Business Process Management Workshops

Reference models provide generic blueprints of process models that are common in a certain industry. When designing a reference model, stakeholders have to cope with the so-called ‘dilemma of reference modeling’, viz., balancing generality against market specificity. In principle, the more details a reference model contains, the fewer situations it applies to. To overcome this dilemma, the contribution at hand presents a novel approach to mining a reference model hierarchy from large instance-level data such as execution logs. It combines an execution-semantic technique for reference model development with a hierarchical-agglomerative cluster analysis and ideas from Process Mining. The result is a reference model hierarchy, where the lower a model is located, the smaller its scope, and the higher its level of detail. The approach is implemented as proof-of-concept and applied in an extensive case study, using the data from the 2015 BPI Challenge.

The emerging concept of multidimensional process mining adopts the ideas of data cubes and OLAP to analyze processes from multiple views. Analysts can split the event log into a set of homogenous sublogs according to its case and event attributes. Process mining techniques are used to create an individual process model for each sublog representing variants of the process. These models can be compared to identify the differences between the variants. Due to the explorative character of the analysis, interactivity is crucial to successfully apply multidimensional process mining. However, current approaches lack interactivity, e.g., they require the analyst to re-perform the analysis steps after changing the view on the data cube. In this paper, we introduce a novel framework to improve the interactivity of multidimensional process mining. As our main contribution, we provide a generic concept for interactive process mining based on a stack of operations.

Many process mining tools and techniques produce output models based on the counting of transitions between tasks or users in an event log. Although this counting can be performed in a forward pass through the event log, when analyzing large event logs according to different perspectives it may become impractical or time-consuming to perform multiple such passes. In this work, we show how transition counting can be parallelized by taking advantage of CPU multi-threading and GPU-accelerated computing. We describe the parallelization strategies, together with a set of experiments to illustrate the performance gains that can be expected with such parallelizations.

In recent years, a multitude of techniques has been proposed for the task of clustering traces. In general, these techniques either focus on optimizing their solution based on a certain type of similarity between the traces, such as the number of insertions and deletions needed to transform one trace into another; by mapping the traces onto a vector space model, based on certain patterns in each trace; or on the quality of a process model discovered from each cluster. Currently, the main technique of the latter category, ActiTraC, constructs its clusters based on a single objective: fitness. However, a typical view in process discovery is that one needs to balance fitness, generalization, precision and simplicity. Therefore, a multi-objective approach to trace clustering is deemed more appropriate. In this paper, a thorough overview of current trace clustering techniques and potential approaches for multi-objective trace clustering is given. Furthermore, a multi-objective trace clustering technique is proposed. Our solution is shown to provide unique results on a number of real-life event logs, validating its existence.

Flexible business processes can often be represented more easily using a declarative process modeling language (DPML) rather than an imperative language. Process mining techniques can be used to automate the discovery of process models. One way to evaluate process mining techniques is to synthesize event logs from a source model via simulation techniques and to compare the discovered model with the source model. Though there are several declarative process mining techniques, there is a lack of simulation approaches. Process models also involve multiple aspects, like the flow of activities and resource assignment constraints. The simulation approach at hand automatically synthesizes event logs that conform to a given model specified in the multi-perspective, declarative language DPIL. Our technique translates DPIL constraints to a logic language called Alloy. A formula-analysis step is the actual log generation. We evaluate our technique with a concise example and describe an alternative configuration to simulate event logs based on an assumed partial execution as well as on properties that are intended to be checked. We complement the quality evaluation by a performance analysis.

The act of retrieving process models from event-based data logs can offer valuable information to business owners. Many approaches have been proposed for this purpose, mining for either a procedural or declarative outcome. A blended approach that combines both process model paradigms exists and offers a great way to deal with process environments which consist of different layers of flexibility. In this paper, it will be shown how to check such models for correctness, and how this checking can contribute to retrieving the models as well. The approach is based on intersecting both parts of the model and provides an effective way to check (i) whether the behavior is aligned, and (ii) where the model can be improved according to errors that arise along the respective paradigms. To this end, we extend the functionality of Fusion Miner, a mixed-paradigm process miner, in a way to inspect which amount of flexibility is right for the event log. The procedure is demonstrated with an implemented model checker and verified on real-life event logs.

