هشتمینفرانس کن ملیجمن اندهی و کنترل فرمان( C4I ) ايران ونم و فننشگاه علو تهران، دا هوايید ستاری شهی، آذرماه9313 C4I2014-PO237 Control Based on Multi-agent System for Distributed Generation Mohammad Reza Vaezi Islamic Azad University Student of Department of Control Engineering, Damavand Branch Damavand,Iran Rezawaezi@gmail.com Reza Ghasemi,Ali Akramizadeh Islamic Azad University Joined the Department of Control Engineering, Damavand Branch Damavand,Iran Reghasemi@gmail.com, akramizadeh@gmail.com Abstract— Islanded microgrid is an independent micro power system. The system has different kinds of distributed generations with different characteristics. Frequency & Voltage stability of microgrids under islanded operation attracts particular attention recently. The islanded microgrid makes its frequence & Voltage control more difficult than the conventional grids. This presents a multi-agent based frequency & Voltage control approach for islanded microgrid and discusses the characteristics of agents. A new cooperative frequency & Voltage control strategy based on centralized multi-agent system (CMAS) is proposed in this study.On This strategy, agents sent data and furthermore each component has its own to center operating decisions (MGCC). After deciding on the information, they are returned. Frequency & Voltage control strategies include primary and secondary frequency control and disposal of multi-stage load in which this study will also provide a method and algorithm for load shedding based fuzzy method.This could also be a big problem for the performance of micro-grid in times of disaster.The simulation results show the promising performance of the proposed structure of the controller based on multi agent systems. I. INTRODUCTION The electric grid is evolving toward what has been defined as the “smart grid paradigm”.[1]. The development of communication infrastructures provides power electronics interfaces with the ability to control complex power systems in efficient and scalable ways and in real time. Multi-agent systems (MAS) are based on distributing information and computing algorithms for complex networks, and are an excellent technological solution for this application.This paper focuses on applications of MAS in power systems and describes how they can be used with other artificial intelligence techniques in order to make the grid smarter and more flexible. In addition to presenting the basics of multi- agent theory, this chapter covers some design procedures and provides several examples, as well as perspectives for future developments of MAS in power systems control. One of the most important parameters is to control the microgrid frequency & Voltage. The change represents a change in the rate of production and consumption And therefore a very important factor in the operation and control of the microgrid. The operation of a network, information on the frequency & Voltage of which shall be recorded the moment and this means that a large volume of information in a record time limit should be examined. It should be noted that the change in frequency & Voltage of more than nominal value in addition to causing damage to the facility's power grid, the means of electricity subscribers will also have detrimental effects Also, if the frequency & Voltage of the micro-grid and allowed optimal control and even collapse the network will not cause instability. Micro-grid frequency & Voltage response correction, especially in the event of an error,one of the decisive factors in the stability and performance of the system. The implementation of a control network based on multi- agent systems where the user is able to decide intelligently convert a lot of research in this area.A multi-agent system is a combination of different factors cooperate to achieve the overall goal of a system. Agents must include a list of protocols that are well defined and represent different ways of communicating with nature (human or other agents) belonging to the system.Multi-agents systems (MAS), which have been applied in computer science studies for years, have characteristics that make them suitable for acting as a basis for building modern distributed control systems.In addition, artificial intelligence techniques can be embedded in some agents with smart features, particularly in automating tasks 711