Agent Definition - Multi Agent Systems - Lecture Slides, Slides of Multiagent Systems

Download Agent Definition - Multi Agent Systems - Lecture Slides and more Slides Multiagent Systems in PDF only on Docsity!

What makes an agent?

Trivial agents

• A thermostat
  • Senses the environment
  • Acts on the environment
  • Has a goal directed behavior
• Unix daemons: e.g. xbiff
• Even for trivial agents, the agent methodology might prove useful
• Even for trivial agents, emerging properties might solve higher level problems.

Communication in agents

Inter-agent communication: Messages

• Communication can be explicit or implicit
  • Explicit: through messages / signals
  • Implicit: through actions which mean something
• Agent systems prefer message based communication - Client-server systems prefer remote procedure call (RPC) - Equivalent, but flavor is different
• Agent communication languages: FIPA ACL, KQML
• Various XML favors

What is an ontology?

• Answer is 42. But what was the question?
• Communicative acts are meaningless if the agents do not have the same understanding about the passed data.
• An ontology defines a common vocabulary for entities who need to share information in a domain. It includes machine- interpretable definitions of basic concepts in the domain and relations among them. - Entities: agents, researchers, experts etc.

Ontology examples

• The Yahoo categorization of web sites
• Dewey decimal notation
  • Used in libraries since 1876
• RDF – resource description framework and its extensions - DAML (DARPA Agent Markup Language) - OWL - A large number of frameworks: check www.semanticweb.org
• Cyc (http://www.cyc.com)

Mobility in the physical and virtual

world

• Physical mobility
  • Ability to move a physical artifact with whom the agent is associated
  • The laws of the physical world always apply:
    • Conservation of matter (!)
    • The laws of movement and dynamics
• Virtual mobility:
  • Ability to move software from machine to machine
  • The laws of physical world does not apply:
    • Agents can move to remote locations without traversing intermediary locations
    • Agents might be duplicated as a failure of movementDocsity.com

Strong and weak mobility for

software agents

• Mobility: ability to move an agent from a machine across (potentially) heterogeneous platforms
• Strong mobility: can move any time
  • Language or virtual machine support needed
  • Telescript, modified JVM’s such as NOMADS
• Weak mobility: can move at certain points during execution - Aglets, D’Agents (formerly AgentTcl), Jade, Bond etc. Docsity.com

Mutability

Why mutability?

• Some applications:
  • Mobile agents, while traversing heterogeneous platforms might need to adapt
  • Remotely building up monitoring agents
• Future:
  • Continuous updating of long-lived agents
  • Self-repair, self-debug
  • Genetic algorithms for agent design