Multi-Agent System Applications--论文代写范文精选

常见的交互类型的例子包括:合作，协调和谈判。交互的多代理系统有别于其他形式的软件,它提供了底层的范式。如前所述,多代理系统信息不完整,数据分散和计算异步。下面的essay代写范文进行详述。

A large variety of Multi-Agent Systems are available on the commercial market and as open-source on the Internet. Agent technologies are not restricted to academia, companies and research labs anymore, but they have entered gradually into our computer systems and homes. In solving problems that concern interactions between people, the application of Multi-Agent Systems has proven that it is a good candidate, i.e. Multi-agent systems are ideally suited to representing problems that have multiple problem solving methods, multiple perspectives and/or multiple problem solving entities. Such systems have the traditional advantages of distributed and concurrent problem solving, but have the additional advantage of sophisticated patterns of interactions. 

  Examples of common types of interactions include: cooperation (working together towards a common aim); coordination (organising problem solving activity so that harmful interactions are avoided or benecial interactions are exploited); and negotiation (coming to an agreement which is acceptable to all the parties involved). It is the exibility and high-level nature of these interactions which distinguishes multi-agent systems from other forms of software and which provides the underlying power of the paradigm. (Jennings et al., 1998, p. 9) First, as mentioned before, Multi-Agent Systems consist of agents that have incomplete information, no centralised control, data is decentralised and computation is asynchronous. Secondly, the weak notion of agency states that agents are exible and in general have the properties of autonomy, social ability, reactivity and pro-activeness. And thirdly, we have distinguished different types of agents, i.e. the cognitive and the social agent. 

  Creating a classication of MultiAgent Systems is a tedious task and we could take an approach that is based on the types of agents applied and their interaction with the environment resulting in a classication of reactive towards social pro-active systems. However, the difference in opinions and denitions about what agents are, automatically results in a problematic classication of MASs. An approach by Wooldridge (2002) is to rst draw the distinction between distributed systemsmultiple agents are nodes in a distributed systemand agents that have the (pro-active) function of assisting users working with applications (Maes, 1994). Next, Luck, McBurney, and Preist (2004) divide MAS into (real-time) Multi-Agent decision systemsagents participate in a system have to make joint decisions, and Multi-Agent simulation systems, where MAS is applied as a model to simulate some real-world domain. The important distinction between both is that with decision systems, there is often a direct outcome ready for application in the real world, while simulation systems require rst an interpretation of the researcher before results can be applied in the real world27 .

  Nevertheless, we follow the approach of Luck et al. by classifying systems in two groups and show in which domains MAS applications are developed. The domains can be structured varying from hardware, such as robotics and manufacturing towards software, e.g. simulation systems and entertainment. In the next sections, 2.6.1 and 2.6.2, we respectively mention examples of multiagent decision systems and multi-agent simulation systems.

  Multi-Agent decision systems The Industrial Application is one of the domains where MAS needs to interact with the physical world, be it humans or other agents. Industrial manufacturing and (real-time) control systems in the form of MASs are introduced to solve coordination problems in factories, air-trafc control, telecommunication networks and transportation systems. The general goal of industrial applications is to manage processes efciently. Examples of industrial applications are the manufacturing system YAMS (Parunak, 1995) and the process control system ARCHON (Jennings & Pople, 1993). Closely linked to manufacturing systems (primary processes) are applications for workow and business process management (secondary processes). 

  An example is the ADEPT project that views the business process []. . . as a collection of autonomous problem solving entities that negotiate with one another and come to mutually acceptable agreements that coordinate their interdependent sub-activities (Jennings, Faratin, Norman, O'Brien, & Odgers, 2000, p. 2). Industrial applications are mostly custom-made solutions for specic problem areas in the Business to Business market, while commercial applications are more directed towards the consumer market. Coordination of interdependent sub-activities, similar to ADEPT, can also be applied in information retrieval and sharing systems. 

  A `web' of agents that are experts in a certain area of interest form a connected network of information sharing nodes. Other agents function as brokers and have knowledge about where to go for what. In P2P applications, e.g. Gnutella, Bit Torrent and Skype, this functionality is already delivering music, video and voice over IP. Following up the information agent network, e-commerce is the next step in which (mobile) agents take over tasks from the user and `walk' over the Internet to compare prices and nd the most suitable deals with supplier agents. For instance, Chavez and Maes (1996), de- ne a web-based multi-agent auction system (Kasbah) in which humans are represented by bidder and seller agents. 

 Multi-Agent simulation systems 
  Agent based simulation systems are used to describe, prescribe and predict realworld scenarios and are applied in many areas, e.g. trafc systems, crowd behaviour and riot control, combat scenarios and many alike. In this dissertation, the (cognitive) agent-based social simulation (Gilbert & Troitzsch, 1999; Sun, 2006a) is applied as a modelling tool for studying the behaviour of agents that represent individual people. In the eld of agent-based social simulation, two broad approaches can be distinguished (Luck et al., 2004). One approachthe emphasis in this dissertationdenes systems that underlie social interaction, and the other focuses on observing social processes and modelling those. （essay代写)

  The combination of both approaches can be an iterative process, by rst analysing the eld, next creating a multi-agent simulation system and nally observing and validating the model. An example of an application is given by Wooldridge (2002) who mentions a social system, the EOS project (Doran & Palmer, 1995) that consists of agents making it possible to describe a number of social phenomena, such as `overcrowding'when too many agents attempt to obtain resources in some locale or `clobbering'when agents accidentally interfere with each other's goals. Another well known example of a simulation system is Swarm (Minar, Burkhart, Langton, & Askenazi, 1996). Swarm is a generic discrete event model used for simulating a collection of independent agents of any kind. The areas in which Swarm already is applied is diverse, e.g. chemistry, economics, physics, anthropology, ecology, and political science. Many simulation toolkits have evolved and the search for adopting a toolkit that is suitable for our current research has been difcult. Therefore, in the discussion of this chapter, we explain why we invest in constructing a new toolkit. （essay代写)

  51Due网站原创范文除特殊说明外一切图文著作权归51Due所有；未经51Due官方授权谢绝任何用途转载或刊发于媒体。如发生侵犯著作权现象，51Due保留一切法律追诉权。
  更多essay代写范文欢迎访问我们主页 www.51due.com 当然有essay代写需求可以和我们24小时在线客服 QQ:800020041 联系交流。-X（essay代写)