=Paper=
{{Paper
|id=Vol-2196/BPM_2018_paper_22
|storemode=property
|title=Event-based Process Simulation of Operating Resources in Assembly Workflows
|pdfUrl=https://ceur-ws.org/Vol-2196/BPM_2018_paper_22.pdf
|volume=Vol-2196
|authors=Philipp Staudt,Sönke Knoch,Daniel Porta
|dblpUrl=https://dblp.org/rec/conf/bpm/StaudtKP18
}}
==Event-based Process Simulation of Operating Resources in Assembly Workflows==
https://ceur-ws.org/Vol-2196/BPM_2018_paper_22.pdf
Event-based Process Simulation of Operating
Resources in Assembly Workflows
Philipp Staudt, Sönke Knoch, and Daniel Porta
German Research Center for Artificial Intelligence (DFKI)
Stuhlsatzenhausweg 3, 66123 Saarbrücken, Germany
{firstName.lastName}@dfki.de
Abstract. Business Process Simulation is a useful and widely adopted
technique that provides process designers and data analysts with the nec-
essary information to estimate the performance impact of business deci-
sions before they are actually deployed. Manufacturers can benefit from
such simulations especially when a strong changeability of the assembly
workflows and systems is required. In order to integrate the simulation
into existing manufacturing plants, simulations need to consider not only
full simulations but also mixed simulations which include virtual and real
operating resources. In this work, we present a new simulation framework
which allows the process designer to model mixed simulations by defin-
ing virtual operating resources. These can be coupled with real hardware
allowing the simulation to interact with manufacturing components to
test the integrated system for robustness at runtime.
Keywords: BPMN, event-driven simulation, mixed simulation.
1 Introduction
Business process simulation (BPS) is a technique to quantitatively analyze the
process in terms of performance measures, such as time, cost, quality and flexi-
bility [2]. BPS generates a large number of process instances executed based on
stochastic information about task durations and resource information. Instead of
creating the simulation model manually, logs from real process executions may
be used to discover those models from historic data using existing process mining
techniques [5]. Different simulators exist integrated with the Business Process
Model and Notation (BPMN) differing regarding their simulation capabilities
[3]. Since the extensibility of such tools is often difficult, [4] suggest an open and
extensible process simulator based on a plug-in structure.
Developments aiming at smarter factories require solutions for the concept of
changeability. Changing the manufacturing system, e.g. by adding new devices
or exchanging existing ones, cost time if the device start-up fails leading to
downtimes and further debugging. Thus, a flexible and extensible simulation
framework is necessary to allow a virtual start-up in the existing manufacturing
system and to test its compatibility with the running system ex-ante.
F. Casati et al. (Eds.): Proceedings of the Dissertation Award and Demonstration, Industrial Track
at BPM 2018, CEUR-WS.org, 2018. Copyright c 2018 for this paper by its authors. Copying
permitted for private and academic purposes. This volume is published and copyrighted by its
editors.
2 Staudt et al.
We propose a framework allowing mixed simulations running virtual and real
resources together. The framework consists of a Web-based dashboard and the
Basis Simulation System (BaSS) which mediates between the BPM engine and
the virtual as well as the real resources. The framework allows an easy integration
of new resources and the adding and exchange of resources at runtime. Differ-
ent states of resources are coupled through event-based communication with a
BPMN process which can be adapted easily by process designers to control the
assembly flow simulated entirely or partly.
In the following, we present the framework’s architecture and the simulation
interface in Section 2. The usage of the system, its graphical user interface and
an exemplary assembly process which are part of this demo, are illustrated in
Section 3. In Section 4, we draw a brief conclusion.
2 The Framework and the Basis Simulation System
The framework’s architecture can be seen in Figure 1. The core of the frame-
work is the BaSS. It contains the BPM Engine Handler, a distributed Data Store
and an undefined number of virtual operating resources. Simple events emitted
by the BPM engine are processed by the BPM Engine Handler and enriched
with assembly-specific data from the data store to form complex events. These
complex events are propagated to the respective virtual resources via a common
communication interface. The receiving virtual resources can be coupled with
real resources. In this case, they act as adapters which transmit the events to
the real resources. If there is no real resource, the virtual resource simulates the
real one based on the implemented simulation for this resource. This mechanism
allows mixed simulations of factory resources. In addition, the Data Store offers a
REST-Application Programming Interface (API) for monitoring and configura-
tion purposes and thus, provides valuable data input for visualizing and logging
simulated processes aiming at evaluating simulations. A Web-based Dashboard
has been implemented to visualize live and historic process data.
Fig. 1. Overview of the BaSS architecture.
Process Simulation of Operating Resources 3
All components within the BaSS are implemented as OSGi services in Java.
OSGi is a modular system and a service platform allowing the modular de-
sign of the framework in so-called bundles. It enables an easy exchange of bun-
dles, among others representing resources, during runtime which avoids a re-
deployment of the framework and thus interruptions of the assembly process.
