When Industry 4.0 meets Process Mining

When Industry 4.0 meets Process Mining

Available online at www.sciencedirect.com ScienceDirect ScienceDirect Procedia Computer Science 00 (2019) 000–000 Procedia Computer Science 00 (2019...

956KB Sizes 0 Downloads 84 Views

Available online at www.sciencedirect.com

ScienceDirect ScienceDirect

Procedia Computer Science 00 (2019) 000–000 Procedia Computer Science 00 (2019) 000–000

Available online at www.sciencedirect.com

ScienceDirect

www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia

Procedia Computer Science 159 (2019) 2130–2136

23rd International Conference on Knowledge-Based and Intelligent Information & Engineering 23rd International Conference on Knowledge-Based Systems and Intelligent Information & Engineering Systems

When Industry 4.0 meets Process Mining When Industry 4.0 meets Process Mining

a a

Cristina-Claudia Osmanaa*, Ana-Maria Ghiranaa Cristina-Claudia Osman *, Ana-Maria Ghiran

Business Informatics Research Center, 58-60 T. Mihali Street, Cluj-Napoca, 400591, Romania Business Informatics Research Center, 58-60 T. Mihali Street, Cluj-Napoca, 400591, Romania

Abstract Abstract Digitization relies on emerging technologies driven by Internet of Things. This not only affects businesses, but also the entire Digitization on smart emerging technologies driven bytransportation. Internet of Things. This not of only affectsunder businesses, butumbrella also theapplies entire society: smartrelies homes, agriculture, and intelligent The connection “things” the same society: smart homes, smartmassive agriculture, intelligent Theindustry connection of “things” the same umbrella also in education through open and online coursestransportation. (MOOCs). Each should adapt tounder the changes brought by applies Digital also education throughanmassive open online (MOOCs). Each industrywith should adapt4.0 to the brought by Mining Digital Era. in This paper makes introduction to the courses challenges that come together Industry andchanges proposes Process Era. This paper makes an introduction to process the challenges cometotogether with Industry 4.0 andadapt proposes Process Mining techniques as a new approach for business analysisthat in order support companies to rapidly to market changes by techniques as a new solutions approach for forcustomers. business process analysis in order to overview support companies rapidly adapt to market changes by offering customized The study provides a brief of Industryto4.0 and gradually introduces Process offering customized for4.0. customers. The study provides a brief overview of Industrybrought 4.0 and by gradually Process Mining as a purview solutions of Industry A case study approach is used to confirm the advantages Processintroduces Mining. Different Mining a purview of Industry describing 4.0. A caseboth studycontrol-flow approach is and usedresource to confirm the advantages brought by Process Mining. Different types ofas graphical visualization perspectives are also presented. types of graphical visualization describing both control-flow and resource perspectives are also presented. © 2019 The Author(s). Published by Elsevier B.V. © 2019 2019 The The Author(s). Authors. Published bybyElsevier B.V. © Published Elsevier B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) This is an open access article under KES the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review Peer-review under under responsibility responsibility of of KES International. International. Peer-review under responsibility of KES International. Keywords: Industry 4.0, Digital Enterprise, Process Mining Keywords: Industry 4.0, Digital Enterprise, Process Mining

1. Introduction 1. Introduction The tremendous development of information technologies has led to the emergence of a new paradigm: Industry tremendous development of information led to in themanufacturing emergence of industry, a new paradigm: Industry 4.0.The Firstly, it was used by German government,technologies in 2011, as ahas strategy being part of the 4.0. Firstly,Strategy it was used German government, in 2011, a strategy in industry,worldwide: being partUnited of the High-Tech 2020byAction Plan [1]. Since then, newas perspectives of manufacturing this concept emerged High-Tech Action Plan [1]. Since new perspectives this concept States refer Strategy to “Smart2020 Manufacturing”, Japan talksthen, about “Innovation 25”,ofChina includes emerged this term worldwide: in “Made inUnited China States refer “Smartdiscusses Manufacturing”, Japan talks about “Innovation China there includes termdifferences in “Made inonChina 2025”, whiletoKorea about “Manufacturing Innovation 2.0”.25”, Although are this slightly how 2025”, while Korea discusses about “Manufacturing Innovation 2.0”. Although there are slightly differences on how * Corresponding author. Tel.: +40264405300 E-mail address:author. [email protected] * Corresponding Tel.: +40264405300 E-mail address: [email protected] 1877-0509 © 2019 The Author(s). Published by Elsevier B.V. This is an open access underPublished the CC BY-NC-ND 1877-0509 © 2019 The article Author(s). by Elsevier license B.V. (https://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of KES International. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of KES International. 1877-0509 © 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of KES International. 10.1016/j.procs.2019.09.386

