Meeting 24/08/2011
Minute of the Meeting
Who
- Jose Santiago Perez Sotelo
- Andrea Omicini
Where
- Room 51, DEIS 1st Floor, viale Risorgimento 2, 40136 Bologna, Italy
When
- 24/8/2011, h. 12:30
What
- Seminar Towards Adaptive Service Ecosystems with Agreement Technologies (internal talk @ Internal Meeting, DEIS, Bologna, 24/08/2011) — José Santiago Pérez-Sotelo (José Santiago Pérez-Sotelo, Carlos E. Cuesta, Sascha Ossowski), by Jose Santiago Perez Sotelo
Details
Jose Santiago Perez Sotelo starts presenting research by himself and his group (link), focussing on Agreement Technologies for Adaptive Service Ecosystems.
During presentation, many points are discussed, mainly to clarify the main concepts and also foresee possible common topics of interest.
In particular (in no particular ordering), the following issues are raised and discussed:
- The notion of ecosystem is apparently key in Agreement Technologies. It is also a key-notion in the SAPERE Project: a comparison between the two is clearly due.
- The “Agreement Technologies Tower” (with Semantics, Norms, Organisation, Argumentation & Negotiation, Trust — bottom up) actually contains much more than the labels say. For instance, reputation is not a "part" of trust but is conceptually placed at the same level, and communication & dialogue clearly belong to argumentation & negotiation. This should be clarified.
- Also, we did some work where the relation between dialogue and argumentation is mediated by the A&A meta-model — see for instance Co-Argumentation Artifact for Agent Societies, Argumentation and Artifacts for Negotiation Support.
- The notion of organisations as first-class entities allow them to work as agents looking for agreements. This leads to a potential multilevel structure of MAS, where agents represent the leaves at the bottom level, the MAS is the top-level as a whole, and OU (organizational units) "fill" the intermediate levels.
- this is clearly related with some notions of #soda(): agent society as first class entity (SODA: Societies and Infrastructures in the Analysis and Design of Agent-based Systems), with zooming (SODA: A Roadmap to Artefacts), and the related layering mechanism for facing complexity of MAS engineering ((Zooming Multi-Agent Systems).
- The notion of adaptivity clearly stems from the idea that local interaction, in terms of achievement of local agreements, produces a global re-organisation which fits more the overall application needs. From the case studies clearly emerge the difference between a potentially centralised solution and a self-adaptive one — which is based on the idea (Ex. 1) that organisations can directly negotiate and reach agreements in a dynamic way.
- The discussion about controls and protocols, with the fourth point on organisations adaptively changing their goals, resembles somehow Activity Theory's modelling of collaboration activities as co-ordination, co-operation and co-construction: see for instance Activity Theory as a Framework for MAS Coordination.
- Dynamics of agreements is essential for adaptivity: they should include some notion of time and validity to "vanish" at some time.
The discussion ends at 14:30 with the definition of the tasks.
Tasks
- for Andrea
- preparing the first version of the MoM (done)
- adding references for José to read (done)
- preparing a presentation for next meeting to introduce our research framework (done)
- for José
- check the MoM (done)
- read papers and references (done)
- preparing a thesis outline (in progress)
- Agreement Technologies for teaching: some notes (done)
What's Next
Next meeting is scheduled for Monday, 29th of August 2011.