Satellite MAS Position Statement FOSE-MAS 2008 – Danny Weyns
Life on satellite MAS
When we engineer multi-agent systems (1) Do we ask ourselves the question why we select MAS as a solution in the first place? –What are the concrete motivations – what are the tradeoffs? What about the typical non-functional (quality) concerns? –Usability, maintainability, …ility, budget, …. –How do they relate to real stakeholders
When we engineer multi-agent systems (2) What about integrating the MAS with its software environment? –Software programming environment, frameworks, middleware, legacy systems, etc. What about the evaluation of the MAS architecture? –How well will the MAS achieve its functional and quality goals?
When we engineer multi-agent systems (3) What is the impact of introducing MAS on an organization? –Correspondence of software architecture with developing organization
Why are we not really concerned with all these issues that are crucial in practice?
Challenges ahead Each of these issues provides a challenge for FOSE-MAS –Why (not) using a MAS architecture? –Qualities of MAS, tradeoffs –Integrating MAS with its software environment –Evaluation of MAS architectures Verification of system-wide properties of decentralized architectures –Impact of MAS on the developing environment
Earth is calling satellite MAS To tackle the increasing problems of software systems, mainstream SE needs our expertise! –E.g. to develop self-managing systems Grounding MAS engineering in mainstream SE is a key to (industrial) adoption of MAS