H. M. W. Verbeek

Business process mining: An industrial application

Contemporary information systems (e.g., WfM, ERP, CRM, SCM, and B2B systems) record business events in so-called event logs. Business process mining takes these logs to discover process, control, data, organizational, and social structures. Although many researchers are developing new and more powerful process mining techniques and software vendors are incorporating these in their software, few of the more advanced process mining techniques have been tested on real-life processes. This paper describes the application of process mining in one of the provincial offices of the Dutch National Public Works Department, responsible for the construction and maintenance of the road and water infrastructure. Using a variety of process mining techniques, we analyzed the processing of invoices sent by the various subcontractors and suppliers from three different perspectives: (1) the process perspective, (2) the organizational perspective, and (3) the case perspective. For this purpose, we used some of the tools developed in the context of the ProM framework. The goal of this paper is to demonstrate the applicability of process mining in general and our algorithms and tools in particular.

An Alternative Way to Analyze Workflow Graphs

At the CAiSE conference in Heidelberg in 1999, Wasim Sadiq andMaria Orlowska presented an algorithm to verify workflow graphs [19]. The algorithm uses a set of reduction rules to detect structural conflicts. This paper shows that the set of reduction rules presented in [19] is not complete and proposes an alternative algorithm. The algorithm translates workflow graphs into so-called WF-nets. WF-nets are a class of Petri nets tailored towards workflow analysis. As a result, Petri-net theory and tools can be used to verify workflow graphs. In particular, our workflow verification tool Woflan [21] can be used to detect design errors. It is shown that the absence of structural conflicts, i.e., deadlocks and lack of synchronization, conforms to soundness of the corresponding WF-net [2]. In contrast to the algorithm presented in [19], the algorithm presented in this paper is complete. Moreover, the complexity of this alternative algorithm is given.

Verification of EPCs: using reduction rules and petri nets

Designing business models is a complicated and error prone task. On the one hand, business models need to be intuitive and easy to understand. On the other hand, ambiguities may lead to different interpretations and false consensus. Moreover, to configure process-aware information systems (e.g., a workflow system), the business model needs to be transformed into an executable model. Event-driven Process Chains (EPCs), but also other informal languages, are intended as a language to support the transition from a business model to an executable model. Many researchers have assigned formal semantics to EPCs and are using these semantics for execution and verification. In this paper, we use a different tactic. We propose a two-step approach where first the informal model is reduced and then verified in an interactive manner. This approach acknowledges that some constructs are correct or incorrect no matter what interpretation is used and that the remaining constructs require human judgment to assess correctness. This paper presents a software tool that supports this two-step approach and thus allows for the verification of real-life EPCs as illustrated by two case studies.

Organizational modeling in UML and XML in the context of workflow systems

Workflow technology plays a key role as an enabler in E-Commerce applications, such as supply chains. Until recently the major share of the attention of workflow systems researchers has gone to the exchange of information in cross-organizational processes. Increasingly the focus is shifting from the exchange of data to support for interorganizational workflow processes. One of the initiatives in this direction has been XRL (eXchangeable Routing Language), an extendible instance-based language having an XML syntax and Petri-net semantics. In this paper, we move to the next level by extending XRL with organizational entities, structures, and rules. Hence, we describe an organizational model first in UML and then convert it into an XML DTD. Our organizational model allows for the specification of non-human resources, collections of resources (e.g., departments, teams, etc.), availability of resources, delegation, and role inheritance. Additional features of our proposal are the tight integration of organizational concepts and routing concepts. An important goal of this work is to create standard for organizational modeling much like the X. 500 standard for directories.

XRL/Flower: Supporting Inter-organizational Workflows Using XML/Petri-Net Technology

In this paper, we present the architecture of XRL/Flower. XRL/Flower is a software tool, which benefits from the fact that it is based on both XML and Petri nets. Standard XML tools can be deployed to parse, check, and handle XRL documents. The Petri-net representation allows for a straightforward and succinct implementation of the workflow engine. XRL constructs are automatically translated into Petri-net constructs. On the one hand, this allows for an efficient implementation. On the other hand, the system is easy to extend: For supporting a new routing primitive, only the translation to the Petrinet engine needs to be added and the engine itself does not need to change. Last, but not least, the Petri net representation can be analyzed using state-of-the-art analysis techniques and tools.

