Ontology Development for Designing and Managing Dynamic Business Process Networks

Recent advances in flexible process design and deployment technologies, such as web services and flexible workflow systems, promise the development of flexible intra- and inter-organizational business process networks with potential economic benefits. However, exploiting this potential requires the development of scalable process-management technologies that provide a consistent viewpoint and common language (for describing properties of the process network) to all stakeholders involved in its management. Current approaches to managing such networks requires one to use a babel of languages (in the formal sense) at different levels of abstraction. The development of explicit process ontologies is one of the potentially viable approaches to provide an unified view. In this paper, we discuss the need for process ontologies, propose a task-based ontological framework synthesizing recent research in the fields of AI planning and software engineering, and illustrate the need to link declarative and procedural process representations. We illustrate the use of this ontological framework for managing processes in developing applications for integrating information from disparate information sources in multiple domains. Further, based on this experience, we discuss potential benefits and areas for future research.     

A web services framework for distributed model management

Distributed model management aims to support the wide-spread sharing and usage of decision support models. Web services is a promising technology for supporting distributed model management activities such as model creation and delivery, model composition, model execution and model maintenance to fulfill dynamic decision-support and problem solving requests. We propose a web services based framework for model management (called MM-WS) to support various activities of the model management life cycle. The framework is based on the recently proposed Integrated Service Planning and Execution (ISP & E) approach for web services integration. We discuss encoding of domain knowledge (as individual models) and utilize the MM-WS framework to interleave synthesis of composite models with their execution. A prototypical implementation with an example is used to illustrate the utility of the framework to enable distributed model management and knowledge integration. Benefits and issues of using the framework to support model-based decision-making in organizational contexts are outlined.