2

Cristina-Claudia Osman et al. / Procedia Computer Science 159 (2019) 2130–2136 Author name / Procedia Computer Science 00 (2019) 000–000

2131

practitioners and researchers comprehend this paradigm, all of them mention Information Systems, Artificial Intelligence, Cloud Computing, 3D Printing, sensors, Internet of Things (IoT) and Big Data. Moreover, Cyber-Physical Production Systems (CPPS) are considered the main actors of Industry 4.0. All industrial revolutions focused on manufacturing sector and technology was the enabler of each Industrial Revolution. Next, we will summarize the changes brought by each revolution with the purpose to emphasize the role of technology. First Industrial Revolution began at the end of 18th Century and also continued in the 19th Century, when people started using machines to perform their work. Second Industrial Revolution started at the beginning of 20th Century and it was marked by the division of labour and the emergence of electricity, while Third Industrial Revolution put the basis of computerization process heading for automation. Finally, the Fourth Industrial Revolution can be seen as a furtherance of the previous one by offering a digital transformation of companies in order to provide a transparent view of business processes. The changes brought by Industry 4.0 have also affected business models. Every new technology changes the business models by providing new products and services. Work automation tools like Robotic Process Automation (RPA) may lead to more productive businesses. Processes’ activities are executed by virtual robots, element that guarantees the correctness of the performed activities. There are some examples of successful implementation RPA tools in companies from Finland [2], Colombia [3] or United Kingdom [4]. Moreover, a recent study highlights the advantages of merging RPA with Process Mining in order to identify the maturity of business processes [5]. A case study of Vodafone implementing RPA and Process Mining is also presented. Nowadays, in the Digital Era, efficiency is downgraded by organization’s agility [6]. Therefore, the capability of an organization to rapidly adapt to market changes, has a bigger impact on business market resistance than its own business model. A recent study from 2016 [7], shows that across Europe there exists several Digital Manufacturing Initiatives (for instance Smart Industry in Netherlands, Usine du Futur in France, Produtech in Portugal, Fabrica Intelligente in Italy and so on). A very well structured and documented study about Industry 4.0 challenges and opportunities is presented in [8]. The authors identify 138 standards and the most often mentioned ones are: Radio Frequency Identification (RFID), eXtensible Markup Language (XML), Unified Modelling Language (UML) and Internet Protocol (IP). MATLAB was mentioned as being the most used software for digital modelling and simulation. The study also shows that Linux and Microsoft Windows are the most used computer operating systems, while Android is in the top preferences concerning mobile operating systems. Surprisingly, none of the approaches include Process Mining techniques among the enablers of Industry 4.0. Although, there are studies discussing the process-centric perspective of Industry 4.0 using Process Mining [9]. The overall structure of the paper takes the form of four sections, including this introductory unit. Next section introduces Process Mining among the enablers of Industry 4.0 by connecting it to Internet of Things, Lean Thinking and Business Process Management (BPM). Third section presents a running example consisting of an event log describing a document management process. First, we propose a series of questions whose answer will be given by applying different Process Mining algorithms. Then, graphical visualizations are extracted from the event log, and both, control-flow and resource perspective are analysed. Heuristics Miner, BPMN Miner, Dotted Chart and Social Networks are used in diagrammatic visualizations’ extraction. The fourth section presents the findings of the research, focusing on the power of new emerging domains like Process Mining. 2. Industry 4.0 2.1. Internet of Things (IoT) Even since 2008, RFID Working Group has defined the Internet of Things as being the world-wide network of interconnected objects uniquely addressable, based on standard communication protocols [10]. There is some evidence suggesting that value creation is the fundamental key of any business model [11]. Due to the unremittingly development of IoT networks, companies must adapt and rethink their own process of value creation [12]. Gartner [13] expects that by 2020 there will be 20 billion internet-connected things. As IoT asks for knowledge from different domains (for example: Business Intelligence, Management, Information Security, Product Design, Networks, Cloud Computing, etc.), companies should rethink departments’ organization. Currently, the administrative role of HR

