Subject-Oriented Business Process Management. Models for Designing Digital Transformations

The subject-oriented business process management (S-BPM) paradigm, with its basic idea that actors form a business processor choreography or orchestration, has impact on many modern trends in BPM. The modelling, encapsulation, communication orientation, and executable environment can not only support trends, but may also provide a new approach to technological and organizational developments to the point of potentially solving open issues in progress. This paper discusses the contribution of S-BPM to particularly important current BPM trends (such as IoT, BPM platforms, NLP, Process Mining) and the solutions that the subject-oriented paradigm can provide.

Agile methods are often used to make initiatives for adaptation or working in business processes more flexible. A particular challenge in the many practices of agile business process management is to establish the necessary self-organization in teams. This is a particular challenge for many companies because the actual state of self-organization is unknown and very many areas (e.g., communication, collaboration, work distribution) are affected. Both challenges can be addressed with the maturity model developed in this paper. The literature-based artifact allows measuring the degree of self-organization in processes in eight design fields for 31 concrete indicators. Using a questionnaire, teams can identify a development path to more agile process management.

While small and medium-sized enterprises (SMEs) make up 99% of registered companies in Germany, only a fraction of them is engaged in Internet of Things (IoT) and Artificial Intelligence (AI) as part of their Industry 4.0 initiatives. Despite the potential of IoT and AI, the prerequisites to use these technologies may not be met by SMEs, or the benefits expected may not be aligned with their needs. This research paper identifies typical characteristics of SMEs in the manufacturing sector through a literature review. In addition, we conducted a brainwriting workshop and discussed the findings among interdisciplinary researchers. Our qualitative research approach revealed 19 distinct barriers classified into three key dimensions. Our findings can assist technology managers and production departments in evaluating their organizations and addressing the identified adoption barriers. Additionally, the results can be used in further research to set up practice-oriented guidelines that support the holistic adoption of IoT and AI in manufacturing SMEs.

This paper discusses the increasing importance of automating business processes, with a particular focus on data integration challenges required for data analytics at a larger organization. In this context, the paper introduces the concept of integration-Platform-as-a-service (iPaaS) solutions and highlights their key principles and advantages. It further presents a practical case study by the data team of Handelsblatt Media Group (HMG) to showcase how iPaaS can significantly increase efficiency of organizations. This paper not only serves as a valuable case study for iPaaS implementation for data teams, but also contributes to the broader discourse on the role of iPaaS technology in enhancing business performance.

Platform economy is well established in the business to customer domain. However, digital platforms are lacking in the business to business realm, especially for SMEs. This work presents a way to create a common frame of reference, that is inevitably needed to efficiently design a digital platform. Following the principles of a Design Thinking approach, models are created, that can be used as communication tool for a large group of diverse stakeholders. The approach includes the two means of initial collection of knowledge: exploratory user stories and capturing of use-case processes. A subject-oriented, unified Synthesis Model is created on his basis. By enriching this model with another layer of information from the user stories and their context, another form of a graphical representation is realized. The presented approach proofs valuable in a big research consortium with about 25 organisations (companies, research institutes, and authorities; mainly SMEs). The models are used to design platform architecture as well as platform processes and link them to one another. In addition intangible assets for digital platform development for SMEs are identified.

Scenarios and stories are widely used across different professional domains. The paper presents a case study on their use in a multidisciplinary project aimed at supporting therapies for patients after stroke. Existing training programs with patients and therapists served as starting point for exploring usage scenarios with the humanoid robot Pepper. Based on a set of envisaged scenarios, implementations for the robot’s behavior in predefined “therapy workflows” were provided. Additionally, the domain-specific language TaskDSL4Pepper was developed that allows therapists to specify their own executable interaction models between the patient and the robot. The case study also reveals the need for involving various domain experts to create richer sets of stories with more alternatives.

‘The Metaverse represents the next generation Internet’ - this industry belief challenges the current understanding of Internet-based interaction, and the way it progresses towards a cyber-physical network. Will development techniques for the Internet-of-Things (IoT), Internet-of-Behaviors (IoB), and Internet of Everything (IoE) deliver what is the essence of Metaverse applications? In this reflection paper this question is tackled through the lens of Digital Process Twins (DPTs), as they are already an effective and efficient technology for Cyber-Physical System (CPS) development. The argumentation of industry for establishing the Metaverse as the next generation Internet allows the identification and specification of design requirements. In this contribution these requirements are derived from conceptual inputs on Metaverse developments. The requirements are used for analyzing subject-oriented Digital Process Twin capabilities on whether they can support design and implementation of Metaverse applications. Benefits that can be delivered by executable Digital Process Twins models do not only concern the design (process) when being supported by subject-oriented abstraction and representation overcoming interoperability barriers, but also by the validation of behavior models and their synchronization at runtime, in particular with physical components increasing the immersion capabilities of Metaverse applications.

