To explain the necessity of ACL in a Multi agent system, it is extremely important to elucidate about the primary way in which an agent differs from an ordinary object.
An agent is an autonomous piece of software written for some specific purposes, to cater some particular services to its end users. The autonomous behavior of an agent depicts the fact that an agent has some inbuilt intelligence by using which it decides for itself its actions based upon what it senses from its environment. The other aspect of being autonomous is the fact that an agent’s decisions can’t be directly controlled by another party. This translates to the fact that an agents functions can’t be invoked directly by a 3rd party, both theoretically and practically.
The above mentioned explanation illustrates a problem that if an agent’s functions can’t be invoked by another agent/entity, then how the two can interact. How can an agent be influenced from, take input from another agent/entity or give its output to another agent/entity.
The solution to this problem is Agent Communication Language (ACL). The ACL package would be responsible for creation of a message that’s understandable by all entities involved in the multi agent system. Through this package all agents will create a message through some pre defined rules as described in FIPA Specifications. And the message will be sent to the required destination. At the reception end, the agent will take its own decision based on the ACL Message.
Agent
Communication Languages:
Agent Communication Languages provides agents with a means of exchanging information and knowledge, which is really the essence of all forms of interaction in multi-agent systems. The result of which, was the FIPA ACL. ACL is an outer language that specifies message format and include descriptions of their pragmatics, that is the communicative acts or intentions of the agents. Furthermore, FIPA also define semantic languages to successfully communicate with each other. FIPA published SL0 in 1997 and included basic semantics. After this SL1 and SL2 is published which is based on SL0 and provides rich semantics. Every agent has common semantics to talk each other that is known as shared ontology.
Current Status:
Currently our group is trying to understand what is ACL and which rules are used to create ACL message, parse message and generate message in proper format which is understand able to each system.
Future Task:
After completing
analysis our group will complete design and then will start development very
soon.
Group Members