Enhancing Flexibility in Business Process Management using Declarative Languages

Traditional business process models are based on specifying exactly how to execute the process, e.g., all possibilities have to enter into the model by specifying its control-flow. The drawback of such models is that a rigid behavior is imposed. It means users of a traditional process model have no option unless the paths specified in such model. This kind of model is more adequate to represent repetitive processes, where the designer has to envision and to specify all the possible paths. However, many processes are characterized by some looseness in terms of its structure (e.g. healthcare, disaster management). Nowadays constraint-based processes approaches have received increased interest. In these processes, the control flow is defined implicitly as a set of constraints or rules, and all possibilities that do not violate any of the given constraints are allowed to be executed. Thus, any execution order of activities is possible provided that the constraints are not violated. Thereby, most of time the process execution is driven by user’s choice, thus reaching a great flexibility. The goal of this paper is to present an initial comparison between three declarative approaches used to represent constraint-based processes: Dynamic Condition Response Graphs, Declare/LTL and Supervisory Control Theory.