Technical changes are an integral part of engineer to order product manufacturing. External as well as internal causes require these modifications throughout the entire product manufacturing process. Various factors prevent full front-loading, which provides for early implementation of the modifications. In this paper a method of partial front-loading is presented. Accordingly, the progression of the process is considered in the development of alternative solutions for change implementation. The data representation of the production processes, the digital shadow, is an essential part of the approach. This enables the filtering of alternatives based on the time progression. With the additional technical evaluation of the options, a two-stage validation is created. The necessary requirements for the extension of the existing engineering change management models are defined. A practical use case serves as a basis for the evaluation of the front-loading approach in engineer to order. In addition to the actual approximation of the front-loading target curve, the percentage of rejects is assessed.

Digital transformation and Industry 4.0 pose challenges for all industries. Small and medium-sized enterprises (SMEs) are particularly affected due to cost pressure and the shortage of skilled workers. Adequate process models are needed to manage data analytics projects (DAP) efficiently and effectively in the face of a steadily growing amount of data. However, existing methodologies in the literature are not widely used in SMEs mainly because they are not addressing their specific needs. In this paper we present a Simplified Reference Model (SRM) for early-stage DAPs and compare it to the well-known Cross-Industry Standard Process for Data Mining (CRISP-DM). Three practical scenarios were used to evaluate the applicability of the SRM and identify weaknesses in the execution of DAPs in manufacturing SMEs. Based on our exploration, the main issues are data availability, insufficient data consistency, and inability to understand complex technical environments. Additionally, the paper highlights the need to develop SME-specific operational guidelines and identify potential barriers to the adaption of advanced technologies.

In the context of digital transformation, digital twin data is increasingly being exchanged across company boundaries. As this data exchange is little or not at all standardized today, this results in major interoperability problems. Consequently, inefficiencies in the transfer of data occur, leading to high efforts in identification, structuring and completion as well as a lack of quality of the digital twin data especially within the engineering phase. Looking at the requirements from industrial experts to solve interoperability in the exchange of digital twin data, the concept of the asset administration shell (AAS) presented by Industry 4.0 offers possible solutions, which have not yet been implemented and validated in practice across organizations. For this purpose, a new, standardized, and cross-company data exchange and collaboration model as well as a prototype implementation in the context of the interoperable digital twin is under development to also analyze the benefits and potentials of the standardization approach. This paper presents the motivation and research problem of the project, and how expert knowledge and relevant existing work can be utilized to develop a solution based on empirical evidence and AAS capabilities.

“Efficiency is doing better what is already being done“ – When economist Peter Drucker came up with this quote, the world was largely unaffected by machine learning (ML) techniques and the way modern artificial intelligence (AI) can facilitate processes. Yet, fraud detection methods had already evolved from manual to data-matching systems in the late 90s. Thereafter, individual risk profiles helped in better understanding the potential roots and causes of fraud. Nowadays, modern AI and ML methods look promising, with applications gaining importance and popularity in various fields. However, the performance, efficiency and ease of implementation may vary substantially for different specifications of fraud detection endeavors. Relating to a general process of customer orders and prompt shipment, ML-supported assessments on fraudulent orders can accelerate transactions. When trying to identify irregular credit card payment, this article highlights the usefulness and limitations of popular ML fraud detection methods in relating them to more traditional approaches. Considering different alternative models, we also assess the volatility in some of the most common modifications in modern ML fraud detection techniques.

Business Process Management is a discipline characterized by a variety of activities and tasks. The process notation BPMN 2.0 is also an activity-and task-oriented modeling language. In this paper, we examine whether PASS, a subject-oriented modeling language, can support today’s requirements for modeling processes. We have gathered relevant research and provided a qualitative evaluation of their results by applying them on a real production use case. The research is showing that the use of PASS as a common modeling language provides a better understanding and overview of complex processes than BPMN 2.0. Additionally, the analysis shows that the modeling of processes via PASS can be simpler. Nevertheless, as our research is providing only a qualitative evaluation more verification is needed for evaluating our results.

