WP1: Meta-model
WP1: Meta-model
Leading Unit: UNITN
Most agent-oriented methodologies are supported by small teams of academic researchers, and as a result, most of them are in an early stage and still in the first context of mostly "academic" approaches for agent-oriented systems development. Moreover, such methodologies are not well documented and very often they are defined and presented as result of a research paper focusing on specific aspects of the methodology. Meta-models are useful for specifying the concepts, rules and relationships used to define methodologies. Although it is possible to describe a methodology without an explicit meta-model, formalising the underpinning ideas of the methodology in question is valuable when checking its consistency or when planning extensions or modifications. A good meta-model must address all of the different aspects of methodologies, i.e. the process to follow, the work products to be generated and those responsible for making all this happen. In turn, specifying the work products that must be developed implies defining the basic modelling building blocks from which they are built.
Meta-models are extremely important to integrate methodologies. Indeed, merging methodologies without a common universal meta-model as a guide on the scope of SDLC (software development lifecycle) concepts can lead to one of two types of errors: assuming differences of concern exist when none exists, or falsely assuming similarity of concern because of common use of terms. The first type of error may lead to repetition and lead to an unnecessarily large and cumbersome methodology, rendering it less accessible to developers. Tolerating errors of the first type, a successful unification effort would result in a large methodology with its bulk concerned with a collection of "exceptional cases" without common structures. We find that this is exactly what happened with UML. The second type of error can create inconsistencies because of inconsistent interpretations of terms. Tolerating such errors, the resultant methodology would produce inconsistent models and lower its usability, as software developers subsequently struggle to deal with problems resulting from inconsistencies and would most likely lead to its abandonment. The two errors are more likely to occur for areas of concern in the SDLC which are not well known and uncommon, where agreement on concepts and approaches is less likely.
The main objective of this workpackege is to define a meta-model that integrates the meta-models of the major agent-oriented methodologies with abstractions and extensions introduced by multi-agent languages and tools. In particular, the activities will be organised in three main activities as described in the following.
A1.1: Study of the state-of-the-art
In this activity we will study the state of the art regarding agent-oriented methodologies and developing tools. The main objective of this activity is to understand and clarify concepts, rules and relationships used in each methodology. In other words, we will study the meta-model associated each methodology and we will compare the methodologies on the base their meta-models. Whenever not available, we will try to reconstruct the meta-model of the examined methodology. The methodologies that we will take in consideration include methodologies with a full direct descendancy from full OO methods (such as GAIA, MESSAGE, INGENIAS, AOR, Prometheus and PASSI) and other methodologies that do not acknowledge any direct genealogical link to object-oriented approaches (such as Tropos and SODA). On the other hand, we will focus on multi-agent systems infrastructures such as JADE, TuCSoN, MARS and TOTA that represents the state of art and are object of research for the groups involved in this project.
— Expected Results
The results of this phase will be very important for the second activity of this workpackage when we will define the meta-model of an integrated methodology. We will produce scientific papers, including critical review of the examined meta-models and comparison of methodologies based on the examined meta-models.
— Begin/End
- month 1 - month 6
— Deliverables
- D1.1 Document describing the state of art about agent-oriented methodologies and multi-agent infrastructures.
A1.2: Definition of an integrated meta-model
This second activity will consist on the definition of an integrated meta-model. Starting from the results produced in the A1.1 activity, the most appropriate meta-models "fragments" will be selected and possibly adapted from existing methodologies and multi-agent systems infrastructures to form the integrated meta-model. Roughly, the meta-model will describe the concepts and notions used by the methodology, the relations between these concepts, the activities carried on in the methodology and the whole process of the methodology. Particular attention will be given to the integration of the methodologies' meta-models with the infrastructures abstractions. The final meta-model will be the integration of at least four major methodologies (likely Tropos, Gaia, PASSI, and SODA) and two multi-agent systems infrastructures (likely JADE and TuCSoN).
To represent a meta-model we need an appropriate framework (or standard) that is generally referred itself as a meta-model. In literature there are many proposals and choosing among them will be another objective of this phase. Our evaluation will mainly focus on OPEN Process Framework (OPF) and OMG SPEM (Software Processing Engineering Meta-model), which are the most used ones.
— Expected Results
The results of this phase will be very important for the other workpackeges of the project, and in particular to the WP2 and WP3 where we will develop the methodology and the agent infrastructures. We will produce scientific papers in which we propose the meta-model to the agent-oriented software engineering community.
— Begin/End
- month 3 - month 12
— Deliverables
- D1.2 Document describing the meta-model
A1.3 Comparison with methodological framework
The research units will contribute to develop the integrated methodology starting from the meta-model proposed in the previous phase. The main objective will be to guide the development of the methodology using the ideas captured in the meta-model and possibly revise the meta- model on the feedback that we will receive from the construction of the methodology. In particular, we will pay particular attention on how our hypothesis of integration between methodologies and infrastructures can work in practice and possibly revise our hypothesis.
Another important objective of this phase will be the verification of the conformity with the meta-model of the implementation of the supporting tools that the other partner will develop in the other phases of the project. This will be done, first of all, using the meta-model to guide the design of the tools and then by means of test cases. The meta-model will be possibly revised after the verification process.
— Expected Results
The result of this phase in the production of internal reports to share our results with the other partners of the project and scientific
— Begin/End
- month 6 - month 24
— Deliverables
+ D1.3 Document presents the comparison between the meta-model and the developed methodology
A1.4: Evaluation and Revision of the Meta-model
Final evaluation with case studies (WP4). Our objective in this activity will be the evolution of our meta-model on the base of the effective usability of the methodology and its supporting tools in a case study. Particular emphasis will be deserved to understand and evaluate the effective utility of integrating in a methodologies concepts and methods typical of agent-oriented methodologies with abstractions introduced by infrastructures and developing tools. Of course this will be an evaluation that will be done jointly with the other partners and will be the major results of our research.
— Expected Results
We will produce scientific papers to divulgate the results of our work. We are confident to publish such papers to the major (AO)SE conferences and international journals on the topic.
— Begin/End
- month 12 - month 24
— Deliverables
- D1.4 Document describing the evaluation of the metamodel and the description of its revised version.