Flexibility and distribution are major challenges of an advanced workflow management system, but have been addressed mostly separately from each other. In this paper, we present an agent-based workflow enactment service which combines flexible and decentralized workflow execution. Every task is coordinated by its own task (coordination) agent which interacts with related task agents by event passing. Realized as reactive agents, the task agents know how to react to state changes as well as to structural workflow changes so that workflow changes can be handled also in a decentralized enactment architecture. Instead of generating different task or workflow objects for the different task types of a workflow schema, the execution behavior of a task agent is extracted from the workflow schema and explicitly represented within the task agents. Finally, on schema level, the behavior definition can be customized in order to express an adequate behavior for heterogeneous and flexible processes. This is not only the basis for defining less restrictive workflows in advance, but also for supporting dynamic workflow changes in every possible situation.

On the basis of new information technologies, today's business evolves towards distributed and cooperative service processing resulting in a network of co- operating organizations. This paradigm also influences engineering and production processes. Both the characteristics of these processes and the interoperability be- tween different organizations challenge the development of enabling technologies to support flexible and inter-organizational product development. Currently available workflow and document/data management technologies fail to support these distrib- uted and collaborative processes. They lack abilities for supporting inter-organiza- tional processes on the level of workflow as well as of document management. Furthermore, only little support is provided for collaborative workflows. In this paper, we present the flow.net approach which has been developed within the MOKASSIN project. It is based on a workflow component which has been designed to support distributed, dynamic and collaborative process management across several organizations. The standard workflow management architecture is extended with integrated document management services including version and workspace control. In order to support inter-organizational processes, our approach provides a distribu- tion layer which supports both decentralized workflow and document management.

In order to apply workflow technology to a wide range of process types, modeling and enacting concepts for heterogeneous, flexible, and distributed processes are needed. Both, a-priori less restrictive workflows and dynamic adaptation due to unpredictable situations, have to be supported - in the presence of distributed workflow enactment. In this paper, we present the flow.net approach which covers these requirements and supports the modeling of flexible workflows and their dynamic modification within a decentralized enactment model. On the basis of this execution model, we outline how flexible rule-based specification and enacting techniques can be combined with the high-level modeling constructs of task graphs.

In this paper, we focus on the definition and adaptation of the execution behavior of a task in order to support flexible workflows in the presence of distributed workflow enactment. We argue that an adequate behavior definition is the basis for both, modeling less-restrictive workflows in advance as well as supporting dynamic workflow changes. We show how different control flow dependency types can be specified in our approach and can be used to define less-restrictive workflows. Furthermore, we discuss the definition of an adequate behavior for dynamic modifications in different situations. In particular, we describe how the application of and reaction to dynamic changes can be adapted in our approach depending on the process context and the behavior of a task itself.

Information technology support for complex, dynamic, and distributed business processes as they occur in engineering domains requires an advanced process management system which enhances currently available workflow management services with respect to integration, flexibility, and adapt ation. We present an uniform and flexible framework for advanced process management on an a bstract level which uses and adapts agent technology from distributed artificial intelligence for both modelling and enacting of processes. We identify two different frameworks for applying agent tec hnology to process management: First, as a multi-agent system with the domain of process manag ement. Second, as a key infrastructure technology for building a process management system. We will then follow the latter approach and introduce different agent types for managing activities, products, and resources which capture specific views on the process.

On the basis of new information technologies, today’s business evolves towards distributed and co operative service processing resulting in a network of c operating organizations. This paradigm also influences engineering processes. Both the characteristics of engi neer ng processes and the interoperability between differen t organizations challenge the development of enabling t echnologies to support global engineering networking. Currently available workflow and document/data managem ent technologies fail to support these distributed and collaborative processes. They lack abilities for supportin g crossorganizational processes on the level of workflow a s well as document management. Furthermore, only little su pport is provided for collaborative workflows. In this paper, we present the MOKASSIN approach. It is based on a workflow component which has been designed to support distributed, dynamic and collabor ative process management across several organizations. Th e standard workflow management architecture is extend ed with integrated document management services includ ing version and workspace control. In order to support crossorganizational processes, our approach provides a d istribution layer which supports both decentralized work fl w and document management.

Process modeling and enacting concepts are at the center of workflow management. Support for heterogeneous processes, flexibility, reuse, and distribution are great challenges for the design of the next generation process modeling languages and their enactment mechanisms. Furthermore, flexible and collaborative processes depend also on unpredictable changes and hence require human intervention. Therefore, high-level process modeling constructs are needed which allow for an easy, adequate, and participatory design of workflows. We present a process modeling language which covers these requirements and is based on object-oriented modeling and enacting techniques. In particular, we outline how tasks and task nets are specified at a high level of abstraction, how flexible and user-adaptable control and data flow specifications are supported, and how reuse of workflow models can be improved. The approach is characterized by the uniform and integrated modeling of workflow schema and instance elements as objects and by the integration of flexible rule-based techniques with the high-level constructs of task graphs. Finally, we present our object-oriented approach for the distributed enactment of workflow models: A workflow is directly enacted by task agents which may be treated as reactive components, which interact by message passing, and whose execution behavior is derived from the context-free and context-dependent behavior of the tasks defined in the workflow schema.

Dynamic evolution of workflow models due to process (re)engineering activities and dynamic changing situations of the real process is one of the most important challenges in workflow management. The authors present an approach for the management of evolving workflow specifications which copes with the evolution of a workflow schema and the dynamic modification of workflow instances. The approach is based on the integrated modeling of workflow schema and instance elements, the separated definition of 'what to do' and 'how to do' in the workflow schema, late binding of workflows at run-time, and the versioning of the workflow schema. On this basis, they support lazy, eager, and selective propagation as well as local customization of instances and their upward propagation. Furthermore, they address the problem of managing consistent configurations of the versioned entities of a workflow schema. In their workflow-specific versioning approach, the consistency of the workflow configuration is guaranteed and hence the version mechanism is transparent to the user.

