Controlling residential co-generation system based on hierarchical decentralized model

This paper deals with a residential co-generation system as an application of the hierarchically decentralized model. We assume the residential co-generation systems consist with autonomous co-generation units that behave autonomously. It is important in controlling co-generation systems to forecast quantity of energy demands, and thus we use Neural Network to forecast them. We propose a simulation model of the residential co-generation system and a model of agents which control co-generation unit autonomously. To simulate the model, we prepare artificial demand data. Through several computer simulations, effectiveness of sharing information between co-generation units is investigated.