| Popis: |
The problem of energy management optimization deals with heterogeneous multi-services having different types of models and divergent preferences, it is distributed by nature. Centralized optimization approaches bring advantages such as optimal possible solution but are not able to take into account this kind of service representation which forces its resolution to be distributed as well. For such needs, solutions based on Multi-Agent Systems (MAS), well suited to solve spatially distributed and opened problems, were proposed to solve this problem. Although the control of the emergent behavior of multi-agent systems is extremely complex, the performance of the centralized system is when possible better than that obtained with distributed systems. An alternative approach, noted mixed approach, proposes a formulation to combine the centralized solving approach for energy management problem in homes with a multi-agent solving system. Recently, Reinforcement Learning (RL) has also been used in the domain of building energy management, especially on Heating, Ventilation, and Air Conditioning systems (HVAC) but just inside simulated environments. |
