Flexible Process Notations for Cross-organizational Case Management Systems

In recent times western economies have become increasingly focussed on knowledge work. Knowledge work processes depend heavily on the expert knowledge of workers and therefore tend to require more flexibility then the processes seen in traditional production work. Over-constrained processes cause frustration and inefficiency because they do not allow workers to use their expert experience to make the best judgements on how to solve the unique challenges they are faced with. However some structuring of their work is still required to ensure that laws and business rules are being followed.

IT Systems for process control have a large role to play in structuring and
organizing such processes, however most of these systems have been developed with a focus on production work and fail to support the more
flexible processes required by knowledge workers. The problem arises at the core of these systems: the notations in which the processes are defined. Traditional process notations are flow-based: control of the process flows from one activity to the next. This paradigm inherently encourages modellers to think in terms of strict orderings instead of supporting flexibility. Flow-based models that do try to capture a large degree of flexibility tend to turn into so-called "spaghetti models", because all possible paths through the process need to be modelled explicitly. Over the last decade new, more flexible, process notations have been researched by using a constraint-based paradigm, where one directly models the rules of a process.

Dynamic Condition Response (DCR) Graphs, which have been developed at
the process and systems models group at IT University of Copenhagen (ITU),
are one such notation. They stand apart from other constraint-based notations
by having a small set of 4 basic constraints, yet offering the full formal expressiveness of both regular and omega-regular languages. They also offer an operational semantics based on transformations of markings, which means that the notation can represent a process at both design- and runtime, facilitating easy reasoning about the execution of the process and techniques for runtime adaptation.

This dissertation reports on the results of the Technologies for Flexible Cross-organizational Case Management Systems (FLExCMS) research project which was started in cooperation between ITU and the company Exformatics A/S. The goals of the project were to strengthen the industrial adoption of constraintbased notations and techniques by further developing DCR Graphs to be industrially applicable, with a particular focus on guaranteeing safety for cross-organizational processes. We will show how DCR Graphs have been extended with new dimensions such as time, data and hierarchy and we will report on the development of techniques for the safe distribution and run-time adaptation of cross-organizational processes based on DCR Graphs. We brought this research into practice at Exformatics by developing tools for modelling and executing DCR Graphs and applying these tools within customer projects, we report on a number of case studies based on these projects. Finally we report on a new angle of research within the Business Process Management field tentatively called Hybrid BPM Technologies, which aims to combine the advantages of both the flow- and constraint-based paradigms.