Communication between the individual components is realized by using the
Message Queue Telemetry Transport Protocol (MQTT) protocol. It is a light-
weight and well adopted communication protocol in the Internet of Things. As
such, it provides an effective way of communicating thanks to the broker-based
publish / subscribe architecture and the Quality of Service (QoS) levels offered.
The loose coupling of the event-based architecture supports the modular archi-
tecture of our framework. High scalability is provided by using MQTT clustering.
A virtual resource is an abstraction of a real operating resource whose API
it implements. In addition, each real as well as virtual resource has a uniform
life-cycle based on Packaging Machine Language (PackML), cf. [1], which is an
industry technical standard developed for the control of packaging machines
providing a generic state model also applicable to other resources. If a resource
is simulated, it must mimic the behavior of the real one. This is realized by the
implementation of the finite state automaton using an OSGi bundle which is
identical to the behavior of the real resource. The implementation is the entry
point for simulation input in form of stochastic data influencing the simulated
behavior of a resource in terms of duration and error-proneness. If not considered
for simulation, the virtual resource acts as an adapter and forwards events from
the BPM engine to the real hardware.
3 User Interface and Assembly Process
The BaSS and a BPM-Engine Handler, which is based on the Camunda BPM
Engine, together with the visualization dashboard shown in Figure 2 is available
from our cloud server 1 . After starting the application, the following two Web
interfaces are available to the user:
Configuration Admin Apache Felix Console for OSGi bundle configuration.
It is used to set-up the BaSS services. It comes with pre-configured bundles
for the MQTT communication and the necessary operating resources.
Live Dashboard Used for process management and real time visualization of
the process flow. It is connected to the underlying BPM-Engine and can ini-
tialize and destroy processes. Important information like the current process
execution or task execution times can be watched easily. In addition, audit
logs are available showing the process history of the selected process.
For the demo to work, the pre-configured operating resources and the com-
munication must be started. This is done in the configuration admin interface
which forms the control center for starting and stopping resources represented by
1
https://cloud.dfki.de/owncloud/index.php/s/sZpJZC9HKiqemBS
4 Staudt et al.
the respective bundle implementations. To enable publish / subscribe communi-
cation, a suitable MQTT broker must be installed on the system, e.g. ActiveMQ.
Afterwards, the implemented processes can be started. There are three pro-
cesses included in this demo (see Figure 3):
Initialization Initializes all operating resources according to their PackML
state. So, each resource must be reset and started for being able to exe-
cute concrete commands.
Main Assembly A press process is simulated in the demo. First, a button
needs to be pushed before the press is able to perform the press action.
Finalization After the main assembly process, the operating resources can be
completed and stopped according their PackML state model.
4 Conclusion
By providing an easy-to-use framework for the mixed simulation of operating
resources in factory plants, operating resources can be tested individually in a
real manufacturing context. This may be crucial to avoid wrong expectations on
the process which is only simulated entirely. Due to the PackML life-cycle, virtual
resources can easily be defined and coupled with real resources. Manufacturers
may also provide virtual resources which act as simulated operating resources
implementing the state machine of the concrete operating resource. Thereby,
simulated and real components share the same API and can be easily integrated
in processes models.
Fig. 2. BPMN Process Dashboard Showing the Simulated Process.
REFERENCES 5
Reset Push- Start Push-
Button Button
Push-Button Push-Button
reseted started
Reset press Start press
Press reseted Press started
Initialization
Activate press
with Do press Raise press
push-button
Starte Prozess pressed press raised
Press Error
React to press
error
Main Assembly
Stop push-
button
push-button
stopped
Stop press
Press stopped
Finalization
Fig. 3. Complete press process including process initialization (top), the main press
process (middle) and the finalization of the operating resources (bottom).
Acknowledgments. This research was funded in part by the German Federal
Ministry of Education and Research (BMBF) under grant number 01IS16022E
(project BaSys4.0). The responsibility for this publication lies with the authors.
References
[1] ANSI/ISA-TR88.00.02-2015. Machine and Unit States: An Implementation
Example of ANSI/ISA-88.00.01. 2015. url: https : / / www . isa . org /
store / ansi / isa - tr880002 - 2015, - machine - and - unit - states - an -
implementation-example-of-ansi/isa-880001-/43761047 (visited on
06/25/2018).
[2] Marlon Dumas et al. Fundamentals of Business Process Management, Sec-
ond Edition. Springer, 2018.
[3] José Luı́s Pereira and António Paulo Freitas. “Simulation of BPMN Process
Models: Current BPM Tools Capabilities”. In: New Advances in Informa-
tion Systems and Technologies. Ed. by Álvaro Rocha et al. Cham: Springer
International Publishing, 2016, pp. 557–566.
[4] Luise Pufahl, Tsun Yin Wong, and Mathias Weske. “Design of an Extensible
BPMN Process Simulator”. In: Business Process Management Workshops.
Ed. by Ernest Teniente and Matthias Weidlich. Cham: Springer Interna-
tional Publishing, 2018, pp. 782–795.
[5] A. Rozinat et al. “Discovering Simulation Models”. In: Inf. Syst. 34.3 (May
2009), pp. 305–327.