This work package will study and identify a sound and uniform set of innovative mechanisms and patterns of self-adaptation and dynamic self-expression, both at the service components and service-component ensembles level, properly modelled and experimentally evaluated. From the scientific viewpoint, the identified innovative adaptation patterns will contribute overcoming the traditional distinction between self-adaptation (top-down) and self-organization (bottom-up), and will also contribute mechanisms and techniques for engineering emergent behaviours in complex service-component ensembles.

The first year of activities has adopted a very pragmatic approach. This has included (big red circle in the figure above):

• Defining a robust conceptual and operational framework, namely SOTA (“State Of The Affairs”), that can be used to elicit and rationally represent adaptation requirements. SOTA has also been used as a mean to verify via model-checking, adaptive requirements, and as a mean to elicit and model knowledge requirements.
• Exploiting SOTA as a way to more formally categorize self-adaptation patterns SOTA, and to help choosing the most suitable to meet specific requirements among a catalogue of adaptation patterns.

Based on the above, during the first year we already started studying a first taxonomy of possible adaptation patterns, as well as the possibility of expressing some of these in terms of SCEL (small red circle in the figure above). Yet, we delayed the finalization of the pattern catalogue to the second year.

During the second year, in line with the approach of the first year, the research activities have included the following activities and key contributions (see the green circles in the figure above):

• We have worked towards integrating and harmonizing the SOTA model with the general ensemble models (GEM) being defined in the context of work package 2. Such integration enables to adopt a single conceptual framework both to model the adaptive requirements of ensembles and their dynamical properties.
• We have completed the catalog of self-adaptive patterns that we started compiling during the first year. Such catalog structurally organizes and describes the most commonly adopted architectural patterns for self-adaptive components and ensembles. 
• We have performed extensive experimentations to test the dynamic behavior of such patterns. On the one hand, we have continued working on the robotics simulator to test the behavior and effectiveness of some self-adaptive patterns on specific problems related to the robotics scenario. On the other hand, we have built (over the IBM rational software architectural simulator) a simulator for the dynamic behavior of self-adaptive patterns, and have tested it with some patterns applied to the e-mobility case study.
• The way we have structured the catalog of patterns enables a natural way to identify self-expression patterns (only quickly analyzed in the first year) and the mechanisms needed to dynamically enforce self-expression in components and ensembles. This has enabled us to clarify the set of self-expression patterns and their relations to self-adaptive patterns. However, a deeper analysis on the extent of applicability of such patterns in real-world scenario and on the mechanisms to control their dynamic execution is still missing, and will be performed in the following periods.

Contact: Franco Zambonelli  This e-mail address is being protected from spambots. You need JavaScript enabled to view it