Effective Role Resolution in Workflow Management

  • Authors:

  • Daniel D. Zeng;J. Leon Zhao

  • Affiliations:

  • Department of Management Information Systems, Eller College of Management, University of Arizona, Tucson, Arizona 85721, USA;Department of Management Information Systems, Eller College of Management, University of Arizona, Tucson, Arizona 85721, USA

  • Venue:
  • INFORMS Journal on Computing
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Workflow systems provide the key technology to enable business-process automation. One important function of workflow management is role resolution, i.e., the mechanism of assigning tasks to individual workers at runtime according to the role qualification defined in the workflow model. Role-resolution decisions directly affect the productivity of workers in an organization, and consequently affect corporate profitability. Therefore it is important to develop effective policies governing these decisions. However, there has not been a formal treatment of role-resolution policies in the literature. In this paper, we analyze role-resolution policies used in current workflow practice and propose new optimization-based policies that utilize online batching. Through a computational study, we examine three workflow-performance measures including maximum flowtime, average workload, and workload variation under these policies in different business scenarios. These scenarios vary by overall system load, task-processing-time distribution, and the number of workers. Based on computational results, we obtain the following insights that can help guide the selection of role-resolution policies. (a) As the overall system load increases, the benefit of using batching-based online optimization policies becomes more significant. (b) Processing-time variation has a major impact on workflow performance, and higher variation favors optimization-based policies. (c) Online optimization has the potential to reduce average workload significantly, and to reduce workload variation significantly as well.