=Paper=
{{Paper
|id=Vol-3098/dc_191
|storemode=property
|title=Aspect Optimalization of Robotic Process Automation (Extended Abstract)
|pdfUrl=https://ceur-ws.org/Vol-3098/dc_191.pdf
|volume=Vol-3098
|authors=Petr Prucha
|dblpUrl=https://dblp.org/rec/conf/icpm/Prucha21
}}
==Aspect Optimalization of Robotic Process Automation (Extended Abstract)==
https://ceur-ws.org/Vol-3098/dc_191.pdf
Aspect Optimalization of Robotic Process
Automation (Extended Abstract)
Petr Průcha
Faculty of Economics
Technical University of Liberec
Liberec, Czechia
petr.prucha@tul.cz
Abstract—Robotic Process Automation (RPA) is an
expanding technology that is used around the world. RPA is A. RATIONALE
used by many organizations for the automation of routine tasks. Theoretical research indicates that there is a research gap in
Due to the easy-to-use environment of RPA technology, it is easy the investigatation of which aspects of RPA are possible to
to deploy new robotic processes in production. As an innovative optimize. Many companies build and deploy RPA robots yet
technology, there is currently a lack of research and analysis to do not optimize the performance of the robots afterwards due
provide a full understanding of the performance of RPA robots.
to a lack of knowledge and resources. RPA developers can
The proposed research investigates which methods lead to better
efficiency, easier maintenance, and crash-free operation. later use the knowledge from future researchers and
implement some aspects into the development process of
Keywords—Robotic process automation, optimization of RPA, RPA. These aspects can be proposed as guiding principles on
performance and efficiency of RPA how to build and design RPA robots. Due to the novelty and
rapid growth of RPA, guiding principles are lacking for how
I. INTRODUCTION to build more efficient robots that are easy to maintain and
Robotic process automation is a software robot that completes crash-free.
routine tasks of a deterministic process. RPA applies the user's Investigation into these research questions mitigate the
front-end and works according to a predetermined scenario cost of maintenance, hypercare and RPA developers, decrease
employing exactly the same user interface as any employee of time spent on the development of RPA, and promote an
a certain organization. [1] environment in which robots can complete more tasks
One of the reasons why RPA is so popular is due to the allowing companies to use resources for further automation
non-invasiveness of RPA into current IT systems, as well as with more advanced technologies.
the possibility to automate the activities of legacy systems
without the need to invest in back-end architecture. Last but II. RESEARCH GOAL AND RESEARCH QUESTIONS
not least, investing in RPA is more economical than The main goal isto find which attributes of RPA robots can be
developing legacy systems. [2], [3] optimized so that RPAs are more efficient, easy to maintain
RPA technology is already very advanced. Most leading RPA and crash-free due to any exceptions.
software vendors can in all probability automate 90 to 95% of From the main goal, several research questions followthat
deterministic user activity.[4] should answer the research goal.
Most RPA manufacturers strive to create a user-friendly
interface with a good user experience to simplify work. A • RQ 1: What are the vital aspects of RPA which
characteristic feature of RPA technology is called "Low- should be optimized?
Code". Low-Code programming uses a minimum of code • RQ 2: What are the options for increasing the
written by the developer and tries to simplify program performance/efficiency of the RPA robots
development by using a graphical user interface, whereas the • RQ 3: What methods and techniques will help to
developer mainly uses pre-prepared functions of the simplify the maintenance of RPA robots?
integrated development environment (IDE). RPA platforms • RQ 4: Which methods will contribute to the possible
mostly use the advantages of Low-Code programming and exceptions of RPA robots?
RPA platforms can also be called Low-code development
platforms (LCDP) [5]. Very often, Low-Code software is III. METHODOLOGY
associated with the "Drag and Drop" function. [6] In the first step, we research RQ 1 to understand the field
This leads to opening coding and automation to citizen of RPA. We conduct a literature search of scientific studies
developers, who could have a lack of skills and training [7]. and analyses of the current state of the art. In addition, in-
The development of custom Low-Code programs simplifies depth interviews will be conducted with RPA developers,
development and makes it possible to automate more head of Center of Excellence (CoE), and RPA leads who have
activities with Low-Code RPA than it would be to automate real development experience. Based on their knowledge and
processes with traditional coding [1], [8]. These two facts lead the literary research, we can better understand the problem
to poorly performing RPA robots, which are easy to build but with greater insight.
hard to maintain [9]. Performance and maintenance are Based on the discovered findings, we prepare the
significant research challenges of the current state of the art in experiments to answer the research questions. For every
the field of robotic process automation [2]. experiment, an appropriate statistical hypothesis will be
formulated with a method to validate the results.
Copyright © 2021 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0
International (CC BY 4.0).
RQ 2: For running the experiment to investigate RQ 2, it based on an analysis of RPA robot logs. An analysis of UI
will be necessary to prepare a neutral testing environment. We logs can show patterns of why robots crash due to system
will try to prepare testing scenarios to reflect reality. For exceptions. Machine learning algorithms will be used for the
testing, the scenarios will be crucial to choose the proper analysis of the logs to find the rules. It is possible to use
applications. We already know some requirements: We need algorithms like C4.5, RIPPER or PART. If these algorithms
at least one application with API support and Graphical User do not fulfill their purpose, it will be necessary to conduct a
Interface (GUI) in one application and an application with a more intensive study of machine learning and AI algorithms
login. The following requirements will be added to the testing for the automatic recognition of patterns in data.