2132

Cristina-Claudia Osman et al. / Procedia Computer Science 159 (2019) 2130–2136 Author name / Procedia Computer Science 00 (2019) 000–000

departments extends to digital-savvy group of employees that are able to recognize the most suitable candidates by identifying their domain knowledge [14]. Sales and marketing are now under the same umbrella and new positions like chief commercial officer emerged, aiming to guarantee a closer internal and external alignment [15]. On the other hand, customers’ feedback is becoming increasingly important and critical in business operations. Data analytics and processing are the essential key in understanding customer behaviour. Starting with the last decade, there was a rapidly growing literature on Digital Twin (DT) concept. The literature shows that DT paradigm was first mentioned by NASA in the context of building new generations of vehicles [16]. The main idea behind this concept is that every physical asset of a vehicle has a digital correspondent. Although, some researchers do not agree with NASA as being the originator of this paradigm as in 2002, a similar notion was introduced to define a conceptual model for Product Lifecycle Management [17]. This led to Information Mirroring Model [18]. There are two main purposes of Digital Twin: to predict future behaviour and to interrogate for the current and past behaviour. 2.2. Lean Thinking and Business Processes Management Lean Thinking is a business methodology that targets a new manner of organizing employees’ activities in order to diminish waste [19]. Davenport was one of the first authors which defined a business process as being a set of activities that interact to produce a business outcome [20]. Business Process Management (BPM) approaches provide solutions of identifying, controlling and optimizing organizational processes. As Lean Thinking, one of BPM aims is to reduce waste. The incredible development of IT technologies help companies to rapidly adapt on customers’ demands. The use of information systems like Customer Relationship Management Systems facilitates a better communication with clients by offering promptly solutions on their requirements. Nowadays, these systems are capable to record all customer-company interaction. The analysis of the generated data can help companies to provide customer tailored experiences as Industry 4.0 technologies promise new ways to develop businesses. Process Mining techniques may offer insights about the interaction of customer with the company, enabling the Digital Twin of a company. More details about Process Mining are provided in the next sub-section. 2.3. Process Mining The joinder of BPM concepts with Machine Learning techniques put the basis of a new domain called Process Mining [21]. The starting point of Process Mining is the event log. An event log consists of several traces and each trace has multiple events. By mapping an event log to a business process model, each event from a trace corresponds to an activity. The attributes of an activity can also be associated to an event. Each event must store at least the following details: the name of the event, the timestamp and the resource performing the action. Optionally, details about costs can also be recorded. ProM Framework is an open source tool that supports the largest number of Process Mining algorithms [22]. There also exists commercial versions of Process Mining tools and the most known is Disco [23]. The latest one provides an improved user-experience through its friendlier UI. Additionally, it supports the main ProM functionalities regarding process discovery and analysis. Process Mining tools, regardless their type, require eXtensible Event Stream (XES) files [24] as input for process models extraction. However, CSV files are also accepted as there are algorithms able to convert them into XES format. The most known Process Mining type is process discovery: the conversion of event logs into process models. This Process Mining type, preponderantly, focuses on the order of activities (control-flow perspective) as the most part of ProM plugins provide visual models like Petri Nets, Causal Nets, BPMN diagrams, etc. Another Process Mining type is represented by conformance checking when models’ quality is evaluated [21]. Mainly, the event log is compared to the obtained process model. The last type of Process Mining is the enhancement and it is used to improve or extend the obtained process model. Next section introduces the analysis of an event log describing a document management process. Moreover, the role of Process Mining as part of Industry 4.0 is also highlighted.