Modelability of processes is a recognized and important characteristic of any modeling language. Nevertheless, it is not always purposeful or easy to create process models for every kind of workflow. This article discusses the opportunities and limitations of modeling agile development projects with SCRUM as an example. For this purpose, a BPMN and an S-BPM model for SCRUM are presented. The discussion along recognized rules for good process models shows that both notations provide possible and accurate insights into the process of SCRUM on the one hand. On the other hand, the models raise questions of necessity, added value, and relevance in practice. Practitioners can use the developed models to technically implement agile project management, while researchers benefit from a discourse on opportunities and limitations of modeling agility.

While there does exist a formal interpreter specification for the subject-oriented process modeling language that is the Parallel Activity Specification Schema (PASS), the existing spec is somewhat out-of-date and does not cover every aspect of current PASS. This work analyzes the shortcomings of the existing specification and furthermore contains a new interpreter specification that does fill the gaps. This new interpreter model is written in PASS itself thus forming a recursive definition upon itself (using PASS to specify PASS). Next to covering the gaps, the main benefit of this approach is its much more easy accessibility and understandability.

Subject-oriented modeling has recently found increased application for the modeling of processes in Cyber-Physical Systems (CPS). While models are generally used as a means of dealing with the complexity of such systems, modeling languages are challenged by various cross-cutting concerns that are inherently difficult to depict in an efficient manner. Common examples include regular logging mechanisms, authentication procedures, and the like. If not appropriately supported by the used language, modeling such concerns can result in duplicated model elements that are scattered across the same or different models. They furthermore become tangled with the core functionality needed to realize a business case. This negatively impacts qualities such as model understandability, reusability, maintainability, and evolution. Aspect-oriented programming and modeling concepts have been applied in different contexts to address such issues. In this contribution, we examine subject-oriented modeling with regard to its capabilities for supporting modularity and separation of concerns using the running example of a logistics CPS. Certain limitations are present that could be addressed by an aspect-oriented enrichment. We outline an initial proposal and discuss related considerations and challenges.

Autonomous cyber-physical systems (CPS) will influence our daily lives more and more. Collaboration between intelligent machines and humans will become commonplace and negotiation of the autonomy of the actors involved may be required. The concept of shared autonomy is used to address this situation. An implementation of a subject-oriented architecture for shared autonomy is implemented in the context of a use case in smart logistics. Based on this use case, we present the generic concept and aim at its technical feasibility.

Process compliance is a major concern in the financial services industry. Today there is no standard approach for ensuring the traceability of data transfers, requiring development and maintenance of customized solutions. This paper shows how one of the key technologies for integrating production systems, the OPC Unified Architecture (OPC UA, IEC 62541), can be used for addressing this issue. OPC UA provides a number of out-of-the-box functionalities including history management and data validation, which are generic and applicable to any type of data. However, to date there are hardly any known applications of this standard outside the manufacturing domain. The paper presents the architecture and implementation of a prototype incorporating an OPC UA communication interface, and illustrates its application in the context of insurance processing. This work provides the potential not only for enhanced data traceability within the financial services industry but also for a more enterprise-wide use of OPC UA leading to cross-domain system integration such as envisioned for Industry 4.0.

For public sector digitisation, the working processes of public administration are recorded, documented and optimised. This paper is concerned with the question of how the resulting business process models can be used for the development of cloud-based systems in a model-driven software development approach. In doing so, it focuses on the formulation of a methodical and technological approach called Towerpark. Towerpark comprises two elements: a process modelling notation and a set of technical tools for the building of specific applications using code generation in the context of public administration. The Towerpark modelling notation (TMN) uses elements from the Business Process Model and Notation (BPMN) 2.0, whereas the Towerpark processing components rely on Kubernetes: from BPMN diagrams, Kubernetes manifests are generated, which can be deployed as specific applications inside a cloud infrastructure. We illustrate and evaluate our approach with a case study and give an outlook on further research activities.

This work is a re-investigation of the concept and developed tools for the modeling approach called, Abstract Layered PASS (ALPS). The paper investigates the details of what is actually two separated extensions to the classical subject-oriented modeling approach with the Parallel Activity Specification Shema (PASS). One is the concept of Layered PASS, a way of modeling multi-behavior subjects and ad-hoc extensions, while Abstract PASS is essentially the theory behind declarative process modeling within the paradigm subject-orientation. Both ideas have been further refined since their first inception and are no longer pure conceptual ideas but rather are supported by functional tools and clear application scenarios. The overall goal of this work therefore is to integrate and unify previously fragmented concepts, the additions that have been made over the last years, often due to practical modeling concerns, and present their now existing formalization with an ontological definition that was developed alongside and is compatible with the formal OWL based exchange standard for PASS.