Jan Martijn E. M. van der Werf

Process Discovery Using Integer Linear Programming

The research domain of process discovery aims at constructing a process model (e.g. a Petri net) which is an abstract representation of an execution log. Such a model should (1) be able to reproduce the log under consideration and (2) be independent of the number of cases in the log. In this paper, we present a process discovery algorithm where we use concepts taken from the language-based theory of regions, a well-known Petri net research area. We identify a number of shortcomings of this theory from the process discovery perspective, and we provide solutions based on integer linear programming.

A business process mining application for internal transaction fraud mitigation

Corporate fraud these days represents a huge cost to our economy. In the paper we address one specific type of corporate fraud, internal transaction fraud. Given the omnipresence of stored history logs, the field of process mining rises as an adequate answer to mitigating internal transaction fraud. Process mining diagnoses processes by mining event logs. This way we can expose opportunities to commit fraud in the followed process. In this paper we report on an application of process mining at a case company. The procurement process was selected as example for internal transaction fraud mitigation. The results confirm the contribution process mining can provide to business practice.

On profiles and footprints --- relational semantics for petri nets

Petri net systems have been successfully applied for modelling business processes and analysing their behavioural properties. In this domain, analysis techniques that are grounded on behavioural relations defined between pairs of transitions emerged recently. However, different use cases motivated different definitions of behavioural relation sets. This paper focusses on two prominent examples, namely behavioural profiles and behavioural footprints. We show that both represent different ends of a spectrum of relation sets for Petri net systems, each inducing a different equivalence class. As such, we provide a generalisation of the known relation sets. We illustrate that different relation sets complement each other for general systems, but form an abstraction hierarchy for distinguished net classes. For these net classes, namely S-WF-systems and sound free-choice WF-systems, we also prove a close relation between the structure and the relational semantics. Finally, we discuss implications of our results for the field of business process modelling and analysis.

Context-Aware compliance checking

Organizations face more and more the burden to show that their business is compliant with respect to many different boundaries. The activity of compliance checking is commonly referred to as auditing. As information systems supporting the organizations business record their usage, process mining techniques such as conformance checking offer the auditor novel tools to automate the auditing activity. However, these techniques tend to look at process instances (i.e., cases) in isolation, whereas many compliance rules can only be evaluated when considering interactions between cases and contextual information. For example, a rule like a paper should not be reviewed by a reviewer that has been a co-author cannot be checked without considering the corresponding context (i.e., other papers, other issues, other journals, etc.). To check such compliance rules, we link event logs to the context. Events modify a pre-existing context and constraints can be checked on the resulting context. The approach has been implemented in ProM. The resulting context is represented as an ontology, and the semantic web rule language is used to formalize constraints.

Business Process Modeling Using Petri Nets

Business process modeling has become a standard activity in many organizations. We start with going back into the history and explain why this activity appeared and became of such importance for organizations to achieve their business targets. We discuss the context in which business process modeling takes place and give a comprehensive overview of the techniques used in modeling. We consider bottom up and top down approaches to modeling, also in the context of developing correct-by-construction models of business processes. The correctness property we focus on is soundness, or weak termination, basically meaning that at every moment of its execution, a process has an option to continue along an execution path leading to termination, which is an important sanity check for business processes. Finally, we discuss analogies between business processes and software services and their orchestrations and argue the applicability of the described modeling techniques to the world of services.

Designing weakly terminating ROS systems

The Robot Operating System (ROS) is a popular software framework to develop and execute software for robot systems. ROS supports component-based development and provides a communication layer for easy integration. It supports three interaction patterns that are essential for control systems: the publish-subscribe pattern, the remote procedure call pattern and the goal-feedback-result pattern. In this paper we apply Petri nets to develop a structured design method for ROS systems, such that the weak termination property is guaranteed. The method is based on stepwise refinement using three interaction patterns and components modeled as state machines. The method is illustrated with a case study of robot ROSE.

A software framework for automated verification

This paper describes a software framework supporting the automated verification of models. The framework allows analyzing different kinds of behavioral models of software systems and business processes like UML activity diagrams and BPEL models. To extend the applicability of the verification tools, a variety of transformation tools have been integrated in the framework.

Web Services, Formal Methods, and Behavioral Types

s of Invited Talks Verification of Data-Aware Processes

Refinement of Synchronizable Places with Multi-workflow Nets

Stepwise refinement is a well-known strategy in system modeling. The refinement rules should preserve essential behavioral properties, such as deadlock freedom, boundedness and weak termination. A well-known example is the refinement rule that replaces a safe place of a Petri net with a sound workflow net. In this case a token on the refined place undergoes a procedure that is modeled in detail by the refining workflow net. We generalize this rule to component-based systems, where in the first, high-level, refinement iterations we often encounter in different components places that represent in fact the counterparts of the same procedure “simultaneously” executed by the components. The procedure involves communication between these components. We model such a procedure as a multi-workflow net, which is actually a composition of communicating workflows. Behaviorally correct multi-workflow nets have the weak termination property. The weak termination requirement is also applied to the system being refined. We want to refine selected places in different components with a multi-workflow net in such a way that the weak termination property is preserved through refinements. We introduce the notion of synchronizable places and show that weak termination is preserved under the refinement of places with multiworks if and only if the refined places are synchronizable. We give a method to decide if a given set of places is synchronizable.