Business process models and ontologies are two essential knowledge artifacts that utilize similar information sources. In this sense, building and managing the relationships between ontologies and business process models provide benefits such as enhanced semantic quality of both artifacts and effort savings. In this study, the PROMPTUM toolset, that enables to model relations between the ontologies and the labels within the process model collections, is presented. In establishing these relations, the PROMPTUM toolset enables definition and management of labels and terms within labels of the process models and the process model elements as resources of domain ontologies. Thus, a related resource is managed as a single resource representing the same real-world object in both artifacts in both creation and maintenance. By providing the required features, the toolset supports not only building business process models and domain ontologies together but also building business process models by using existing ontologies and developing ontologies by using business process model collections.

Verification methods to detect errors in the behavior of process models can be formal or informal. The former are based on formal languages, whereas the latter are based on heuristics. The main advantage of informal methods with respect to the formal ones is their short run-time. However, heuristics may lead to false positives, i.e. they may detect errors in a process model even though such model is correct. In this work, we propose using ontologies to formalize heuristics that avoid false positive scenarios. With ontologies it is possible to avoid ambiguities in heuristics that may lead to inaccurate implementations and to enable their execution by ontology reasoners. To this aim, we propose a set of false positive scenarios and define SWRL rules and SPARQL queries to formalize heuristics for such scenarios by means of ontologies. In addition, we identified three requirements that should be met in order to formalize heuristics and their false positive scenarios.

Business process architectures allow to organize business processes and their relations. The entity-centric approach for business process modeling may offer new insights on this field. We present an entity-centric procedure for deriving a business process architecture baseline using the following core ideas: (i) domain model entities may be interpreted as business entities at a higher level of abstraction than usually used by entity-centric approaches, (ii) domain model relationships provide useful information for deriving a business processes architecture. We present the procedure in combination with an application example. The resulting business process architecture baseline specifies: (i) decomposition, specialization, and trigger relations, and (ii) core and support classification of business processes. These results show the potential of our approach. The main contribution of our work is providing guidelines to obtain business process architectures that may be used as reference models for companies within the same industry.

The different needs and domains of enterprises and how they employ EA modelling languages and tools can give rise to heterogeneity at the syntactical, structural and semantical levels. In particular, models dealing with the process perspective are becoming increasingly complex and hetereogeneous. This raises difficulties in managing and reusing EA models.To address the heterogeneity of EA models, we propose an approach that relies on encoding the models as OWL ontologies, then applying ontology matching techniques to map them. By using ontology-based techniques applied to the heterogeneity analysis of EA models.We have applied our approach to a well known benchmarking data set (EMISA-PMMC) encoded in BPMN. This involved the creation of a novel ontology matching algorithm specifically designed for business processes and their integration in a state of the art ontology matching system, AgreementMakerLight.

There is a broad consensus that design decision making is important for Business Process Management success. Despite many business process design approaches and practices that are available, the quality of business process analysis and design relies heavily on human factors. Some of these factors concern cognitive biases. In this paper, we explore the role of cognitive biases in four key issues regarding the design-time phases of the business process management lifecycle. We outline some research directions that may help us understand and improve the effects of cognitive biases in the design-related practices of business process management.

Agile methodologies are established in software development projects. Agility emphasizes, e.g. rapid development and facilitates communication among all stakeholders. Therefore, these principles might be useful in Business Process Modeling projects too. Till now, it is not clear how these principles could be applied in business process modeling. The contribution of this paper is a mapping of challenges in BPM projects and agile principles from software engineering that were derived from literature and own industrial projects. This comparison provides a basis of decision making which agile methods could be applied in a BPM project to face specific challenges.

In their effort to control and manage processes, organizations often create process models. The quality of such models is not always optimal, because it is challenging for a modeler to translate her mental image of the process into a formal process description. In order to support this complex human processing task, we are developing a smart process modeling method. This paper describes how we have built the underlying prescriptive theory, which is constructed from existing evidence about successful information processing techniques in cognitive psychology.

One of the challenges envisioned for eGovernment is how to actively involve citizens in the improvement of public services, allowing governments to offer better services. However, citizen involvement in public service design through ICT is not an easy goal. Services have been deployed internally in public organizations, making it difficult to be leveraged by citizens, specifically those without an IT background. This research moves towards decreasing the gap between public services process opacity and complexity and citizens’ lack of interest or competencies to understand them. The paper discusses game design as an approach to motivate, engage and change citizens’ behavior with respect to public services improvement. The design of a sample serious game is proposed; benefits and challenges are discussed using a public service delivery scenario from Brazil.

