=Paper=
{{Paper
|id=Vol-2072/paper4
|storemode=property
|title=The Chances of including Extrinsic Factors in Business Process
Management
|pdfUrl=https://ceur-ws.org/Vol-2072/paper4.pdf
|volume=Vol-2072
|authors=Christian Sturm
|dblpUrl=https://dblp.org/rec/conf/zeus/Sturm18
}}
==The Chances of including Extrinsic Factors in Business Process
Management==
https://ceur-ws.org/Vol-2072/paper4.pdf
The Chances of including Extrinsic Factors in
Business Process Management
Christian Sturm
University of Bayreuth
Christian.Sturm@uni-bayreuth.de
Abstract. Business Process Management (BPM) focuses on organizing
and improving business processes, whereby a process is a set of tempo-
rally ordered activities to achieve a defined goal. Additional perspectives
besides the activity are considered to better adapt the process environ-
ment. However, the process environment is currently restricted to these
perspectives. In the era of digitalization, the amount of produced and
stored data expanded massively, but the information is not used in BPM
yet. Hence, it must be investigated how these new information or extrinsic
factors influence process execution and can be utilized with respect to
BPM in general.
Keywords: Process Modelling, Process Discovery, Business Process Manage-
ment, Context Aware Information Systems
1 Introduction and Motivation
A process model contains the execution order of activities of a process as a mini-
mum requirement. However, most of the agile and flexible processes are described
more accurate by including additional information. The flexibility of processes is
investigated in [7]. To improve BPM, multiperspective process models are aiming
at a more fine-grained adaption of the process’ environment as this environment
tend to be in fact rather varying than static. Currently, the multiperspective
process model concerns five perspectives including for instance who is executing
a task, auxiliary tools for completing a task or certain data attributes which are
produced or consumed by activities [5]. The question that arose in this case asks,
if the current process design is sufficient or if the consideration of even more
information bear chances to lift BPM to a new level.
The research area of BPM is several years old and certainly the world has
underwent a huge change since then. The volume of produced data is explod-
ing reasoning from recent developments in electronic data processing (scalable
computation and storage, IaaS). Using this technical infrastructure as catalyst,
digitalization marks the start of a new era in industry and affects our personal
and professional lives. The internetworking of physical devices in production
(IoT) in conjunction with the acquisition and storage of a vast amount of data
opens up new possibilities. The analysis of this Big Data has entered many areas
N. Herzberg, C. Hochreiner, O. Kopp, J. Lenhard (Eds.): 10th ZEUS Workshop, ZEUS 2018,
Dresden, Germany, 8-9 February 2018, published at http://ceur-ws.org/Vol-2072
Extrinsic Factors in Business Process Management 21
in economics and science yet, but is not excessively investigated in BPM since
now. [6] provides an overview of potential synergy effects concerning IoT and
BPM. The data affecting the process environment is more specified in Sect. 2
and suggestions for a target-oriented analyses w.r.t. BPM are proposed.
2 Extrinsic Factors and enclosed Chances and Challenges
To convey an initial feeling for whereof extrinsic factors can be generated from, a
few examples of data sources are given in this section with a first proposal of a
possible categorization.
Process Related Data is produced near the task execution itself regarding the
physical proximity as well as from a logical point of view. For instance, IT-
systems can capture information like the specification of devices in use for the
task execution (performance indicators, display sizes, operating systems, etc.)
or the bandwidth capacity when mobile devices are used within a network.
Organization Internal Data comprises all information belonging to the supe-
rior entity of process execution, e.g. a company or departments of government
institutes. Examples are the prevailing temperature within a production hall,
which affects the manufacturing process (quality of products, cycle times,
etc.), conditions of stocks in internal warehouses or shift schedules coming
from ERP-systems.
Organization External Data are not further restricted. OpenData-Initiatives
follow the trend to provide public access to certain data. Several governments
published demographic data, which can be included in process management
as well as traffic data or tendencies on (finance) markets. In general, every
data source not related to the superior entity of process execution can be
used.
The following paragraph points out the dissociation of extrinsic factors and
common data attributes. Some characteristics of extrinsic factors are shared
with the commonly known data attribute in the multiperspective process model.
Despite this certain resemblance, it is totally worth and necessary to distinguish
data attributes from extrinsic factors for mainly two reasons.
Semantic Point of View. The core peculiarity of data attributes is the strong
correlation to the process instance and the executed activity respectively.
For instance, assume a fine management system for road traffic violations.
Based on the level of fee to pay, the admonished person may prefer different
payment options. The data attribute (fee) is thereby decisive for further
execution and further mandatory for a valid process instance. In contrast,
extrinsic factors have a rather global point of view on the process and are
present even without a certain execution of the process. However, the factors
may still influence the process. Moreover, the data attribute constitutes often
one single value per one executed activity.
