Automatic Configuration of Queueing Network Models from Business Process Descriptions

Abstract

The purpose of this study was to develop a transformation scheme for automatic configuration of queueing network models from business process descriptions. Such a transformation scheme would enable queueing analysis of process models within an enterprise modeling framework, which is theory driven and links business process descriptions with formal qualitative and quantitative analyses in an integrated manner. A framework of this kind is suitable for next generation enterprises such as e-businesses, virtual enterprises, and global supply chains. The aim of this research was to extend the analysis capability of such an enterprise modeling framework. The transformation scheme uses the process control flow and task resource requirements to create a view where activities belonging to one or more process instances flow through a network of resources. Two alternative approaches were explored to automatically configure a queueing network model from a business process description. The first approach called the single-step approach generates a queueing network model from a business process markup language description. The second approach called the multi-step approach generates a queueing network model from a formal Petri-net based business process representation, which is described using the Petri Net Markup Language (PNML). An XML-based interchange format, called the Queueing Network Markup Language (QNML), was developed as part of this research effort to store queueing network descriptions. As a proof of concept, the transformation scheme was implemented as part of the DIME framework, developed at the Center for Computer Integrated Manufacturing Enterprises, Oklahoma State University. The DIME framework was a result of research funded by the National Science Foundation through grant # DMI-0075588, under the Scalable Enterprise Systems Initiative. Two feasible approaches were developed to automatically configure queueing network models from business process descriptions. This research also led to the development of QNML, a Queueing Network Markup Language, which is a XML-based format to describe queueing network models. This work has also enabled the queueing analysis capability within the DIME framework.