Bibliographie

Prototypische Realisierung einer Integration von Agenten und Workflows

Thomas Wagner.
Prototypische Realisierung einer Integration von Agenten und Workflows.
Diplomarbeit, Universität Hamburg, Department Informatik, Vogt-Kölln Str. 30, D-22527 Hamburg, 2009.

Abrégé: Agent technology can be used to specify the structure of a software system by defining a set of entities. These entities, called agents, provide the entire functionality of the system. They possess certain characteristics, like autonomy and encapsulation, which facilitate their use in distributed software systems.
Workflow technology, on the other hand, can be used to describe the behavior of a set of interacting entities within a system. These entities, called resources, execute tasks and exchange data according to a predefined process. Systems, designed with workflow technology, are used to support the business processes of a company.
It is possible to design large and complex systems with both technologies. They both contain a strong focus on different aspects of a system. While agent technology focuses on the structure and distribution of the different entities, the workflow technology focuses on the behavior and dataflow between them. The integration of these two, almost opposite, views is the main objective of this contribution. Workflow technology will be used to enhance agent technology and vice versa. The integration is based on the theoretical, overall architecture proposed in [Reese09]. This architecture consists of five tiers. Beginning with regular agent or workflow systems, each tier introduces new concepts, which integrate both technologies more and more.
This contribution describes a number of prototypical implementations, which can be classified within the overall architecture. The prototypes are based on the Mulan and Capa agent architectures and use workflow nets to model processes. The first system, which is described, is an agent-based Workflow Management System (AgWFMS). It can be classified as a system of the second tier of the overall architecture. This AgWFMS provides workflow-functionality by solely relying on agents and is used as the technological basis for the other two prototypes. These other prototypes are the structure-agent-workflow (S-AgWf) and the protocol dispatcher (ProDi) system. The functionality needed by both systems is provided by a combination of the AgWFMS and the Mulan and Capa agent architectures. This combination, though not a separate prototype, can be classified as a system of the third tier of the overall architecture. These S-AgWf and ProDi prototypes represent the two variations of the fourth tier of the overall architecture. For application development the S-AgWf system offers the capabilities of a workflow management system. Internally though, the system strongly relies on agent technology. Through this use of agent technology the system benefits in certain areas, for example distributed workflow execution. The ProDi system, on the other hand, offers the capabilities of an agent management system, which can be used to design and implement agent applications. Internally, it uses workflow technology to model and support the interactions between the different agents of the system. Because of this, the interactions are more flexible and the system becomes more fault-tolerant, compared to the regular execution of interactions.
All prototypes are described using the Paose-development approach and are discussed in detail. Practical examples of their use are given.

BibTeX enregistrement

@MASTERSTHESIS{Wagner09a,
  author = {Wagner, Thomas},
  title = {{Prototypische Realisierung einer Integration von Agenten und Workflows}},
  school = FBIUniHHab2006,
  year = 2009,
  type = diplomarbeit,
  address = FBIUniAdresse,
  abstract = {Agent technology can be used to specify the structure of a software
 system by defining a set of entities. These entities, called agents,
 provide the entire functionality of the system. They possess certain
 characteristics, like autonomy and encapsulation, which facilitate
 their use in distributed software systems.

 Workflow technology, on the other hand, can be used to describe the
 behavior of a set of interacting entities within a system. These
 entities, called resources, execute tasks and exchange data according
 to a predefined process. Systems, designed with workflow technology,
 are used to support the business processes of a company.

 It is possible to design large and complex systems with both technologies.
 They both contain a strong focus on different aspects of a system.
 While agent technology focuses on the structure and distribution
 of the different entities, the workflow technology focuses on the
 behavior and dataflow between them. The integration of these two,
 almost opposite, views is the main objective of this contribution.
 Workflow technology will be used to enhance agent technology and
 vice versa. The integration is based on the theoretical, overall
 architecture proposed in [Reese09]. This architecture consists of
 five tiers. Beginning with regular agent or workflow systems, each
 tier introduces new concepts, which integrate both technologies more
 and more.

 This contribution describes a number of prototypical implementations,
 which can be classified within the overall architecture. The prototypes
 are based on the Mulan and Capa agent architectures and use workflow
 nets to model processes. The first system, which is described, is
 an agent-based Workflow Management System (AgWFMS). It can be classified
 as a system of the second tier of the overall architecture. This
 AgWFMS provides workflow-functionality by solely relying on agents
 and is used as the technological basis for the other two prototypes.
 These other prototypes are the structure-agent-workflow (S-AgWf)
 and the protocol dispatcher (ProDi) system. The functionality needed
 by both systems is provided by a combination of the AgWFMS and the
 Mulan and Capa agent architectures. This combination, though not
 a separate prototype, can be classified as a system of the third
 tier of the overall architecture. These S-AgWf and ProDi prototypes
 represent the two variations of the fourth tier of the overall architecture.
 For application development the S-AgWf system offers the capabilities
 of a workflow management system. Internally though, the system strongly
 relies on agent technology. Through this use of agent technology
 the system benefits in certain areas, for example distributed workflow
 execution. The ProDi system, on the other hand, offers the capabilities
 of an agent management system, which can be used to design and implement
 agent applications. Internally, it uses workflow technology to model
 and support the interactions between the different agents of the
 system. Because of this, the interactions are more flexible and the
 system becomes more fault-tolerant, compared to the regular execution
 of interactions.

 All prototypes are described using the Paose-development approach
 and are discussed in detail. Practical examples of their use are
 given.}
}

Copyright Notice

This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.