Declarative process models are increasingly used since they fit better with the nature of flexible process-aware information systems and the requirements of the stakeholders involved. When managing business processes, in addition, support for representing time and reasoning about it becomes crucial. Given a declarative process model, users may choose among different ways to execute it, i. e., there exist numerous possible enactment plans, each one presenting specific values for the given objective functions (e. g., overall completion time). This paper suggests a method for generating optimized enactment plans (e. g., plans minimizing overall completion time) from declarative process models with explicit temporal constraints. The latter covers a number of well-known workflow time patterns. The generated plans can be used for different purposes like providing personal schedules to users, facilitating early detection of critical situations, or predicting execution times for process activities. The proposed approach is applied to a range of test models of varying complexity. Although the optimization of process execution is a highly constrained problem, results indicate that our approach produces a satisfactory number of suitable solutions, i. e., solutions optimal in many cases.
|Conference||24th International Conference on Advanced Information Systems Engineering|
|Period||25/06/2012 → 29/06/2012|
|Series||Lecture Notes in Business Information Processing|
- Clinical guidelines
- Constraint programming
- Declarative models
- Temporal constraints