Flexibility in Process-aware Information Systems (ProFlex) Workshop Report
Barbara Weber
University of Innsbruck
A-6020 Innsbruck
Austria
Jan Mendling
WU Vienna
A-1090 Vienna
Austria
Manfred Reichert
University of Twente
NL-7500 AE Enschede
The Netherlands
Abstract
The goal of the ProFlex workshop is to bring together
practitioners and researchers from different communities
who share an interest in flexibility of process-aware infor-
mation systems and team support. This workshop report
gives an overview of the presented papers that address var-
ious flexibility issues of process-aware information systems.
1. Introduction
The economic success of an enterprise more and
more depends on its ability to flexibly and quickly re-
act to changes, e.g., in its market, technology, or legal
environment. For this reason companies are develop-
ing a growing interest in new concepts, systems, and
solutions which help them to flexibly align their organi-
zation and their business processes to new requirements
and to optimize interactions with customers and busi-
ness partners. While there has been major progress in
disciplines that are interested in structured and unstruc-
tured business processes, the agile enterprise is still a
vision. Agility in this context refers to the ability of an en-
terprise to rapidly set up new business processes and
projects in order to quickly adapt to changes in the en-
vironment and aligning its existing information systems
to support them. In order to meet its business objec-
tives, the agile enterprise continuously re-aligns its busi-
ness processes as well as the interactions with its partners
and customers to meet the current requirements.
The goal of the ProFlex workshop is to bring together
practitioners and researchers from different communi-
ties such as business process management (BPM), software
engineering, artificial intelligence, computer supported co-
operative work (CSCW), and Groupware who share an
interest in flexibility of process-aware information sys-
tems and team support. The workshop aims at discussing
the current state of ongoing research, sharing practical ex-
periences, and demonstrating advanced research proto-
types. Workshop topics include adaptive processes, agile
management of business processes, case handling, con-
figurable processes, dynamic composition of processes,
emergent workflows, knowledge-intensive processes,
process-aware groupware, process evolution, workflow es-
calation and compliance management, process mining and
learning, workflow flexibility.
2. Overview of the Papers
The invited paper by Reijers discusses what the author
calls the flexibility promise of workflow management sys-
tems. Since workflow systems capture coordination logic
separated from applications, it has been argued that this de-
sign facilitates a flexible adaptation. In contrast to that, the
practice of workflow usage shows that once a workflow de-
finition is operational, it is hardly touched. Organizations
benefit from reduced cycle time, better resource utilization,
and other logistic parameters rather than from an assumed
flexibility of workflow systems.
The paper by Minor, Koldehoff, Schmalen, and
Bergmann presents ongoing work on an adaptive work-
flow management system for digital design projects. This
work is motivated in particular by the difficulty to bal-
ance tight time to market requirements with the need to set
up an error free production process which still allows flexi-
ble adaptation to meet customer needs. The authors propose
to create a process instance from a default workflow defi-
nition and offer an adaptation in three ways: add or delete
a task, split or bundle instances, and reschedule an in-
stance. In this context, only deviations from the standard
process are modelled. The context model allows to as-
sign attribute-value-pairs to the process for efficient re-
trieval, risk management, and monitoring.
The paper by Wild, Wirtensohn, and Weber presents Dy-
namic Engines as an approach to flexible process-oriented
application development by combining rule engine and
workflow technology. At the core of Dynamic Engines
the so-called dynamic logic engine supports the execu-
tion of process logic and business rules. Several further
engines provide calculation, data storage, and secu-
rity services. In Dynamic Engines users define so-called
bricks that capture process semantics and which are exe-
cuted by the dynamic logic engine. Versioning of bricks
permits a flexible adaption to process changes. As such, Dy-
namic Engines supports high- and low-level process logic
as well as business rules.
The paper by Seel, Delfmann, and Rieke introduces a
concept for the introduction of enterprise systems with con-
trolling enabled configurative models. The need for ef-
ficient and effective customization of enterprise systems
stems from major consulting costs both for adapting the sys-
tem and the organization. In this context, configurable ref-
erence models are promising to streamline customization,
e.g., based on model projections. The authors extend the ref-
erence model life cycle with a controlling phase and iden-
tify what extensionsto the meta-model of model projections
are needed to support controlling.
The paper by Freßmann discusses user requirements of
fire service organizations and how adaptive workflow tech-
nologies can be tailored to support them. In fire service or-
ganizations incident commanders play a central role in dy-
namic decision making under time pressure and risk con-
siderations. User requirements comprise information sup-
port, mobility of work, easy to use technology, consolida-
tion of different information sources, search facilitation, and
dynamic adaptation of best practices. On a technical level,
the CAKE system meets these requirements by combin-
ing computer supported collaborative work (CSCW) speech
dialogue and search technology and by offering adaptive
workflows that can be reused using case based reasoning.
The paper by Rinderle, Kreher, Lauer, Dadam, and Re-
ichert presents an approach to represent process changes in
adaptive process management systems and a respective im-
plementation in the ADEPT2 system. Changes to process
templates and process instances have to be captured in a
way which avoids inconsistencies between them. So called
unbiased process instances (i.e., instances that have not
been subject to changes) can be migrated by creating a
copy of the old process template. So called biased process
instances (i.e., instances that have been changed, e.g., be-
cause of exceptional situations) are migrated using a delta
layer that represents the changes to the original template.
These two migration concepts have been implemented in
the ADEPT2 system as a proof-of-concept.
3. Conclusion and Future Research Topics
The different papers illustrate various flexibility needs of
process-aware information systems. Flexibility issues arise
at the level of process definition or process model, at the
level of process instances, and regarding the interrelation of
both. Furthermore, adaptation blurs the tight association be-
tween both run-time and process instance as well as build-
time and process model:
•The case based reasoning concepts used in the work
of Freßmann show how workflow instances become a
template for new instances as best practice recommen-
dations.
•Model configuration such as discussed in Seel et al.
takes a configurable model to build a design-time,
company-specific model which is then utilized to in-
stantiate individual cases of a process.
•The work of Rinderle et al. highlights how changes of
an instance have impact on the template and vice versa.
•The works of Wild et al. and of Minor et al. raise the
question whether a distinction between process tem-
plate and process instance is necessary.
Following these arguments, it appears as if the relation-
ships between run-time and build-time as well as between
process model and process instance have to be rethought in
order to provide the degree of process flexibility which is
needed in practice. This conclusion is also supported by the
paper of Reijers. The inflexibility thathe identifies is caused
by a strict association of build-time with process models and
run-time with process instances. Future research should in-
vestigate the relationship between these two poles in more
detail by leaving behind the classical separation of process
design and process execution as introduced by the work-
flow reference model of the Workflow Management Coali-
tion.
