The Agent Realization And Intelligent Supervision Of Home Energy System Based On MAS

Along with social information's fast development, people's work and life connect with communication and information close increasingly. Information society changes people's life style and working custom, at the same time also challenges conventional home. People's concept changes a lot along with the progress of society, technology and economy. And then, people's requirement of home is physical space no longer, but an intelligent, safe, convenient and comfortable home is even more attentive. In order to satisfy the demand, the conception of intelligence is brought in home. Therefore, the conception of intelligent home is generated. Intelligent home can be defined that takes dwelling house as flat roof, combines architecture, network communications, equipment automatization and information home electrical equipment, integrates system, stricture, service and supervision, realizes habitation which is highly active, comfortable, safe, convenient, propitious environment and save energy and recreational roboticized system.Intelligent Agent technology as one of artificial intelligence fruits is gradually introduced by scholars in China and foreign countries and becomes a new direction of discussing home intelligentized. The original intention of Agent is"surrogate"or"instrument", and Agent goes by the name of"intelligent main block"or"intelligent surrogate", it is a sort of computing entity which can simulate human being's behavior, has vicegerent property, intelligent property, and is able to provide corresponding service. Multi-Agents which have different aims are combined rationally, that is to say, Mufti-Agent System (MAS) are formed. MAS as the typical appliance of distributed artificial intelligence, has the mode of operation which needs cooperation according to require, incompact coupling, these traits accord with the demand of the intelligent home system completely. Therefore, in the thesis, home energy appliance integration system as one of intelligent home system is built on base of MAS technology. Home energy equipment not only obtains intelligent control but also can gain the ends of saving energy.In the thesis, the deficiency of current home energy equipment centralized control's settling scheme is analyzed. Multi-Agent-based Home Energy System (MAES) are built combining multi-Agent technology, and the primary composing structure of the model is given. At the same time, diversified Agents of the model are concluded, and the function, structural model and implementation mode of diversified Agents are designed.In this thesis, the problems of multi-Agent communications and learning in Home Energy System as emphases are studied. Existing multi-Agent communication mode and communication language are studied systematically first of all, and on base of it, combining the practical situation of home environment, multi-Agent communication model based on communication server mode are put forward. And then, Agent learning problems are studied deeply. In the home energy system, the learning capacity of some Agents which have special functions is increased, and the performance of the system will be improved. And then, a sort of multi-Agent cooperation learning arithmetic based on long-term benefit is put forward. This arithmetic takes long-term beneficial estimated value as behavioral selective policy. The behavior corresponding to maximal estimated value is the best one currently. Because this arithmetic makes every Agent consider sequential estate and behavioral gain, and take the long-term gain of team work as learning clue, the best excellent team work policy is more propitious to be searched. In this thesis, this arithmetic is applied to the constant temperature system of water heater. The results of the simulation experiment indicate that this arithmetic makes learning process build on the better learning foundation, consequently approach the best excellent estate much more swiftly, and achieve the aim of quick constant temperature.The study on home energy system based on MAS technology benefits the analysis and realization of intelligent home system, can settle the problems of plug and play in intelligent home, at the same time is useful to reduce the cost of home group network, and help to quickening the establishment of intelligent home's standards.