Comparing and Aligning Process Representations

In this introductory chapter, Sect. 1.1 motivates the need for the alignment and comparison of process information in organizations. Then, Sect. 1.2 introduces the main contributions of our research conducted on this topic. Finally, Sect. 1.3 provides an overview of the remaining chapters.

This chapter discusses background essential to the alignment of process information contained in different informational artifacts. The chapter consists of three parts, each of which focuses on an aspect relevant to any alignment task, as visualized in Fig. 2.1. The input to an alignment task is formed by two informational artifacts related to a business process. Section 2.1 reflects on these input artifacts by discussing the role and presence of process information in organizations.

Organizational stakeholders involved in business processes have different preferences regarding the representation format used to visualize process information. As considered in Sect. 2.1.3, a user’s cognitive style, their experience, and their purpose all influence the suitability of a particular format. As a result, some stakeholders and applications scenarios are better served by process information represented in the form of process models, whereas, in other cases, textual process descriptions are more suitable. For this reason, organizations have recognized the value of maintaining textual process descriptions alongside process models [158]. However, as shown in Sect. 2.1.4, the usage of two representation formats for the same process also comes with the risk of having to deal with inconsistencies between them. As a result of such inconsistencies, users may execute processes based on incorrect information. This can lead to inefficient process execution [31] as well as compliance issues [27]. Therefore, it is crucial that organizations keep their process descriptions in sync [117]. However, the effort required to identify and clear up conflicts for an entire process repository, potentially consisting of hundreds or even thousands of processes [227], is hardly manageable in a manual way. Against this background, this chapter introduces an approach that automatically detects inconsistencies between process models and textual process descriptions. In this way, the approach supports organizations by reducing the effort required to clear up any inconsistencies.

Conformance-checking techniques enable organizations to automatically determine whether business processes are executed according to their specifications. Particularly, they check if observed behavior, as recorded in an IT system and represented in the form of event logs, conforms to the allowed process behavior, typically captured in a process model [20]. The importance of conformance checking has been recognized in various contexts, such as legal compliance [229] and auditing [17]. Due to this importance, numerous conformance-checking techniques have been developed (cf. [12, 20, 191]). A crucial requirement for all these techniques is that the events contained in an event log can be related to the activities of a process model [11]. Without knowing the relations between events and model activities, it is not possible to determine if the behavior within a trace conforms to the behavior specified by a process model. Despite this dependence on the existence of a, so-called, event-to-activity mapping, establishing these mappings is a highly complex task. In particular, mapping techniques face considerable challenges caused by, among others, cryptic event names, non-conforming behavior, and noise [33]. As a result, the goal of mapping techniques is to choose the best mapping from a number of potential ones [167]. This introduces the risk that the selected mapping does not correctly capture the relations between traces and a process model. In the context of conformance checking, selecting an incorrect mapping is particularly harmful. If the selected mapping is incorrect, the results obtained through conformance checking can become incorrect as well. To overcome this issue, this chapter presents a conformance-checking technique that can be applied in spite of an uncertain mapping of events onto activities. Our technique assesses the conformance of a trace by considering the entire spectrum of potential mappings, rather than focusing on a single one. As a result, our conformance-checking technique avoids the risks associated with the selection of an incorrect mapping.

Monitoring Process Performance Indicators (PPIs) represents an important prerequisite for organizations in their strive for continuous process optimization [143]. Therefore, a key task for managers is to define suitable PPIs that are aligned with the strategic business objectives of the organization [223]. Typically, managers achieve this by describing relevant PPIs using natural language descriptions [225, 270]. This has the great advantage that PPIs can be easily specified and understood by all stakeholders [150]. However, to be able to actually monitor PPIs, they must be defined in a way that supports their automated computation and relates them to the technical implementation of a business process in a Workflow Management or Enterprise Resource Planning System [90]. Although there are structured notations that achieve this (cf. [209, 223, 270]), they are not at all similar to the unstructured natural language descriptions used and preferred by managers.

Alignments between processes provide an important basis for a variety of application scenarios and techniques. These alignments are, among others, used for the detection of differences between models [148], the harmonization of process model variants [149], process querying [123], and the propagation of process changes [265]. The accuracy and, therefore, usefulness of such techniques is highly dependent on the correctness and completeness of the alignments that are established by process model matching techniques. However, despite the existence of a plethora of matching techniques, it has been shown that their results leave room for improvement [54]. A possible cause for this is that existing process model matching techniques focus exclusively on information related to the specification of processes, typically by just considering the information contained in process models themselves. Therefore, they ignore information that relates to the actual execution of the processes, as captured in event logs. These logs provide valuable information on data attributes, event durations, and other aspects specifically associated with the enactment of a process.

In this final chapter, Sect. 8.1 provides a summary of the main results, Sect. 8.2 discusses the implications of the presented work for research and practice, and Sect. 8.3 provides an outlook on directions for future research.