The DIET project proposes to specify, design, implement and validate a novel information processing and management framework via a bottom-up, ecosystem approach. DIET would provide a set of new technologies including a complex information ecosystem tool suite with infohabitants exhibiting self-organising and emergent behaviour. The whole ecosystem simulation will be validated by getting different societies of infohabitants to interact within a rich information environment (the WWW) whilst providing useful information management services for human infohabitants. DIET would also provide intelligent, interactie simulation and visualisation tools to facilitate understanding the complex interactions in the ecosystem. DIET ultimately aims to develop technologies which will replace brittle, expensive, unadaptive software with robust, inexpensive and adaptable solutions to ever increasing complexity in a more networked era.
To study, implement and validate a novel information processing and management framework via a "bottom up" and ecosystem-oriented design approach leading to an open, robust, adaptive and scalable environment.
This will involve:
1. The conceptualisation, analysis, design and construction of an architecture/platform and environment in which automated infohabitants and their societies can exist;
2. To research into the effects of alternative forms of interaction among different species of infohabitants under an ecological;
3. Finally to validate and demonstrate the usefulness of the platform and the infohabitant societies included in it, via four tasks. These will include information retrieval, filtering, mining, and trading. These objectives will contribute to coping with the increasing complexity of the current information society, and towards new approaches to computation and control based on ecosystems.
DESCRIPTION OF WORK
The project will involve the theoretical study, implementation and validation of a novel information management framework which will use ecosystem metaphors. Initially this will involve the design of an overall framework in which infohabitants - entities which can process information; - can interact and coexist in societies set in an information environment.
The framework will permit a focus on emergent properties, flexible interactions, adaptation and self-organisation. We will validate the architecture/societies/associated DIET claims by experimenting on four tasks:
1. an information retrieval task which will instantiate an "informaton pull" scenario;
2. an information alerting task which will instantiate an "information push" scenario;
3. a trend analysis task which will instantiate a data mining scenario; and finally;
4. an economic interaction task. This will implement various economic interactions between infohabitants. In order to address these tasks, the architecture will be developed to group infohabitants in a hierarchy of societies with differing properties: harvesting societies, containing infohabitants concerned with information gathering; mining societies, containing infohabitants concerned with processing of information gathered; and brokering societies, containing infohabitants concerned with facilitating services between other societies/infohabitants. These structures will be developed in conjunction with a set of interactive and visualisation tools, which will allow the application of the completed software package to a wide range of real-world problems, assisting the further exploitation of this research.
Work package 1 (overall software framework): Construction of a decentralised, scalable, lightweight and robust agent platform that provides a basis for experimentation and application development throughout the rest of the project.
Work package 2 (Information retrieval, filtering and mining): Development of information processing layer (I-Gaia) based on the DIET platform, and extension of this to carry out information push/pull, meta-information processing and information mining activities.
Work package 3 (Information brokering): Investigation of data models and query languages for information brokering. Development of distributed peer-to-peer architecture for information brokering, and of applications based on brokering between agents.
Work package 4 (Economic interaction): Development of algorithms for fair resource allocation in peer-to-peer systems, and investigation of related models of economic interactions between agents. Development of algorithms to regulate agent behaviour through resource accounting, inspired by economics and/or ecology. Development of bartering mechanism within I-Gaia information processing layer, to enable access to meta-information.
Work package 5 (Visualisation): Construction of intelligent and interactive simulation, control and visualisation tools for use in development and debugging using the DIET software platform.
Work package 6 (Exploitation and dissemination): Dissemination of project activities through workshops, conference and journal publications and presentations. Development of demonstration applications running on the DIET platform and drawing upon research results from other work packages. Investigation of exploitation directions for software platform and applications. Open sourcing of DIET software platform. Investigation of exploitation directions for DIET.
Work package 7 (Project management): Management of collaboration between partners, organisation of regular meetings, production of deliverables, and of reporting of project results.
Funding SchemeCSC - Cost-sharing contracts
28903 Getafe (Madrid)