The assignment of tasks to human performers is a critical component in people-centric business process management systems. Workflow management systems typically assign work items using strategies that only consider qualified resources. There are, however, situations, where this approach falls short. For instance, in emergency response situations, tasks need to be carried out by resources that are available immediately, even if they do not match all skill requirements. This paper compares the performance of a set of six task assignment mechanisms for workflow applications using a scenario from the emergency management domain. In particular, we develop and simulate assignment strategies inspired by stimulus/response models derived from swarm intelligence, and benchmark these strategies against conventional task assignment strategies. Our findings show that swarm intelligence-based approaches outperform the traditional assignment of tasks in ad-hoc organizations, and that workflow-based emergency management systems could benefit significantly from these novel task assignment strategies.
|Name||Lecture Notes in Computer Science|
|Conference||5th International Conference on Business Process Management (BPM 2007), September 24-28, 2007, Brisbane, Australia|
|Abbreviated title||BPM 2007|
|Period||24/09/07 → 28/09/07|