Inference and propagation of performance constraints from abstract to concrete business workflows

E-manufacturing, E-business and Virtual Enterprises

by bluezio on November 3, 2010 - 14:43

Inference and propagation of performance constraints from abstract to concrete business workflows

Antonio García-Domínguez*, University of Cádiz; Inmaculada Medina-Bulo, University of Cádiz; Mariano Marcos-Bárcena, University of Cádiz

Abstract: Currently, manufacturing organizations of all sizes need to collaborate with others in order to meet their goals. However, achieving an adequate level of communication between their people and their information systems is difficult. Organizations are urged to formalize their processes and implement their information systems flexibly. These tasks can be achieved using workflows and service-oriented architectures, respectively. Abstract business-level workflows can be refined into executable workflows which compose services from the established service-oriented architecture, integrating disparate systems. However, it is hard to ensure that the target performance of the business process is achieved from the concrete software components. In this work, we present two algorithms which infer performance constraints for tasks in a workflow from its global and local restrictions. These tasks can be then further decomposed into nested subgraphs automatically. By repeatedly decomposing the tasks and inferring their constraints, target key performance indicator values can propagate from abstract business processes down to concrete activities.

Download paper

Download presentation

Thankyou for your paper

November 18, 2010 - 19:28 - by Richard Barton Dear authors, thankyou for your paper. I have a question about the relative performance of 'formalised' business processes and those which are left to the control of owner/operators. (flexible service providers!?) In many companies, the interface with customers and vendors is subject to considerable variation and often needs a dgree of flexibility to cope with what is often a lack of understanding of the business process. How could this variation be reduced to impeve applicability of the workflow modelling ? Richard.

Improving applicability of workflow modelling

November 21, 2010 - 09:21 - by bluezio Dear Richard, Thank you for your question. The issue you point out is quite common, not just in workflow modelling, but in any case where tacit knowledge needs to be formalized. In this particular case, I would suggest to formalize the process gradually, starting from the better known parts. The parts which are not formalized yet would have to be manually performed, but they could use support from other tools, such as a CRM system. The data recorded by the CRM could be useful for creating the next revision of the workflow. The workflow managemen system can also provide part of the required flexibility. Some systems allow users to partially rewind the process to a previous state and override some decisions, if required. There are systems which also allow for migrating processes between different versions of a workflow. Antonio

Antonio, Thankyou for the

November 23, 2010 - 17:50 - by Richard Barton Antonio, Thankyou for the reply. It would often seem that the 'customer' interface is a difficult one to model. Of the techniques you mention, CRM is a useful tool to capture the customers behaviour. Are you aware of the latest state of the art in this field that could be a useful approach to get customers to directly contribute to the modelling process ? (I am not up to speed in this sort of area but probably should be!) richard.

CRM is for data collection

November 24, 2010 - 09:14 - by bluezio CRM does not have much to do with collaborative business process modelling. However, its contact management facilities can help collect accurate and up-to-date information, and quite a few solutions allow for defining simple workflows with human tasks, that are triggered when some conditions are met in the system. This helps partially automate some processes. These cannot be defined collaboratively, though, and more advanced workflows will probably need a full Workflow Management System. The current trend in Business Process Modelling is using BPMN 2.0 (either as a notation or running it directly), and there are quite a few tools for modelling in that language. Some of them allow for collaborative editing. SAP has an impressive prototype called Gravity, based on Google Wave: http://www.youtube.com/watch?v=FaNhXPSCQWo.

impressive!

November 28, 2010 - 21:26 - by Richard Barton wow, that is indeed impressive, and the context is similar to many business process re-engineering examples in multi-national companies / corporations. Thankyou for the link! richard

welcome

November 18, 2010 - 19:09 - by Richard Barton Hello to the contributing authors of, and visitors to, the E-manufacturing, E-business and Virtual Enterprises session. I am pleased to be able to act as co-chair of this session and have already enjoyed reading the contributed papers. I hope that there will be some active contribution and that we can see some exchange of ideas, constructive comments and also some less formal commentary and questions around motivation and activity in these areas. If you have any queries then please do let me know, and I look forward to reading all of your comments! Best regards, Richard.