Category Power Electronics for Renewable and Distributed Energy Systems
The collection of roles, authority relationships, data flow, resource allocation, and coordination patterns that guide the behaviors of all agents are defined as an organizational topology. The major topologies used in MAS include hierarchies, holarchies, coalitions, teams, congregations, societies, federations, markets, and matrix organizations . Each has its own strengths and weaknesses, and some topologies are more appropriate than others depending on the application:
• A hierarchy (Fig. 15.17) is the earliest and the most widely used topology, in which agents are arranged in a tree-like structure. Agents higher in the tree have a more global view than agents below them...Read More
Since the beginning of artificial intelligence in the 1950s, scientists have been developing computer programs with the purpose of learning how to perform a
specific task. Machine learning became a very exciting branch of artificial intelligence, for it aims at designing machines that can improve their performance without having to actually program them explicitly.
Instead of defining algorithms as functions that map inputs into correct outputs, with the risk of forgetting some inputs when defining the function, machine learning advocates the design of algorithms that enable a machine to learn a function using empirical data. Such data can be either gathered prior to the learning phase of the task at hand, or collected by the machine as it interacts with its environment, i. e...Read More
As described by Russell and Norvig , the internal architecture (or anatomy) of an agent defines its behavior, i. e., there is a function mapping every sequence of percepts (what the agent perceives) into actions. This function can be simple, as in reactive agents, or quite complex as in cognitive agents. An architecture such as subsumption may be used for reactive agents . Cognitive agents may be based on the Soar architecture , while the work of Albus on the reference model architecture  is applicable for hybrid types. There are four main classes of agents classified in accordance to their level of autonomy, the way percepts are used, and how they can be modeled:
• Reflex agents perform simple actions based on current percepts (Fig. 15...Read More
FIPA specification SC00023  defines two levels in how agents should be managed, exist, and operate. The first level, called ‘‘agent level’’, corresponds to each agent itself, while the second relates to groups of agents and how they interact with each other, called the ‘‘MAS level’’. At the agent level, the specification defines the life cycle of the agent, and at the MAS level, it proposes agent management services and a message transport system (MTS). These services are essential in enabling the MAS to operate in a distributed and flexible manner.
The first level defines how agents have their life cycles, from their creation to their end. During their lifetime, agents can be in five possible states:
• Initiated, just after their creation...Read More
A common language and vocabulary is required for agents to communicate. Therefore, languages such as ACL and knowledge query and manipulation language (KQML) were created and can be used for MAS development. ACL is a FIPA specification  and is usually preferred. Similarly to protocols such as TCP, each message is given several attributes, including the content of the message, and information about the participants and the ownership of the conversation. The structure of an ACL message should contain the following parameters:
• A parameter defining the type of communication, called “performative”; this field indicates whether the message is a request, a reply, an information, etc.
• Participants in the conversation, with information on the sender, receiver(s), and reply-to fiel...Read More
The is an on-going development of the ‘‘Internet of things’’ is expected to enable all kinds of products (including energy sources and loads) to be interconnected. These products will need to use a common language to interact coherently. Interoperability between products from different designers and vendors is an important concern if the designed system needs to communicate with other devices to be used or maintained on a regular basis, which is usually the case for power systems applications. The benefits of using standards are important: they enable lower development costs, faster development and better flexibility, interoperability and integration with the existing systems...Read More
With a growing number of applications of MAS in various disciplinary fields, several methodologies have been created to enable developers to follow a formal process when designing MAS. These methodologies describe the tasks and activities that take place during the development process.
Similarly to most standard software development methodologies, or life cycles, most MAS development methodologies follow the same steps, sometimes referred to as the waterfall model: requirements specification, architecture definition and design, implementation, testing/verification, deployment, and maintenance (Fig. 15.10). However, due to the distributed nature of MAS, such methods require some adaptations, which are directly related to how agents are structured and how they interact with each other...Read More
This section gives the fundamentals on how MAS can be applied to power systems. It is described how existing standards, methodologies, languages, tools, architectures, and choices to make can support design specifications such as when and how should which agents interact (cooperate and/or compete) to successfully meet their design objectives. The following items are discussed:
• MAS design methodologies and their characteristics and protocols.
• Standards that MAS should adhere for power systems control.
• Communication protocols between agents.
• Types of services in agent platforms.
• Reactive and intelligent agents structures.
• Use of learning techniques for MAS.
• MAS for planning actions.
• MAS development platforms to facilitate MAS design and prototyping.