Multi-agent system
In a multi-agent system or MAS is a system composed of several identical or different specialized trading units, software agents, collectively solve a problem.
Multi-agent systems exist in biology (natural multi-agent systems ) as well as in the art. An example family of biological multi-agent systems provide ant colonies dar. Some of the processes occurring in ant colonies algorithms ( ant algorithms ) provide heuristic solution methods for complex optimization tasks represent and are in addition to their fundamental interest within theoretical biology is also a model for the optimization of technical processes. This is known as distributed, in the field of technology of artificial intelligence or DAI (distributed artificial intelligence ).
Outside Europe, particularly in the U.S., the term Agent -based Modeling and Simulation ( ABM) has prevailed for MAS. The term is also used for a special kind of knowledge logic. In the knowledge logic is called the carrier of each modeled knowledge ( eg, people, player, processors) as "agents ". It should be noted that systems of mobile agents (mobile agent systems ) often also abbreviated MAS. Mobile agents are software agents, in which the transfer of the design to other nodes has particular importance in a network.
Multi- agent systems in engineering
Multi-agent systems are a sort of expression of characteristics and architecture of agents arrangements. Generally speaking, they are a research area of distributed artificial intelligence that deals with how autonomous, distributed, and "intelligent" systems vote as a unit specific knowledge, goals, abilities and plans to coordinate to act or solve problems.
There are e.g. collaborative agent systems that are distributed confident in their architecture, so as to meet flexible and reliable tasks as a single, local system. It may separate from the interaction of the actuators out a problem, for which a single unit would never be able to. These units are each responsible for an activity, a higher-level control is not necessary and they find together by self-organized coordination solutions.
Brief explanation of the types of agents:
- Collaborative agents: goals are achieved through cooperation and negotiations with other agents, usually built as multi-agent systems. Cooperativity and autonomy are important, but they are often self-learning.
- Interface agents: Communicate often with a human as a system user. Help on the Internet to find deals and specialize to negotiate. Not to these agents include Help Wizard.
- Smart Agents: Do all properties and are able to cope with the variety of tasks.
The main types of communication in multi-agent systems are:
- Point -to-point
- Broadcast
- Announcement
- Signal
The news form most commonly used in dynamic networks is the broadcasting, so that all agents receive each message, and then decide whether and how they take action.
Cooperation and coordination
From the insolubility of the overall problem by a single agent out a cooperation between agents is necessary. In the case of different types of task, the actions must be coordinated and planned in their timing, in order to remain effective and efficient. If a task is answered by an agent, it must immediately return this information so that not more agents want to do the same and interfere with each other in the extreme. Occur, for example, waiting times, must also be considered with the. There are additional tasks that are not directly productive but organizational and necessary when multiple autonomous agents in a shared environment pursue their own goals. Occur, for example conflicts, this can be solved by coordination mechanisms. In particular, include automated negotiation between software agents to the appropriate organizational coordination mechanisms to reduce transaction costs.
PAGE Description
Agents can also be described by PAGE (acronym for percepts, actions, goals, environment = percepts, actions, goals, environment).
BDI Description
Another characterization used the acronym BDI, which stands for beliefs, desires and intentions.
Applications
Robocup, web crawlers, production planning and control, software agent, multi-agent simulation