Process modeling and enacting concepts are at the center of workflow manageme nt. Support for heterogeneous processes, flexibility, reuse, and distribution are great challe nges for the design of the next generation process modeling languages and their enactment mechanisms. Furthermore , flexible and collaborative processes depend also on unpredictable changes and hence require human int ervention. Therefore, high level process modeling constructs are needed which allow for an easy, adequate, and participatory design of workflows. We present a process modeling language which covers these requirements and i s based on objectoriented modeling and enacting techniques. In particular, we outline how tasks and task nets are specified at a high level of abstraction, how flexible and user-adaptable control and dat a flow specifications are supported, and how reuse of workflow models is enhanced. The approach is charac terized by the uniform and integrated modeling of workflow schema and instance elements as obje cts. Further characteristics of our approach are the separated definition of ‘what to do’ and ‘how t o do’ in the workflow schema, the versioning of the workflow schema, and the separate definition of context-free and contextdependent behavior of tasks within a workflow. Finally, we present our object-oriented approach for the enactment of workf low models: Instead of interpreting a workflow instance by a workflow engine, a workflow is dire ctly enacted by task instance objects which may be treated as reactive components, which interact by message passing, and which correspond to the defined process steps in the workflow model. The inter action patterns between these task objects are derived from the workflow schema resulting in a scal able and distributed architecture which is realized as a distributed object system.

Business processes are mostly very complex and take place in dynamic and distributed field. Information technology support for these processes requires an adv anced process management system which enhances currently available workflow management service s with respect to integration, flexibility, and adaptation. In order to apply intelligent agent technologie s to process management two different basic approaches are possible: Multi-agent systems with the dom ain of process management, or intelligent agents as a key infrastructure technology for building proce ss management systems. We present an uniform and flexible framework for advanced process managemen t which follows the latter approach and introduce different agent types for managing activities, pr oducts, and resources which cover specific views on the processes.

Software process dynamics challenge the capabilities of process-centered software engineering environments. Dynamic task nets represent evolving software processes by hierarchically organized nets of tasks which are connected by control, data, and feedback flows. Project managers operate on dynamic task nets in order to assess the current status of a project, trace its history, perform impact analysis, handle feedback, adapt the project plan to changed product structures, etc. Developers are supported through task agendas and provision of tools and documents. Chained tasks may be executed in parallel (simultaneous engineering), and cooperation is controlled through releases of document versions. Dynamic task nets are formally specified by a programmed graph rewriting system. Operations on task nets are specified declaratively by graph rewrite rules at a high level of abstraction. Furthermore, editing, analysis, and execution steps on a dynamic task net, which may be interleaved seamlessly, are described in a uniform formalism.

Change management is a core problem of software development. Management of changes means managing the process of change as well as managing all artifacts of an evolving software system. Both challenges have been focused extensively in the field of software process modeling and software configuration management, respectively. In this paper, we motivate that change management needs an integrated approach to process and configuration management. A fundamental prerequisite to provide such a comprehensive support is the integration of the underlying representation formalism of process and version models. To cope with this problem, we propose a conceptual framework to provide a common conceptual basis of process and version modeling concepts and identify the requirements for comprehensive process support for change management. Furthermore, we classify and evaluate different approaches of how to integrate process and version models from a conceptual point of view.

Engineering workflow management has to cope with a great variety of processes. Well structured and a priori defined activities are mixed with creative, less structured, and continuously evolving process steps. The classical coordination support provided by workflow management systems has to be enhanced to cope with dynamic, product centered, and cooperative processes. We concentrate on cooperation support within workflow management on both document level and workflow level. First, we propose an integrated and flexible approach to process and document management which is based on an object oriented modeling framework. We outline how different kinds of processes can be supported on this basis. Finally, we introduce an approach for flexible control and data flow modeling which supports cooperation on the workflow level and takes versioning and different forms of data interchange between activities into account.

Integration of concurrently acting humans and systems is a great challenge which is characterized by the great diversity of involved actors, activities, and artifacts. Based on the process modeling idea , several integration approaches follow a process-oriented and model-based integration strategy. In this paper, we focus on the application of this model-based integration approach to engineering processes and identify the main integration aspects of an integrated process and product management approach. Furthermore, we provide a classification of such integrated approaches on a conceptual level.

Managing the software development and maintenance process has been identified as a great challenge for several years. Software processes are highly dynamic and can only rarely be planned completely in advance. Dynamic task nets take this into account. They are built and modified incrementally as a software process is executed. Dynamic task nets have been designed to solve important problems of process dynamics, deciding product-dependent structure evolution, feedback, and concurrent engineering. In order to describe editing and enactment (and their interaction) in a uniform way, task nets are formally defined by means of a programmed graph rewriting system.

Managing the software development and maintenance process has been identified as a great challenge for several years. Software processes are highly dynamic and can only rarely be planned completely in advance. We present an approach to software process management which is based on hierarchical nets of processes connected by data and control flow relations. Editing and execution of process nets are highly intertwined. Dynamic process nets are formally defined in PROGRES, a specification language which is based on programmed graph rewriting systems. Graph rewriting systems are a natural choice for several reasons. In particular, process nets are complicated graph structures, and editing as well as execution operations may be specified in a uniform way by graph rewrite rules. The graph rewriting system will form the foundation of a sophisticated process management system.