Agent and Multi-Agents Architecture

Artificial Immune Model

We have defined a specific agent architecture based upon Artificial System. The metaphor considers stimulations/requests as antigens and selected antibodies as reactions/answers. Each antibody is activated by specific antigens and stimulated and/or inhibited by other antibodies. The immune system rewards (respectively penalizes) selected antibodies, which constitutes a good (respectively wrong) answer to a request. This mechanism enables an agent to choose from a set of possible behaviors, the one that seems the best fitting for a specific context. This architecture has already been applied to the robot soccer problem and to the problem of power management in a multi-sources and multi-loads context.

Physics-inspired Model

Because of its intrinsic properties (simple agent design, close link with real problem characteristics, easy application to continuous environment...), physics inspired model are one of the most widespread model used to design interaction between agents and with their environment. Using this approach, we defined a model based on artificial potential fields in order to cope with the facility location problem. In this model, situated agents are influenced by their local perceptions and attractive/ repulsive obstacles and destinations generated fields. This model has been applied to transport networks optimization. Another model has been developed to deal with the car platoon problem. This model, based on attraction/repulsion forces, uses a virtual impedance/control (spring and damper) link between vehicles in order to keep safety and secure properties. Virtual link parameters can change in run time taking into account speed and curvature for instance. A formal/physical study of this model has been performed to prove lateral and longitudinal stability of the platoon. This model has been applied in the context of CRISTAL project. Attraction/Repulsion force based models have also been applied to various fields such as localization and tracking problems for mobile robots, obstacle avoidance, path following,...

This page was last modified on 29 July 2010, at 13:46. This page has been accessed 8,358 times.
Copyright 2010-2020 © Laboratoire Connaissance et Intelligence Artificielle Distribu√©es - Université Bourgogne Franche-Comté - Privacy policy