Towards filling the gap between AOSE methodologies and infrastructures: requirements and meta-model Fabiano Dalpiaz¹, Ambra Molesini², Mariachiara Puviani³, Valeria Seidita⁴ ¹ Università degli Studi di Trento ² Alma Mater Studiorum ³ Università degli Studi di Modena e Reggio Emilia ⁴ Università degli Studi di Palermo WOA08 - Palermo Outline • Introduction to MEnSA • Our approach – – – – – Assembling a Metamodel Requirements Selection of fragments Conceptual map Metamodel • Conclusions and future work F. Dalpiaz, A. Molesini, M. Puviani, V. Seidita 2 The MEnSA project: why? • The development of complex systems requires a new software engineering paradigm – Agent-oriented methodologies – Paradigmatic shift (from OO) at conceptual and technical level – No need to reinvent the wheel – There are many agent-oriented methodologies – Each methodology has different specificities and application areas F. Dalpiaz, A. Molesini, M. Puviani, V. Seidita 3 The MEnSA project: what? • MEnSA – “Methodology for the Engineering of complex Software systems: Agent based approach” – Filling the gap – analysis and design, and implementation – methodologies and infrastructures – Metamodel-based approach – Integration of existing fragments • 3+1 partners – Alma Mater Studiorum (Cesena), Università degli Studi di Modena e Reggio Emilia, Università degli Studi di Trento, ICAR-CNR Palermo e Università degli Studi di Palermo. F. Dalpiaz, A. Molesini, M. Puviani, V. Seidita 4 Our Approach 1. Definition of a set of requirements for the methodology – – Generic requirements Specific requirements derived from the selected methodologies 2. Elicitation and analysis of a set of fragments that satisfy the requirements – starting from Tropos, GAIA, SODA and PASSI F. Dalpiaz, A. Molesini, M. Puviani, V. Seidita 5 Our Approach 3. Definition of a semantic conceptual map – To find out synonyms and inter-level relations between concepts from different abstraction levels 3. Assembly of an integrated metamodel on the base of the selected fragments F. Dalpiaz, A. Molesini, M. Puviani, V. Seidita 6 General Requirements • Transformational approach: from high-level abstraction to low-level concrete entities • Support for traceability • Support for functional and non-functional requirements • Support for goal-oriented and functional-oriented analysis • Precise and compact modeling constructs for the concept of agency – Agent, Agent’s rationale, Situated agent, Social Agent F. Dalpiaz, A. Molesini, M. Puviani, V. Seidita 7 Specific Requirements (1) • Transformational process – Requirement agent → Design agent → Implementation agent • Layering (supported by SODA) – Zooming and Projection mechanisms • Goal oriented analysis (Tropos) before functionaloriented analysis (Passi) • Interaction – Semantic communication + Ontology + compliance with FIPA ACL F. Dalpiaz, A. Molesini, M. Puviani, V. Seidita 8 Specific Requirements (2) • Organizational Rules (Supported by GAIA) • Environment and topology modeling – SODA artifact and workspace • Modeling of non-functional requirements – Tropos soft-goals • Modeling of Agent Plan – Should not constrain to a specific kind of agent F. Dalpiaz, A. Molesini, M. Puviani, V. Seidita 9 Assembling a meta-model • Composition patterns: – Elements from existing meta-models present the same name but have different meanings – Elements have the same meaning but different names – Elements present totally disjoint names and definitions, requiring just a simple composition • Additional concepts and relations act as glue • Outcome: Conceptual map + Glossary of term F. Dalpiaz, A. Molesini, M. Puviani, V. Seidita 10 The Conceptual Map Horizontal relations link synonym concepts Vertical relations define inter-level links (realization) F. Dalpiaz, A. Molesini, M. Puviani, V. Seidita 11 MEnSA metamodel F. Dalpiaz, A. Molesini, M. Puviani, V. Seidita 12 MEnSA metamodel: requirements F. Dalpiaz, A. Molesini, M. Puviani, V. Seidita 13 MEnSA metamodel: design F. Dalpiaz, A. Molesini, M. Puviani, V. Seidita 14 Conclusion • A process to define an integrated agent-oriented meta-model – Identification of the requirements for the target methodology – Selection of a list of fragments from the four considered methodologies. – Construction of a glossary – Definition of a conceptual map of methodologies abstractions • Initial version of the meta-model F. Dalpiaz, A. Molesini, M. Puviani, V. Seidita 15 Future Work • Definition of the meta-model’s implementation phase, extracted from a set of MAS infrastructures • Refine the metamodel – as a result of the work on the methodological aspects and the validation phase over a case study – splitting the two phases into different and more detailed sub-phases F. Dalpiaz, A. Molesini, M. Puviani, V. Seidita 16 Thanks for your attention [email protected] [email protected] [email protected] [email protected]