Verification of XRL: an XML-based workflow language

XRL (eXchangeable Routing Language) is an instance-based workflow language that uses XML for the representation of process definitions and Petri nets for its semantics. Since XRL is instance-based, workflow definitions can be changed on-the-fly and sent across organizational boundaries. These features are vital for todays dynamic and networked economy. However, the features also enable subtle, but highly disruptive, cross-organizational errors. On-the-fly changes and one-of-a-kind processes are destined to result in errors. Moreover, errors of a cross-organizational nature are difficult to repair. In this paper, we show soundness properties of XRL constructs by using a novel constructive approach. We also describe a software tool based on XML and Petri-net technologies for verifying XRL workflows.

Decomposed Process Mining: The ILP Case

Over the last decade process mining techniques have matured and more and more organizations started to use process mining to analyze their operational processes. The current hype around “big data” illustrates the desire to analyze ever-growing data sets. Process mining starts from event logs—multisets of traces (sequences of events)—and for the widespread application of process mining it is vital to be able to handle “big event logs”. Some event logs are “big” because they contain many traces. Others are big in terms of different activities. Most of the more advanced process mining algorithms (both for process discovery and conformance checking) scale very badly in the number of activities. For these algorithms, it could help if we could split the big event log (containing many activities) into a collection of smaller event logs (which each contain fewer activities), run the algorithm on each of these smaller logs, and merge the results into a single result. This paper introduces a generic framework for doing exactly that, and makes this concrete by implementing algorithms for decomposed process discovery and decomposed conformance checking using Integer Linear Programming (ILP) based algorithms. ILP-based process mining techniques provide precise results and formal guarantees (e.g., perfect fitness), but are known to scale badly in the number of activities. A small case study shows that we can gain orders of magnitude in run-time. However, in some cases there is tradeoff between run-time and quality.

YAWL in the Cloud: Supporting Process Sharing and Variability

The cloud is at the centre of attention in various fields, including that of BPM. However, all BPM systems in the cloud seem to be nothing more than an installation in the cloud with a web-interface for a single organisation, while cloud technology offers an excellent platform for cooperation on an intra- and inter-organisational level. In this paper, we show how cloud technology can be used for supporting different variants of the same process (due to “couleur locale”), and how these organisations can aid each other in achieving the completion of a running case. In this paper we describe how we have brought a BPM system (YAWL) into the cloud that supports variants.

An experimental evaluation of passage-based process discovery

In the area of process mining, the ILP Miner is known for the fact that it always returns a Petri net that perfectly fits a given event log. Like for most process discovery algorithms, its complexity is linear in the size of the event log and exponential in the number of event classes (i.e., distinct activities). As a result, the potential gain by partitioning the event classes is much higher than the potential gain by partitioning the traces in the event log over multiple event logs. This paper proposes to use the so-called passages to split up the event classes over multiple event logs, and shows the results are for seven large real-life event logs and one artificial event log: The use of passages indeed alleviates the complexity, but much hinges on the size of the largest passage detected.

Divide and conquer: a tool framework for supporting decomposed discovery in process mining

In the area of process mining, decomposed replay has been proposed to be able to deal with nets and logs containing many different activities. The main assumption behind this decomposition is that replaying many subnets and sublogs containing only some activities is faster then replaying a single net and log containing many activities. Although for many nets and logs this assumption does hold, there are also nets and logs for which it does not hold. This paper shows an example net and log for which the decomposed replay may take way more time, and provides an explanation why this is the case. Next, to mitigate this problem, this paper proposes an alternative way to abstract the subnets from the single net, and shows that the decomposed replay using this alternative abstraction is faster than the monolithic replay even for the problematic cases as identified earlier. However, the alternative abstraction often results in longer computation times for the decomposed replay than the original abstraction. An advantage of the alternative abstraction over the original abstraction is that its cost estimates are typically better.