AUML——Projects

来源：http://www.auml.org/

 

Projects

Multiple FIPA AUML Modeling projects are currently underway. Each project is partitioned into various work areas. These projects and their work areas are listed below. This list also indicate those individuals who volunteered to produce the assigned work products. (Anyone wishing to actively participating in any of these project, please send an email to the Contact address at the bottom of this page.) For a schedule of these document, see Schedule & Events.

To download any of these documents, see Working Documents.

Class and Interaction Diagrams specifications

The Modelling TC participants have initially identified two area for developing detailed specifications These specifications and the individuals who have volunteered to work on these areas are as follows:

• Class Diagrams - Specify the internal behavior of an agent and relating it to the external behavior of an agent using and extending UML class diagrams. (Marc-Philippe Huget, Bernhard Bauer, Jim Odell, Renato Levy, Paola Turci, Radovan Cervenka, Misty Nodine, Stephen Cranefield, and Hong Zhu)

• Interaction Diagrams - A generic term that applies to several types of diagrams that emphasize object interactions. These include collaboration diagrams, sequence diagrams, and the interaction overview diagram. (Marc-Philippe Huget, Bernhard Bauer, Jim Odell, Renato Levy, Paola Turci, Radovan Cervenka, and Hong Zhu)

Modeling Areas

The Modelling TC participants have initially identified modelling areas that may be useful for representing and specifying agent-based systems. These areas and the individuals who have volunteered to work on these areas are as follows:

• Multi- vs. single agent - The multiagent level of abstraction looks at several agents together with their relationships and/or interactions. The single-agent level of abstractions deals with one agent at a time. It describes the agent at the low, internal level before coding. For example, in a structural representation we have the agent main-class and the set of classes that realize the agent's tasks/behaviors . (Massimo Cossentino, Paola Turci, and Marc-Philippe Huget)

• Agent “class/component” and implementation structure - (subsumed on the Class Diagrams specification)

• Goal & soft goals - Recommending representations of formalized goals and nonformal "soft goals" that can be useful for specifying agent-based systems. There are several ways in which goals might be expressed and several existing notations. (Radovan Cervenka and Marc-Philippe Huget)

• Use cases - (as specified in UML for now)

• Social aspects - MAS system design can be inspired by human social phenomena. Furthermore, by computationally modelling social phenomena we can provide a better understanding of them. “Social” does not means only organization, roles, communication and interaction protocols, norms (and other forms of coordination and control); but it should be taken also in terms of spontaneous orders and self-organising structures. (Radovan Cervenka, Misty Nodine, Massimo Cossentino, Jim Odell, and Hong Zhu)

• Environment - Without an environment, an agent is effectively useless. Cut off from the rest of its world, the agent can neither sense nor act. An environment provides the conditions under which an entity (agent or object) can exist. It defines the properties of the world in which an agent will function. Designing effective agents requires careful consideration of both the physical and communicational aspects of their environment. (Renato Levy, Radovan Cervenka, Jim Odell, Marc-Philippe Huget, Paola Turci, and Hong Zhu)

• Workflow/Planning - MAS planning has been in the literature for over 15 years. However, there is no standard representation of it. By extending the UML 2.0 Activity Diagram, both planning and a MAS-based workflow approach can be expressed. (Misty Nodine and Jim Odell)

• Services (Subsumed by Environment and Agent “class”)

• Levels of abstraction - Agent systems can be seen by several angles. A possible view is how the system fits within the requirements. A second view would be how the agents in this system coexist, or in other words the basis of their society. A third view is the actual composition of the agent itself. And finally, how these agents are constructed in the implementation sense (classes and data structures). The levels of abstraction, then, indicates clearly in which angle the diagram is to be seen, and furthermore to define how different levels relate across diagrams. (Renato Levy and Hong Zhu)

• Temporal constraints - temporal properties such as how long an interaction must take; see Live Sequence Charts, UTC120) and Domains of time (e.g., real time vs simulation time) (Renato Levy and Keven Kearney)

• Policies (Referenced, but policy-specification language is out of scope for this group)

• Deployment and Mobility - Mobile agent programming has been mainly technology driven, with a focus on implementing mobile agent platforms and small programming applications. This group will work on an extension of UML to provides language concepts for modeling mobility in analysis and design phase. (Paola Turci)

The document format for each of these areas can be downloaded

 

Modeling Notation Sources

The Modelling TC participants have initially identified sources of notations that should be considered for a FIPA AUML. These sources and the individuals who have volunteered to work on these areas are as follows:

• UML 2.0 (Jim Odell)
• AOR (Jim Odell)
• PASSI (Massimo Cossentino)
• MESSAGE (Radovan Cervenka)
• Tropos (includes i* and GRL) (Radovan Cervenka)
• Adelfe (Marc-Philippe Huget)
• Gaia (Alfredo Garro)
• BRIC (Jim Odell)
• Styx (Radovan Cervenka)
• Prometheus (Radovan Cervenka)
• Madkit (Renato Levy)
• OPM (Marc-Philippe Huget)