Dwelling time probability density distribution of instances in a workflow model

  • Authors:
  • Liu Sheng;Fan Yushun;Lin Huiping

  • Affiliations:
  • CIMS Center, Department of Automation, Tsinghua University, Beijing, China;CIMS Center, Department of Automation, Tsinghua University, Beijing, China;School of Software and Microelectronic, Peking University, Beijing, China

  • Venue:
  • Computers and Industrial Engineering
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper presents a method to judge whether a business process is successful or not. A business process is deemed successful if a large enough proportion of instances dwell in a workflow (wait and be executed) for less than given period. By analyzing instances' dwelling time distribution in a workflow, the proportion of instances which dwell in the workflow for less than any given period will be achieved. The performance analysis of workflow model plays an important role in the research of workflow techniques and efficient implementation of workflow management. It includes the analysis of instances' dwelling time distribution in a workflow process. Multidimensional workflow net (MWF-net) includes multiple timing workflow nets (TWF-nets) and the organization and resource information. The processes of transaction instances form a queuing model in which the transaction instances act as customers and the resources act as servers. The key contribution of this paper is twofold. First, this paper presents a theoretical method to calculate the instances' dwelling time probability density in a workflow where the activities are structured and predictable. Second, by this method the analysis of instances' dwelling time distribution and satisfactory degree based on dwelling time can be achieved. The service time of an instance is specified by the firing delay of the corresponding transition (executing time of the corresponding activity). It is assumed that the service request (processing of a transaction instance) arrives with exponentially distributed inter-arrival times and the firing delay of a transition (executing time of the corresponding activity) follows exponential distribution. Then, the instances' dwelling time probability density analysis in each activity and each control structure of a workflow model is performed. According to the above results a method is proposed for computing the instances' dwelling time probability density in a workflow model. Finally an example is used to show that the proposed method can be effectively utilized in practice.