Multi-agent System Architecture And Agent Interaction

The essence of agent technology is researching how to make one or several entities perform delegated task without the need for human supervision, autonomously, and using various possible methods and techniques. Ideally, an agent has the capability of apprehending user's exact intentions; perform various kinds of goals-drove actions (such as social intercourse, learning, reasoning, and cooperation); sensing and adapting to complex or ever-changing dynamic environments, making full use of available data, knowledge, information, and computation resource, providing agile, exact and satisfied help for users.In an environment where multiple agents exist, they would cooperate and collaborate with each other to solve the problem of the user. This cooperation exist because the domain of the problem may be too big to be handled by a single agent or the problem require agents, having different knowledge or skills, to contribute to the final solution of the problem. Agents communicate to cooperate to solve a problem, to resolve conflicts, and to achieve coordination in the system. An agent may also interact with other agents to satisfy a personal agenda in its plan, where they may interact by negotiating or bidding to perform or delegate a task. Agent-based architectures offer modularity, robustness, separation of concerns, and other advantages of a distributed system. These characteristics have led agent technology to gain applications in wide areas. Electronic Commerce is one of the main domains for multi-agent system application, it provides an ideal domain in which to test concepts in agent to agent trade, negotiation, communication and cooperation.In this paper, firstly we introduce the background, definition, and correlative theories of agent technology. Furthermore, we analyzed the construction and support environment of multi-agent systems; Secondly, we discussed the interaction techniques and theories of an agent, such as communication languages, communication protocols, coordination model and negotiation model, within multi-agent systems; Finally, this paper presents a Multi-Agent System based Electronic Commerce Model.