4

Cristina-Claudia Osman et al. / Procedia Computer Science 159 (2019) 2130–2136 Author name / Procedia Computer Science 00 (2019) 000–000

2133

3. Running example 3.1. Questions to which Process Mining should answer Process Mining is a strategic key of Industry 4.0 as Process Mining algorithms are able to provide process models based on history of events. We propose 6 questions whose answers will be given using Process Mining techniques.      

How the process-as-is look like? Which are the activities included into the process model? How many activities has every instance of the process model? Which are the start/end activities of the process model? Which are the resources involved into the process model? How long does a case take? The next two sub-sections will provide answers to the questions addressed earlier.

3.2. Control-flow perspective In what follows, a brief description of the event log is introduced. The event log used in this research describes a document management process [25]. Before starting the analysis of the event log, we filtered out start events. Therefore, the event log has 18352 traces and 149693 events. The process starts on 11th of July 2016 and ends on 31st of August 2016. There are 8 types of resources executing the activities. Resources from group 1 execute more than 20% of the activities belonging to the process. Each trace starts with Register activity. We assume that the activity named End is an artificial one, in fact each trace ends with one of the following activities: Place documents in an existing Case or Generate and send an outgoing Document. The process consists of 10 activities and there are 3 activities executed in all possible process instances: Register, Receive a Document, and End (see Fig. 1).

Fig. 1. Discovered process model variants

More than half of the traces (51,13%) describe the following path: Register, Receive a Document, Create a new Case, Invest Document into a new Case, Mark Case, Work on the Case, Generate and send an outgoing Document, and End. There are two more possible variants: a) one consisting of four activities: Register, Receive a Document, Place documents in an existing case, and End, respectively b) one consisting of nine activities: Register, Receive a Document, Create a new Case, Invest Document into a new Case, Mark case, Work on the Case, Confirm of the work on the case, Generate and send an outgoing Document, and End. According to process model variants, activities Work on the Case, Confirm of the work on the case may repeat within process execution. Heuristics Miner [26] provides an extended visualization of Alpha Miner [27] based on frequencies (see Fig. 2). The discovered process is generally sequential, but it also contains decisions and loops. After the registration ends and the document is received, in 26,01% of cases the documents are included in existing cases, otherwise a new case is created and the case ends. On the second possible scenario, four activities are executed in a row: Invest document in a new case, Mark case, Work on the Case and Confirm work on the case. The last two activities repeat in 22,86% of cases.

2134

Cristina-Claudia et al.Computer / ProcediaScience Computer Science 159 (2019) 2130–2136 Author nameOsman / Procedia 00 (2019) 000–000

Fig. 2. Heuristics Net obtained from the event log

Another Process Mining algorithm that provides process model describing the control-flow perspective is BPMN Miner. It discovers a BPMN diagram and the graphical model presents the same behaviour as Heuristics Net (see Fig. 3). Decisions and loops are equally identified: two XOR-splits, respectively two XOR-joins.

Fig. 3. BPMN diagram obtained from the event log

The Dotted Chart confirms that activities are executed only on working days, from Monday to Friday. Mainly, the resources are dedicated to execute specific activities, for instance the registration activity is automatically performed by the System. Resources belonging to group 5 confirm new cases, while the artificial end activity is executed by resources either from group 6 or 7. Mostly, these resources perform the last two possible activities from the process model. The only resources executing two activities are those belonging to group 1 (Create a new Case, respectively Receive a Document). This plugin also provides statistics concerning the duration of cases. The mean duration of cases is almost 65,8 minutes (3949545,717088056 milliseconds), while the shortest case takes approximatively 2 minutes and the longest case takes nearly 63 hours.

Fig. 4. Dotted Chart

3.3. Resource perspective The social network analysis is performed using two algorithms: Working-Together, respectively Handover-of-Work [28]. First algorithm shows that nearly all resources are working within the same case, the only exception being the

6

Cristina-Claudia Osman et al. / Procedia Computer Science 159 (2019) 2130–2136 Author name / Procedia Computer Science 00 (2019) 000–000

2135

resources from group 7, which only interact with the System and resources from group 1. On the other hand, Handoverof-Work connects the resources that interact within the whole process, for instance two resources are connected if they perform any two subsequent activities within a case. Both social networks suggest that resources from group 7 perform the lowest number of activities. The assertions made based on the Dotted Chart are strengthened by this visualization (for example, artificial End activity is executed by resources either from group 6 or 7). Furthermore, each case is initiated by the System, then the work is transferred to resources from group 1. Additionally, the social network also illustrates the work transfer from group 4 to group 5 and contrariwise.