22 Christian Sturm
Technical Point of View. Most solutions reach their limits w.r.t. an efficient
discovery of a model in the context of Process Mining, especially when
considering a second perspective like the organizational perspective or the
data attribute. Taking into account much more information will probably
cause performance drops so considerations concerning this matter are required.
The following gives an overview of open questions and challenges regarding
the implementation of extrinsic factors in BPM.
Selection of Extrinsic Factors. One core challenge is the selection of extrin-
sic factors, which actually do have an impact on the process. As stated,
extrinsic factors comprise a variety of possible influence factors, whereas
the characteristic of a process is usually restricted to a single domain. The
variety forbids a brute-force approach to execute correlation analysis of any
factor with the process model regarding the performance – at least with
conventional computation methods.
Skilful combinations of multivariate statistical approaches with Visual Ana-
lytics techniques (see below) must be employed.
Semantic Challenges. What questions can we answer with the help of extrinsic
factors? This problem touches almost every application case in the world of
data science, and so is necessary to ask in the field of BPM. In [2] and [1]
the process environment is used to optimize the control flow of processes. As
stated, extrinsic factors do not correspond to single activities. Thus, they may
also affect different perspectives of activities, like resources or the lifecycle
of an execution (throughput or idle times). The extrinsic factors may also
prevent processes from its execution while completely different processes are
running. It has to be investigated, how current BPM supports this variety of
information gain or how design or discovery purposes must be adapted.
Technical Challenges. Especially in context of declarative process mining,
including extrinsic factors into the calculations poses a challenge regarding
the performance. Even without concerning extrinsic factors, most of current
solutions cannot finish in reasonable time. Hence, the computing time must
be taken into account. A first approach to tackle this issue using Big Data
technology was presented in [8].
Data Quality. Most of the time when using Data Science technology on a vast
volume of data, the analysis is accompanied with trust issues. Algorithms
are used as a black box and the computation model often bears a lack of
transparency. The input information is badly subjected to quality services,
although input information is untrustworthy and wrong or even too less
input data causes wrong output information. Furthermore, the quality of the
output data can suffer from bias, for instance when using a limited sample
set from the whole recorded data in statistical analysis like in [4].
Visual Analytics. Visual Analytics is an interdisciplinary field providing in-
teresting approaches helping analysts to investigate big-sized and especially
multivariate data. These ideas can help to support conveying discovered
Extrinsic Factors in Business Process Management 23
correlations and dependencies or can even applied in discovery procedures. A
first approach is presented in [3].
Process Life Cycle. The vast amount and variety of data is not manageable
in the design phase. However, in a re-design phase, when extrinsic factors
were determined and the influence was disclosed within a preceded process
mining step, process models may incorporate extrinsic factors. Therefore, it
has to be investigated, if current modelling approaches must be extended
to be able to denote the additional information. The (technical) difficulties
of process mining were stated above and in an optimization step, extrinsic
factors and process information can be included both in machine learning
algorithms to automatically submit proposals for improvements to process
scientists.
3 Conclusion
Business Process Management focuses on optimizing business processes. However,
the high availability of data through digitalization or IoT is exploited only
marginally since now. Current research considers process environment but shows
drawbacks w.r.t. performance, quality of output data or the scope of the analysis.
Thus, it has to be investigated how to boost the performance and which additional
questions can be answered by BPM with extrinsic factors.
References
1. Carvalho, J.d.E.S., et al.: Learning context to adapt business processes. In:
CSCWD (2013)
2. Carvalho, J.d.E.S., et al.: A method to infer the need to update situations in
business process adaptation. Comput. Ind. 71(C), 128–143 (Aug 2015)
3. Dixit, P.M., et al.: Enabling interactive process analysis with process mining and
visual analytics. In: BIOSTEC. pp. 573–584 (2017)
4. Hompes, B.F.A., et al.: A generic framework for context-aware process perfor-
mance analysis. In: OTM (2016)
5. Jablonski, S., Bussler, C.: Workflow management – modeling concepts, architec-
ture and implementation. International Thomson (1996)
6. Janiesch, C., et al.: The internet-of-things meets business process management:
Mutual benefits and challenges. CoRR abs/1709.03628 (2017)
7. Schonenberg, H., et al.: Process flexibility: A survey of contemporary approaches.
In: Advances in Enterprise Engineering I (2008)
8. Sturm, C., et al.: Distributed multi-perspective declare discovery. In: BPM
Demo Track and BPM Dissertation Award (2017)