Agent-based architecture for designing hybrid control systems

When designing very complex control strategy using hybrid technology, one usually faces the challenge of balancing effective realization of multi-control modeling with design simplicity. To better manage this difficulty we have used the agent paradigm as a simple and powerful bridge between asynchronous/distributed computation and Matlab environment. The proposed architecture has been used to design a complex hybrid control environment using multi-objective, fuzzy c-means, and genetic algorithms optimization to design hybrid control strategies suitable for the energy flows management on board of hybrid electric vehicles.