Fig. 5. (a) Social Network (Working-Together), (b) Social Network (Handover-of-Work)

4. Conclusions Towards the increasing amounts of produced and consumed data within information systems, this paper highlights the power of emerging tools and technologies aiming to support operational processes in Digital Era. The data generated by information systems (for instance ERP, CRM, BPM systems) can be converted into a specific format of event logs. There also exists Process-aware Information Systems (PAISs) that automatically generate event logs. These files are the elementary ingredient of Process Mining. Definitely, digitization, a key factor in Industry 4.0 is changing business processes. This research has shown that Process Mining is a purview of Industry 4.0 by using a case study of a document management process. The findings of this research answered six questions introduced in sub-section 3.1. In this way, two perspectives of the process are investigated: control-flow and resource by the instrumentality of graphical visualizations. The findings show that the discovered process models are partially automated as there exists activities executed by a resource called System. The engagement of resources within the process was also discussed. References [1] Kagermann, Henning, Wahlster Wolfgang, and Helbig Johannes. (2013) “Recommendations for implementing the strategic initiative INDUSTRIE 4.0: Securing the future of German manufacturing industry; final report of the Industrie 4.0 Working Group.”, Forschungsunion. [2] Asatiani, Aleksandre, and Esko Penttinen. (2016) “Turning robotic process automation into commercial success–Case OpusCapita.” Journal of Information Technology Teaching 6 (2): 67-74. [3] Aguirre, Santiago, and Alejandro Rodriguez. (2017) “Automation of a business process using robotic process automation (rpa): A case study.” Workshop on Engineering Applications, Springer, Cham, 65-71. [4] Lacity, Mary, Leslie P. Willcocks, and Andrew Craig. (2015) “Robotic process automation at Telefonica O2.”, The Outsourcing Unit Working Research Paper Series, 3-19. [5] Geyer-Klingeberg, Jerome, Janina Nakladal, Fabian Baldauf, and Fabian Veit. (2018) “Process Mining and Robotic Process Automation: A Perfect Match.” Proceedings of the Dissertation Award, Demonstration, and Industrial Track at BPM 2018, 9-14. [6] Sommerfield Basil, Roxana Moise-Cheung, Deloitte. (2016) “The digitally-fit organization”, Inside magazine, Issue 12, Part 01 – From a

2136

Cristina-Claudia Osman et al. / Procedia Computer Science 159 (2019) 2130–2136 Author name / Procedia Computer Science 00 (2019) 000–000

