The objective of this activity consists in refining the problem ontology described during POD activity by adding new concepts related to the agent-based solution and refining the existing ones. Concepts, predicates and actions of this ontology are now also intended to be used for describing information exchanged during communications among roles. This implies the definition of all the predicates that are used to exchange knowledge in communications as well as the actions that can be done by Holons/Agents and affect the status of the world they live in (as represented in ontology by concepts). The introduction of actions in ontology is not new and is compliant with a FIPA specification [1].
The presence of actions in our ontology allows to model the complete knowledge space of our autonomous entities, in terms of the concepts they can understand, the predicates they can assert about the status of those concepts, and the actions they can perform/conceive in order to affect the status of concepts.
This activity, although positioned at the beginning of this phase, is not supposed to be executed once. Rather because of our iterative/incremental design approach, this activity will probably be executed several times. The need for new concepts, predicates and actions can arise at any moment in the design activities and can justify an iteration aimed at improving ontology with the desired new elements.
Contents |
Goal
The objective of this activity consists in refining the problem ontology described during POD by adding new concepts related to the agent-based solution and refining the existing ones. Concepts, predicates and actions of this ontology will also be used to describe information exchanged during communications among roles.
Input
The solution ontology is firstly deduced from the problem ontology. Role plans and capacity identification allow to refine the needs of each role and organization, and thus allow the identification of new concepts.
Output
The ontology is described in terms of concepts, predicates, actions and their relationships. And a stereotype for class diagram is used to represent an ontology.
MAS Meta-Model Elements
Define(Ontology), Define(Predicate), Define( Action), Define(Concept), Relate(Ontology,(Concept, Action, Predicate)), Relate(Concept,Action), Relate(Concept,Predicate)
Work to be done
The first task consists in refining existing concepts description and identifying new ones (related to the agency-level solution that is under development). Actions and predicates should also be added.
Methodological Guidelines
For concepts identification, a guideline consists in looking into previously identified organizations, the hierarchical composition of two organization generally hides a composition between concepts. This can also be done by looking into role plans and scenarios. Identification of actions and predicates can be facilitated by results of the capacity identification activity. As we have already said, if a capacity deals with some ontological knowledge, manipulated concepts should be connected to an action in the corresponding ontology. Moreover the solution ontology will also be exchanged in agent communications; an indication on which knowledge will be necessary to roles behaviors and in which activities it will be used can be found in role plans and scenarios. From these, actions and predicates can generally easily be identified. Of course this description of the ontology is also an iterative process, and it is generally refined during the communication ontology description (the second activity of the Agent Society Phase).
References
- [1] Fipa Rdf Content Language Specification
- Foundation For Intelligent Physical Agents.
Experimental, XC00011B, 2001.
