This site is dedicated to documenting research and experiences in the area of business process configuration. This study has been conducted by QUT's Business Process Management (BPM) Research Group - Brisbane, Australia, in collaboration with TU/e's BPM Reseach Group - Eindhoven, The Netherlands, and supported by the Australian Research Council (ARC) Discovery Grant "Next Generation Reference Process Models". Some of the case studies have been elaborated with the Australian Film Television & Radio School (AFTRS), as part of the ARC Centre of Exellence "BPM for the Creative Industries".

This research project has led to the development of a number of tools for process configuration, which can be found in the Tools section. The Documentation section provides an updated list of the academic contributions related to this project, while the Links section lists the contributors to the project, among providing some interesting links to the process configuration topic.

Business process configuration - an introduction

Business process configuration relates to the life cycle of configurable reference process models, from the design of languages to capture process variability, to the support of process configuration and execution.

A reference process model is an integrated representation of multiple variants of a business process, which is designed to be individualized to meet specific requirements. As such, reference process models promote systematic reuse of proven or common practices.

Reference process models are therefore general solutions in a given domain (e.g. logistics, human resources management), which need to get aligned to the business rules of a specific setting, i.e. an enterprise or a new project.

The best way to understand the concept of configurable reference process model, is by means of an example. Let's consider two common practices for the Post-production process in a screen business project: i) Tape shooting, and ii) Film shooting. As shown by the following picture, both the practices share some commonalities. In fact, whether we are interested in shooting on Tape or on Film, we always need to Prepare the footage for edit and carry out an Offline edit: these two activities are mandatory. However, shooting on Tape implies to perform an Online edit after the first two activities, while shooting on Film implies to perform a Negmatching. The possibility to choose between Online edit and Negmatching, according to certain conditions (in this case the shooting medium), represent a variability in the Post-production process.

The reference process model is thus the combination of both the proven practices in a unique artifact. Here, an OR after activity Offline edit represents a variation point, i.e. a point in the process in which a decision needs to be taken, whether to choose the Tape variant (represented by the left-hand side branch of the OR) or the Film variant (the right-hand side branch).

Therefore, a reference process model is an artifact in which variation points are explicity represented by means of e.g. graphical aids, and associated to a set of variants. The selection of the most suitable variant for each variation point is called configuration, while the process of transforming the reference process model into a derived model, according to a configuration (e.g. by dropping some branches), is called individualization.

The reference process model lifecycle

The decision of which variant to assign to each variation point, needs to be taken before deploying and possibly executing the derived process model. In other words, this is not a run-time decision, i.e. a decision relying on the availability of some data, but rather a design-time decision. The traditional process models lifecycle is made up of four stages: i) design, ii) implementation, iii) execution and iv) diagnosis. Hence, to leverage reference process models in the process lifecycle, we need to introduce a new stage, namely configuration, in which these decisions can be taken. This new stage will follow the design phase, in which reference process models are constructed by process modelers with the help of domain experts, as shown by the following picture.

Once all variation points have been assigned a variant, the individualization process can take place, thus producing a derived process model. At this state, the obtained model can be fine-tuned, before being deployed for execution.

Aims of this research

Nowadays, the benefits of reusing reference process models compared to modeling business processes from scratch, are widely accepted. The systematic adoption of reference process models can reduce the modeling costs and increase the use of proven or common practices. For example, enterprise system packages such as SAP ERP, provide business processes covering common functions such as invoicing, financial reporting and controlling. Analysts and developers configure these processes to meet the requirements of individual customers.

However, reference process models in commercial use still suffer from the following limitations:

1. Lack of usability by non-modeling experts - the process of configuring a reference model to the demands of an organization requires users to have a thorough understanding of both the domain and the modeling language the reference model has been constructed in. They must be capable of estimating the impact of configuration decisions on the model. Consequently, the user who carries out the configuration must not only be a domain expert but also be skilled in reading and configuring reference models. This assumption is unrealistic in application domains where users are unfamiliar with modeling notations.

2. Lack of configurability - commercial products tend to utilize available general process modeling techniques. The unavailability of a dedicated reference modeling language leads to limitations in the expressive power. Firstly, it is not obvious which configuration alternatives exist while a system is being implemented. Secondly, the models do not provide any decision support for the actual selection of an alternative. They do not guide the user as to what might be a recommended configuration given the user's environment. Furthermore in existing languages for reference process models, configurability is concerned only with the control-flow aspects of the process model and tends to oversimplify other factors participating in the process, such as business objects and human roles.

3. Lack of executability - current reference process models are designed as conceptual models in order to facilitate discussions in the early phases of a BPM project. As such, they are intuitive suggestions rather than a precise specification of IT requirements. Also, the methods employed to configure reference process models are manual and error-prone. In particular, analysts are left with the burden of ensuring the correctness of the individualized process models and to manually fix errors. This leads to limitations, since these models cannot be directly used by developers to derive executable workflow specifications. As a result, implementations usually diverge from the original models.

This research aims at addressing the above shortcomings, by designing, implementing and validating a framework for reference process configuration, which:

• relies on the use of interactive questionnaires to facilitate the configuration of reference process models by domain experts, without requiring process modeling knowledge;

• allows an incremental configuration, with the purpose to ensure the correctness of the derived process models, with respect to both syntax and behavioral semantics. This in turn reduces the effort of making the derived models ready for execution;

• extends the configuration to other process elements, such as business objects and human roles, beyond the process control-flow, and capture their interplay.