digital perspective, Online: https://www2.deloitte.com/content/dam/Deloitte/lu/Documents/technology/lu_digitally-fit-organization.pdf, Accessed February 15, 2019. [7] European Parliament, “Industry 4.0”, Online: http://www.europarl.europa.eu/RegData/etudes/STUD/2016/570007/IPOL_STU(2016)570007_EN.pdf, Accessed February 15, 2019. [8] Liao, Yongxin, Fernando Deschamps, Eduardo de Freitas Rocha Loures, and Luiz Felipe Pierin Ramos. (2017) “Past, present and future of Industry 4.0-a systematic literature review and research agenda proposal.” International journal of production research 55(12): 3609-3629. [9] Halaška, Michal, and Roman Šperka. (2018) “Process Mining–the Enhancement of Elements Industry 4.0.” In 2018 4th International Conference on Computer and Information Sciences (ICCOINS), 1-6. [10] INFSO D.4 Networked Enterprise & RFID INFSO G.2 Micro & Nanosystems in Co-operation with the Working Group RFID of the ETP EPOSS. (2008) “Internet of Things in 2020, Roadmap for the Future”, Online: https://docbox.etsi.org/erm/Open/CERP%202008060910/Internet-of-Things_in_2020_EC-EPoSS_Workshop_Report_2008_v1-1.pdf, Accessed February 15, 2019. [11] Hui, Gordon. (2014) “How the internet of things changes business models.” Harvard Business Review 92(7/8): 1-5. [12] Naqvi, Syed Ahsan Raza, Syed Ali Hassan and Fatima Hussain. (2017) “IoT Applications and Business Models”, Internet of Things, SpringerBriefs in Electrical and Computer Engineering. Springer, Cham, 46-51. [13] Gartner (2017) “Leading the IoT”, Online: https://www.gartner.com/imagesrv/books/iot/iotEbook_digital.pdf, Accessed February 15, 2019 [14] Schulman, Don, Shiv Iyer, Gerarda E. Van Kirk, and Mohammed Hajibashi. (2017) “Be the new digital enterprise”, Accenture, Online: https://www.accenture.com/t20171024T083850Z__w__/us-en/_acnmedia/Accenture/cchange/digital-enterprise/docs/Accenture-DigitalEnterprise-POV.pdf, Accessed February 15, 2019. [15] Gonzalez, Francisco (2015) “Reinventing the company in the Digital Age”, Online: https://www.bbvaopenmind.com/wpcontent/uploads/2015/01/BBVA-OpenMind-book-2015-Reinventing-the-Company-in-the-Digital-Age-business-innovation.pdf, Accessed February 15, 2019. [16] Glaessgen, Edward, and David Stargel. (2012) “The digital twin paradigm for future NASA and US Air Force vehicles.” Proceedings of 53rd Structures, Structural Dynamics and Materials Conference, Special Session on the Digital Twin, 1818-1832. [17] Grieves, Michael (2005) “Product Lifecycle Management: Driving the Next Generation of Lean Thinking: Driving the Next Generation of Lean Thinking: Driving the Next Generation of Lean Thinking” McGraw Hill Professional. [18] Grieves, Michael (2012) ”Virtually indistinguishable”, Systems engineering and PLM. In L. Rivest, A. Bouras, & B. Louhichi (Eds.), Product lifecycle management: Towards knowledge-rich enterprises, 226–242. [19] Womack, James P., and Daniel T. Jones. (1996) “Lean Thinking”, Simon and Schuster. New York. [20] Davenport, Thomas H. (1993) “Process innovation: reengineering work through information technology” Harvard Business Press, 1993. [21] Van Der Aalst, Wil. (2016) “Process Mining: Data science in action.” Springer, Berlin, Heidelberg. [22] Verbeek, H. M. W., J. C. A. M. Buijs, B. F. Van Dongen, and Wil MP van der Aalst. (2010) “Prom 6: The process mining toolkit.” Proceedings of BPM Demonstration Track 615, 34-39. [23] Günther, Christian W., and Anne Rozinat. (2012) “Disco: Discover Your Processes.” BPM (Demos) 940, 40-44. [24] Günther, Christian W., and Eric Verbeek. (2012) “Standard definition.” Online: http://www.xesstandard.org/_media/xes/xesstandarddefinition-2.0.pdf, Accessed February 15, 2019. [25] Djedović, Almir (2018) “Document Processing Event Logs”, 4TU.Centre for Research Data, Online: https://doi.org/10.4121/uuid:6df27e596221-4ca2-9cc4-65c66588c6eb, Accessed January 17, 2019. [26] Weijters, A.J.M.M., van Der Aalst, W.M. and De Medeiros, A.A. (2006) “Process mining with the heuristics miner-algorithm” Technische Universiteit Eindhoven, Tech. Rep. WP166:1-34. [27] Van der Aalst, Wil, Ton Weijters, and Laura Maruster (2004) “Workflow mining: Discovering process models from event logs.” IEEE Transactions on Knowledge and Data Engineering 16(9): 1128-1142. [28] Van der Aalst, Wil MP, and Minseok Song. (2004) “Mining social networks: Uncovering interaction patterns in business processes.” Proceedings of International conference on Business Process Management, Springer, Berlin, Heidelberg, 244-260.