Predicting Temporal Exceptions in Concurrent Workflows

Predicting Temporal Exceptions in Concurrent Workflows

Iok-Fai Leong (University of Macau, Macau), Yain-Whar Si (University of Macau, Macau) and Robert P. Biuk-Aghai (University of Macau, Macau)
DOI: 10.4018/978-1-4666-0023-2.ch011
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

Current Workflow Management Systems (WfMS) are capable of managing simultaneous workflows designed to support different business processes of an organization. These departmental workflows are considered to be interrelated since they are often executed concurrently and are required to share a limited number of resources. However, unexpected events from the business environment and lack of proper resources can cause delays in activities. Deadline violations caused by such delays are called temporal exceptions. Predicting temporal exceptions in concurrent workflows is a complex problem since any delay in a task can cause a ripple effect on the remaining tasks from the parent workflow as well as from the other interrelated workflows. In addition, different types of loops are often embedded in the workflows for representing iterative activities, and presence of such control flow patterns in workflows can further increase the difficulty in estimation of task completion time. In this chapter, the authors describe a critical path based approach for predicting temporal exceptions in concurrent workflows that are required to share limited resources. This approach allows predicting temporal exceptions in multiple attempts while workflows are being executed. The accuracy of the proposed prediction algorithm is analyzed based on a number of simulation scenarios. The result shows that the proposed algorithm is effective in predicting exceptions for instances where long duration tasks are scheduled (or executed) at the early phase of the workflow.
Chapter Preview
Top

Overview

A workflow management system can be used to host several simultaneously executing workflows. Conflicts usually occur when tasks from instances of different workflows compete for limited resources (e.g. human resources, financial resources, etc).

Figure 1 (a) shows an example of conflict involving tasks from two concurrent workflows when only four units of resources are available for sharing. In this example, resource conflicts occur at tasks T22, T23, T24 and T25. Furthermore, these resource conflicts are linked to the underlying temporal constraints of a task within the workflow.

Figure 1.

(a) Resource conflict in workflows; (b) Temporal constraints of task T

Complete Chapter List

Search this Book:
Reset