exformatics A/S. The approach is achieved by extending the Dynamic Condition Response (DCR) graph model with interfaces and composition along those inter-faces. Both refinement and sub-processes are then constructed in terms of that composition. Sub-processes take the form of hierarchical (complex) events, which dynamically instantiate sub-processes. The extensions are realised and supported by a prototype simulation tool.
Thesis
Full-text available
Current business process technology is pretty good in supporting well-structured business processes and aim at achieving a fixed goal by carrying out an exact set of operations. In contrast, those exact operations needed to fulfill a business pro- cess/workflow may not be always possible to foresee in highly complex and dynamic environments like healthcare and case management sectors, where the processes ex- hibit a lot of uncertainty and unexpected behavior and thereby require high degree of flexibility. Several research groups have suggested declarative models as a good approach to handle such ad-hoc nature by describing control flow implicitly and there by offering greater flexibility to the end uses. The first contribution of this PhD thesis is to formalize the core primitives of a declarative workflow management system employed by our industrial partner Result- maker and further develop it as a general formal model for specification and execution of declarative, event-based business processes, as a generalization of a concurrency model, the classic event structures. The model allows for an intuitive operational semantics and mapping of execution state by a notion of markings of the graphs and we have proved that it is sufficiently expressive to model ω-regular languages for infinite runs. The model has been extended with nested sub-graphs to express hierarchy, multi-instance sub processes to model replicated behavior and support for data. The second contribution of the thesis is to provide a formal technique for safe distribution of collaborative, cross-organizational workflows declaratively modeled in DCR graphs based on a notion of projections. The generality of the distribution technique allows for fine tuned projections based on few selected events/labels, at the same time keeping the declarative nature of the projected graphs (which are also DCR graphs). We have also provided semantics for distributed executions based on synchronous communication among network of projected graphs and proved that global and distributed executions are equivalent. Further, to support modeling of processes using DCR Graphs and to make the formal model available to a wider audience, we have developed prototype tools for specification and a workflow engine for the execution of DCR Graphs. We have also developed tools interfacing SPIN model checker to formally verify safety and liveness properties on the DCR Graphs. Case studies from healthcare and case management domains have been modeled in DCR Graphs to show that our formal model is suitable for modeling the workflows from those dynamic sectors. This PhD project is funded by the Danish Strategic Research Council through the Trustworthy Pervasive Healthcare Services project (www.trustcare.eu).
Conference Paper
Full-text available
Petri nets efficiently model both data- and control-flow. Control-flow is either modeled explicitly as flow of a specific kind of data, or implicit based on the data-flow. Explicit modeling of control-flow is useful for well-known and highly structured processes, but may make modeling of abstract features of models, or processes which are highly dynamic, overly complex. Declarative modeling, such as is supported by Declare and DCR graphs, focus on control-flow, but does not specify it explicitly; instead specifications come in the form of constraints on the order or appearance of tasks. In this paper we propose a combination of the two, using colored Petri nets instead of plain Petri nets to provide full data support. The combined approach makes it possible to add a focus on data to declarative languages, and to remove focus from the explicit control-flow from Petri nets for dynamic or abstract processes. In addition to enriching both procedural processes in the form of Petri nets and declarative processes, we also support a flow from modeling only abstract data- and control-flow of a model towards a more explicit control-flow model if so desired. We define our combined approach, and provide considerations necessary for enactment. Our approach has been implemented in CPN Tools 4.
Conference Paper
Full-text available
Declarative modeling has attracted much attention over the last years, resulting in the development of several academic declarative modeling techniques and tools. The absence of empirical evaluations on their use and usefulness, however, raises the question whether practitioners are attracted to using those techniques. In this paper, we present a study on what practitioners think of declarative modeling. We show that the practitioners we involved in this study are receptive to the idea of a hybrid approach combining imperative and declarative techniques, rather than making a full shift from the imperative to the declarative paradigm. Moreover, we report on requirements, use cases, limitations, and tool support of such a hybrid approach. Based on the gained insight, we propose a research agenda for the development of this novel modeling approach.
Conference Paper
Full-text available
Declarative workflow languages have been a growing research subject over the past ten years, but applications of the declarative approach in industry are still uncommon. Over the past two years Exformatics A/S, a Danish provider of Electronic Case Management systems, has been cooperating with researchers at IT University of Copenhagen (ITU) to create tools for the declarative workflow language Dynamic Condition Response Graphs (DCR Graphs) and incorporate them into their products and in teaching at ITU. In this paper we give a status report over the work. We start with an informal introduction to DCR Graphs. We then show how DCR Graphs are being used by Exformatics to model workflows through a case study of an invoice workflow. Finally we give an overview of the tools that have been developed by Exformatics to support working with DCR Graphs and evaluate their use in capturing requirements of workflows and in a bachelor level course at ITU.
Conference Paper
The declarative-procedural dichotomy is highly relevant when choos-ing the most suitable process modeling language to represent a discovered pro-cess. Less-structured processes with a high level of variability can be described in a more compact way using a declarative language. By contrast, procedural pro-cess modeling languages seem more suitable to describe structured and stable processes. However, in various cases, a process may incorporate parts that are better captured in a declarative fashion, while other parts are more suitable to be described procedurally. In this paper, we present a technique for discovering from an event log a so-called hybrid process model. A hybrid process model is hierarchical, where each of its sub-processes may be specified in a declarative or procedural fashion. We have implemented the proposed approach as a plug-in of the ProM platform. To evaluate the approach, we used our plug-in to mine a real-life log from a financial context.