The Multi-Agent System is a type of software in computer science. In traditional programs, each usually has one of the following: either a single designated worker, or a designated leader with an arbitrary number of workers led by that leader. Multi-agent systems function differently according to this: there is an arbitrary number of workers but none of which is a designated controller, and none can view all the information of the workers while all of which are autonomous.
Real Life Information[]
Detailed Description[]
All multi-agent systems consists of agents and their environment. Typically multi-agent research refers to software agents. However, the agents in a multi-agent system could equally well be machines, humans or human teams. A multi-agent system may contain combined human-agent teams.
Every agent in a multi-agent system share the following traits:
- Decentralized: There is no designated agent whose job is to control the other agents. Multi-Agent Systems have no form of government, only members.
- Autonomy: The agents are mostly independent of each other. If one fails, the others aren't affected and no agent needs to contact another in order to act.
- Localized: No individual agent has full access to the entire system. They can only access their own data and designated shared data between the system.
- Self-Organizing: Each agent only needs to follow a single communication standard for interacting with other agents, thus requiring no outside assistance.
Applications[]
Multi-Agent Systems are usually used in environments where having a single program perform each task isn't feasible or would be cumbersome. Common uses include:
- Simulations
- Artificial Intelligence
- Distributed Computing & Robotics
- Economics
- Fault Tolerance
- Coordinated Defensive Systems
- Logistics
- Complex Graphics
- Geographical Data Processing
Examples of Multi-Agent Programs in Code Lyoko[]
- Project Carthage: According to Franz Hopper's diary, Project Carthage did some work with Multi-Agent Systems.
- X.A.N.A.: It was programmed sometime between the late 80s to early 90s, making it a little earlier than real life multi-agent systems. X.A.N.A. combines the multi-agent concept with more sophisticated A.I.
- The Anti-X.A.N.A. Program: This program combined the multi-agent concept with malware, destroying a multi-agent system with another multi-agent system.
Note: Despite the fact that the Marabounta was called a multi-agent system, this was in actuality false since there was a designated leader; so while itt does follow the Agent-Based Model, it's not a true multi-agent system.