Communication between Agents

Knowledge sharing and exchange is particularly important in multi-agent systems (MAS). An agent is usually described as a persistent entity with some degree of independence or autonomy that carries out some set of operations depending on what he perceives. An agent usually contains some level of intelligence, so it has to have some knowledge about its goals and desires. The whole multi-agent system is created to be capable of reaching goals that are difficult to achieve by an individual agent or a monolithic system. In multi- agent systems, an agent usually cooperates with other agents, so it should have some social and communicative abilities.

In order to communicate, agents must be able to:

deliver and receive messages - at this physical level, agents must communicate over agreed physical and network layers to be able to deliver and receive strings or objects that represent messages
parse the messages - at the syntactic level, agents must be able to parse messages to correctly decode the message to its parts, such as message content, language, sender, and also must be able to parse the content of the message
understand the messages - at the semantic level, the parsed symbols must be understood in the same way, i.e., the ontology describing the symbols must be shared or explicitly expressed and accessible to be able to decode the information contained in the message

For multi-agent systems the first physical level as well as the second syntactic level is well standardized by the Foundation for Intelligent Physical Agents (FIPA), for example by agent management specification and agent communication language specification. As for the third level, semantics, standard exists that describe the content languages and that describe usage of ontologies.