The term Decision Mining has been put forward in literature to cover numerous applications in a diverse set of contexts. In the business process management community, it typically reflects the way processes and data required for decision purposes in those processes are blended into one model during discovery. However, the upcoming field of decision modeling and management requires the term to be repositioned in order to obtain a better understanding of the interplay of processes and decisions. In this paper, the different approaches that are currently available are delineated and a case is made for a new type of decision mining: one that separates the control flow and decision perspective in a less stringent form compared to existing approaches.

Business process management systems incorporate the possibility of monitoring the behaviour of a company, by observing their business process indicators. Depending on the process executed, and the order of their performances, certain KPIs can be modified to render the company more competitive. This paper proposes the creation of a model-based fuzzy logic that can represent the relation between KPIs and the business processes of the companies. The use of this graph enables business experts to simulate the evolution of the business according to the decisions taken in the governance process, thereby helping in governance activities.

Modeling business processes has become standard among companies to efficiently organize their business operations. Case Management approaches have been proposed to overcome the limited flexibility of traditional business process languages as BPMN when it comes to supporting knowledge-intensive processes. One of such approaches is Chimera, in which business scenarios are captured by a set of process fragments, a domain model, and object lifecycles.When modeling the real-world ITIL incident handling process, we observed that although Chimera is in general well-suited to capture this process, it misses the functionality to restore states of data objects. Such functionality is useful to undo errors of the case manager during process execution or to perform planned rollbacks. Therefore, we extend Chimera with state variables that memorize previous data states, making it possible to restore those states in the further course of the case. Our extension is validated using the real-world incident handling process.

Choreography in business processes is used as a mechanism to communicate various organizations, by providing a method to isolate the behaviour of each part and keeping the privacy of their data. Nevertheless, choreography diagrams can also be necessary inside an organization when a single instance of a process needs to interact and be synchronized with multiple instances of another process simultaneously. The description, by business experts, and the implementation, by developers, of these choreographed models are highly complex, especially when the activities involved in the processes exchange various data objects and with different cardinalities. We propose the automatic detection of the synchronization points, when a choreographed process model is needed. The choreography will be derived from the analysis of the process model, data objects consumed and generated through the process, and the data conceptual model that relates the data objects. A graphical tool has been developed to support where the synchronization points must be included, helping to decide about the patterns that describe how a single model can be transformed into a choreographed model.

In recent years, microservice architectures have emerged as an agile approach for scalable web applications on cloud environments. As each microservice is developed and deployed independently, they can be developed in the platform and programming language that best suite their purposes, using a simple communication protocol, as REST APIs or asynchronous event-based collaborations, to compose them. In this paper, we argue that process engines provide an excellent platform to develop microservices whose business logic involves complex work flows or processes so that a Business Process language can be used as high-level language to develop these services and a process engine to execute it. We identify the requirements for integrating a process engine in a microservice architecture and we propose how the communication and deployment in a microservice architecture can be handled by the process engine.

Crosscutting concerns in business processes have been addressed, among other forms, under the aspects orientation paradigm. The goal is reducing visualization complexity, allowing reuse and improving maintainability. Literature presents techniques that address aspects in BPM lifecycle stages of modeling and implementation. However, those techniques adopt different semantical representations, making the integration between those stages very difficult. This paper proposes a service identification method to select an implementation for aspects in order to meet goals set in the modeling stage. We describe an artifact produced with this purpose within an application scenario where Web Services are discovered and selected during a process execution. We conclude that aspects’ behavior can be flexible and adaptable at runtime .

Process model similarity measures can be utilized for searching process model collections, which is also called similarity-based search. While there are quite a lot of approaches, most of them base on an underlying alignment between the activities of the compared process models. Yet, according to the results of the process model matching contests conducted in recent years, such an alignment seems to be quite difficult to achieve. The Latent Semantic Analysis-based Similarity Search approach described in this paper circumvents the matching challenge by not requiring such a matching. Instead, it uses a Latent Semantic Analysis-based Similarity Measure to query model collections and retrieve similar models. An evaluation with a collection of 80 models resulted in very good results in terms of Precision, Recall, and F-Measure. The best F-Measure value obtained during the experiments was 0.92.