scenarios. For more relevant results, we will run the Handling business exceptions is more difficult because it
experiments on the most commonly used RPA software, is usually caused by human error. A common error is not
which is UiPath, Automation Anywhere and Blue Prism [10]. adding all alternatives of situations to the process design
One hypothesis for testing RQ 3 is that API is more stable document (PDD) or not fulfilling the assignments from the
without change compared to GUI. In a certain period of time, PDD in the solution design document SDD or later in the
we will be observing changes in the code of API and also in development stage. [11] The problem of discovering all
the code of GUI and after some time we will compare what is routine traces is typical for process mining. Finding all routine
more stable. In the experiment we again choose an application traces has already been solved. The greater challenge is to
with API and GUI. The relevant application can be social discover the candidates of routine traces for automation [12],
media or enterprise resource planning ERP programs. We will [13].
prepare a set of these applications and deploy RPA robots to
perform tasks and after a certain period, we will compare ACKNOWLEDGMENT
results to discover whether robots performing tasks via API This research was sponsored by the Technical University of
are more stable than robots performing tasks via GUI. Liberec via the SGS project with evidence number SGS-2021-
RQ 4: Exception handling of RPA robots is mostly solved 1033. This research was conducted with the help of Pointee.
just by restarting the process and if it is system exceptions, We want to thank all participants of the in-depth interviews
then it usually helps. Many system exceptions are caused by for their time and help.
waiting for loading the certain application or web service. In
this research project we will try to analyze RPA logs and try REFERENCES
to extract patterns from data with a focus on exceptions [1] L. P. Willcocks, M. Lacity, and A. Craig, 'The IT function and robotic
caused by loading the application. process automation', London School of Economics and Political
Science, LSE Library, Oct. 2015. [Online]. Available:
As for other ideas of how to improve and optimize RPA, https://ideas.repec.org/p/ehl/lserod/64519.html
it is possible to iterate a new round of literary research and in- [2] R. Syed et al., 'Robotic Process Automation: Contemporary themes
depth interviews. and challenges', Comput. Ind., vol. 115, p. 103162, Feb. 2020, doi:
10.1016/j.compind.2019.103162.
IV. CURRENT STAGE OF THIS RESEARCH PROJECT [3] A. Asatiani and E. Penttinen, 'Turning robotic process automation
into commercial success – Case OpusCapita', J. Inf. Technol. Teach.
The first phase of this research project has been completed Cases, vol. 6, no. 2, pp. 67–74, Nov. 2016, doi: 10.1057/jittc.2016.5.
including the collection of information from experts by means [4] X. Lhuer, 'The next acronym you need to know about: RPA (robotic
of in-depth interviews and the literature analysis. The first process automation)', 2016.
[5] C. Silva, J. Vieira, J. C. Campos, R. Couto, and A. N. Ribeiro,
experiment will be testing RQ 2 on how to increase 'Development and Validation of a Descriptive Cognitive Model for
performance. The hypothesis is that the RPA robot completes Predicting Usability Issues in a Low-Code Development Platform',
the task faster through API than GUI. This should decrease Hum. Factors J. Hum. Factors Ergon. Soc., May 2020.
the throughput time, and the RPA robots will be able to handle [6] S. Agostinelli, A. Marrella, and M. Mecella, 'Research Challenges for
Intelligent Robotic Process Automation', in Business Process
more tasks. Also, the use of the API should help with the Management Workshops, Cham, 2019, pp. 12–18.
maintenance of RPA robots since one of the other hypotheses [7] M. Oltrogge et al., 'The Rise of the Citizen Developer: Assessing the
is that more changes are released on GUI than API, and the Security Impact of Online App Generators', in 2018 IEEE Symposium
new versions of GUI can cause the robot to crash. on Security and Privacy (SP), San Francisco, CA, May 2018, pp.
634–647. doi: 10.1109/SP.2018.00005.
RQ3: We will run the experiment described in the [8] S. Anagnoste, 'Robotic Automation Process - The next major
methodology to test that API is more stable than GUI. Another revolution in terms of back office operations improvement', Proc. Int.
interesting fact that has to be tested is that when we are Conf. Bus. Excell., vol. 11, no. 1, pp. 676–686, Jul. 2017, doi:
building RPA robots, we should reuse parts of the robot for 10.1515/picbe-2017-0072.
[9] D. Kedziora and E. Penttinen, 'Governance models for robotic
the same task. If the process has the same parts in at least a process automation: The case of Nordea Bank', J. Inf. Technol. Teach.
few steps, for example, the login to an application, it would Cases, vol. 11, no. 1, pp. 20–29, Jun. 2020.
be better to reuse the robots for the part with the login. The [10] 'Robotic Process Automation (RPA) Software Reviews 2021 |
advantage of this is that if changes occur in any application Gartner Peer Insights', Gartner https://www.gartner.com/market/
robotic-process-automation-software (accessed Aug. 19, 2021).
and RPA robots crash, it would be necessary to re-code the [11] J. Schuler and F. Gehring, 'Implementing Robust and Low-
RPA robot to fix the error. If RPA developers reuse the robot, Maintenance Robotic Process Automation (RPA) Solutions in Large
they only have to make the repair on one place/one robot. It is Organisations', SSRN Electron. J., 2018, doi: 10.2139/ssrn.3298036.
similar to guidelines for programing (Do-not repeat yourself - [12] A. Bosco, A. Augusto, M. Dumas, M. La Rosa, and G. Fortino,
"Discovering Automatable Routines From User Interaction Logs,"
DRY). in BPM'19, Forum track, 2019.
Exception handling is in the early stage of the research [13] V. Leno, A. Augusto, M. Dumas, M. La Rosa, F. Maggi, and A.
project. The research project aims to improve errors with Polyvyanyy, 'Identifying candidate routines for Robotic Process
system exceptions but we also research some improvements automation from unsegmented UI logs', ArXiv200805782 Cs, Aug.
2020.
for business exceptions of RPA robots. A possible solution is
Copyright © 2021 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0
International (CC BY 4.0).