Autonomic Electronic Institutions
Agent-based technologies were originally focused on the features that enabled individual agents to interact, namely agent rationality and communication. The appearance of multi-agent systems (MAS) brought forward the social component of the interactions of groups or populations of agents. Although emphasis in MAS research has largely been put in the internal architecture of those agents and the communication capabilities, issues like task allocation, cooperative and competitive behavior, coordination mechanisms and their constitutive elements have gradually become part of the mainstream.
The notion of electronic institution (EI) focuses on the environment where agent interactions take place and structures them mirroring the way traditional institutions have articulated interactions among human agents. EIs have been conceived as coordination artifacts where the social elements of the coordination are made explicit and the subjective elements are left to the individual agents that join the electronic institution. The term electronic institution was coined at the IIIA to refer to the computational analogue of traditional institutions. Those concerns and ideas behind that proposal have motivated alternative and complementary developments. Nowadays, the notion of virtual organizations, and electronic institutions in particular, has become one of the standard topics of interest in MAS. In the meanwhile the original IIIA proposal has evolved over several years and is stable enough to deserve revisiting its foundations and goals.
It should be noted that, so far, research on EIs (not only at IIIA) and on virtual organisations has focused on structural aspects of EIs and the state of the art is limited to the proposal of languages and architectures to implement centrally designed static institutions. Even though these efforts have had practical and interesting outcomes, we may now go beyond by designing institutions that work under more dynamic conditions. This endeavour aligns with a recent trend in the development of information systems. Indeed, the growing complexity of advanced information systems in the recent years, characterized by being distributed, open and dynamical (and thus, EIs can be regarded as a particular case), has given rise to interest (present in computation since Charles Babbage's work) in the development of systems capable of self-management.
Therefore, the main goal of the IEA project consists in studying how to endow an electronic institution with autonomic capabilities that alllow it to yield a dynamical answer to changing circumstances, through the adaptation of its interaction conventions. Among all the characteristics that define an autonomic system, the IEA project focuses on the study of self-configuration and re-configuration as pointed out in [49] as a second characteristic:
An autonomic computing system must configure and reconfigure itself under varying (and in the future, even unpredictable) conditions. System configuration or "setup" must occur automatically, as well as dynamic adjustments to that configuration to best handle changing environments.
In particular, in the IEA projet we intend to pursue the following goals:
- To elaborate self-configuration models for electronic institutions able to automatize the description, discussion and agreement on the interaction conventions interaction conventions (norms or behaviour rules) to be used by a set of autonomous agents. To this end, we will investigate: (i) efficient interaction mechanisms, (ii) trust models and (iii) flexible negotiation models based on argumentation and coping with imprecise, uncertain or contradictory information.
- To elaborate reconfiguration models for electronic institutions that allow the adaptation of their interaction conventions under varying conditions in order to guarantee the fulfillement of the goals they were designed for. To this ens, we will study efficient algorithms for learning models under different working conditions.
- To propose new protocols for combinatorial auctions making use of more information in the auction mechanism and to develop new scalable algorithms to determine the solve the winner determination problem when the problem is over-constrained due to constraints imposed by the auctioneer.
- To implement a development environment for autonomic elctronic institutions. Such development environment is conceived as an extension to EIDE, the development environment for electronic institutions developed at IIIA. The new environment will allow the specification, simulation and execution of autonomic electronic institutions.
- To develop two proof-of-concept prototypes, based on the development environment for autonomic electronic institutions, for the designed mechanisms of self-configuration and re-configuration.
To summarise, the main goal of the IEA project is to make headway in the next generation of electronic institutions, namely those endowed with self-* capabilities, where the * sign indicates a variety of properties: self-organization, self-configuration, self-diagnosis, self-repair, etc.