The size of execution data available for process mining analysis grows several orders of magnitude every couple of years. Extracting and selecting the relevant data to be analyzed on each case represents an open challenge in the field. This paper presents a systematic literature review on different approaches to query process data and establish their provenance. In addition, a new query language is proposed, which overcomes the limitations identified during the review. The proposal is based on a combination of data and process perspectives. It provides simple constructs to intuitively formulate questions. An implementation of the language is provided, together with examples of queries to be applied on different aspects of the process analysis.

To work efficiently with and unlock the potentials of business process models, measuring their similarity is a basic requirement. Thus, many automatic similarity measurement approaches have been developed during the last years, which utilize very different aspects of a model. At the same time, it is unclear which measures can be meaningfully applied in which context and how they behave in general. Hence, this paper analyzes how the values of existing similarity measures correlate and how corresponding implementations perform with respect to their resource consumption. The results of our analysis show that the similarity values of most measures highly correlate while their performance prohibits the usage of more than 50% of the measures in practice.

Predicting the final state of a running process, the remaining time to completion or the next activity of a running process are important aspects of runtime process management. Runtime management requires the ability to identify processes that are at risk of not meeting certain criteria in order to offer case managers decision information for timely intervention. This in turn requires accurate prediction models for process outcomes and for the next process event, based on runtime information available at the prediction and decision point. In this paper, we describe an initial application of deep learning with recurrent neural networks to the problem of predicting the next process event. This is both a novel method in process prediction, which has previously relied on explicit process models in the form of Hidden Markov Models (HMM) or annotated transition systems, and also a novel application for deep learning methods.

Current automated methods to identify erroneous or malicious executions of a business process from logs, metrics, or other observable effects are based on detecting deviations from the normal behavior of the process. This requires a “single model of normative behavior”: the current execution either conforms to that model, or not. In this paper, we propose a method to automatically distinguish different behaviors during the execution of a process, so that a timely reaction can be triggered, e.g., to mitigate the risk of an ongoing attack. The behavioral classes are learned from event logs of a process, including branching probabilities and event frequencies. Using this method, harmful or problematic behavior can be identified during or even prior to its occurrence, raising alarms as early as undesired behavior is observable. The proposed method has been implemented and evaluated on a set of artificial logs capturing different types of exceptional behavior. Pushing the method to its edge in this evaluation, we provide a first assessment of where the method can clearly discriminate between classes of behavior, and where the differences are too small to make a clear determination.

Imperative languages like BPMN are eminently suitable for representing routine processes and are likewise cumbersome in case of flexible processes. The latter are easier to describe using declarative process modeling languages (DPMLs). However, understandability and tool support of DPMLs are comparatively poor. Additionally, there may be an affinity to a particular language caused by existing company infrastructure or individual preferences. Hence, a technique for automatically translating process models between different languages is required. Process models usually describe several aspects of a process, such as activity orderings and role assignments. Therefore, our approach focuses on translating resource-aware process models. We utilize well-established techniques for process simulation and mining to avoid the definition of cumbersome model transformation rules. Our implementation is based on a discussion of general configuration principles and a concrete configuration suggestion. The whole translation approach is discussed and evaluated at the example of BPMN and DPIL.

Software processes usually include activities involving several people playing different roles. SPEM provides primitives for defining all the roles involved in each activity. Software process specification notations are not executable and thus supporting tools cannot provide this functionality. Therefore, even having a formal software process specification we cannot achieve all the potential benefits: people have difficulties in following their responsibilities, resulting in a low productivity. The business process domain provides notations that can be executed on a BPMS. There have been attempts to transform SPEM specifications into BPMN. However, there is no natural way to model multi-role tasks in BPMN, and therefore none of these proposals has solved this issue. In this paper we discuss two promising alternatives for modeling multi-role software activities in BPMN: defining compound roles and modeling multi-role tasks as independent processes. We provide an XSLT transformation for automatically generating each of these solutions from a software process specification. We use a real world running example to illustrate the approach.

Team recommendation is a key and little-explored aspect within the area of business process management. The efficiency with which the team is conformed may influence the success of the process execution. The formation of work teams is often done manually, without a comparative analysis based on multiple criteria between the individual performance of the resources and their collective performance in different teams. In this article, we present a multi-criteria framework to allocate work teams dynamically. The framework considers four elements: (i) a resource request characterization, (ii) historical information on the process execution and expertise information, (iii) different metrics which calculate the suitability of the work teams taking into account both individual performance as well as collective performance of the resources, and (iv) a recommender system based on the Best Position Algorithm (BPA2) to obtain a ranking for the recommended work teams. A software development process was used to test the usefulness of our approach.

Business process management is an approach to improve business processes continuously. While factors like cost, quality and time are usually considered, sustainability considerations often fade into the background. With this study, we try to support the improvement of business processes with regard to sustainability. We present a process model that describes an approach of how to integrate sustainability aspects into business process management. We also present a catalogue to support the identification of improvement potential. We execute the process model and the catalogue using an existing business process. Finally, we present a summary and outline shortly the limitations and opportunities for future work.

Business process modeling and business process management has been used to capture, support and improve a large variety of processes and practices in the private and public sector. Traditionally what is regarded as a good business process is strongly related to economic dimensions. With the increasing importance of assuring sustainable development, BPM techniques should to an increasing degree be able to be used to support the goal of sustainability of the supported or automated solution. This paper provides results from a case study in the Carbon Track and Trace (CTT) project on supporting the compilation and reporting of data on greenhouse gas (GHG) emissions on the city level in the form of GHG inventories. Although basic BPM-techniques are applicable on this levels, we have identified a number of challenges and potential improvements to represent the relevant aspects in such cases to support automated and semi-automated solutions.

The transformation of proven methods and knowledge to reusable artifacts in the form of patterns or anti-patterns is very common in many areas of IS research, for instance in Business Process Management (BPM). An area that lacks in support on the basis of patterns is the area of Green Business Process Management. With this paper, we introduce the concept of sustainability patterns in BPM that can be used for the improvement of existing processes or for the design of new processes in due consideration of ecological goals such as the reduction of resource consumption during the executing of these processes. We further present the results of a qualitative analysis that we conducted to extract a total number of 26 Ecological Process Patterns from real-world processes and a catalog with generic process weakness patterns. The identified patterns indicate a latent potential for process enhancement once the patterns are applied to real-world processes.

The impact of human activities has become one of the greatest concerns of our society, leading to global efforts to achieve sustainable development. Business organizations play an important role on this endeavor. Business Process Management (BPM) offers a comprehensive approach for designing, implementing, executing and monitoring business processes. This position paper aims to raise the discussion about how BPM can be extended to consider the direct and indirect effects of the business processes in the environmental, economic and social dimensions of sustainability.

Considerable progress regarding impact factors of process model understandability has been achieved. For example, it has been shown that layout features of process models have an effect on model understandability. Even so, it appears that our knowledge about the modeler’s behavior regarding the layout of a model is very limited. In particular, research focuses on the end product or the outcome of the process modeling act rather than the act itself. This paper extends existing research by opening this black box and introducing an enhanced technique enabling the visual analysis of the modeler’s behavior towards layout. We demonstrate examples showing that our approach provides valuable insights to better understand and support the creation of process models. Additionally, we sketch challenges impeding this support for future research.

Research on the process of process modeling (PPM) studies how process models are created. It typically uses the logs of the interactions with the modeling tool to assess the modeler’s behavior. In this paper we suggest to introduce an additional stream of data (i.e., eye tracking) to improve the analysis of the PPM. We show that, by exploiting this additional source of information, we can refine the detection of comprehension phases (introducing activities such as “semantic validation” or “problem understanding”) as well as provide more exploratory visualizations (e.g., combined modeling phase diagram, heat maps, fixations distributions) both static and dynamic (i.e., movies with the evolution of the model and eye tracking data on top).

To visualize information about process behavior over time, typically timeline based visualizations are used in contemporary analysis tools. When an overview over a large range of process instances and possible repetitive behavior is to be displayed, however, the timeline projection comes with several limitations. In this article, an alternative to the common timeline projection of process event data is elaborated, which allows to project series of time-related events and regularities therein onto a circular structure. Especially for comparing process rhythms in multiple sets of event data, this visualization comes with advantages over timeline projections and provides more flexibility in configuration. A conceptual elaboration of the approach together with a prototypical implementation is presented in this paper.