Indeterminacy Reduction in Agent Communication Using a Semantic Language

Abstract

In recent years, the importance of vagueness and uncertainty in the messages exchanged between agents has been highlighted, mainly due to the ubiquitous nature of the (artificial or human) agents’ communication. The imprecision in the communication becomes more significant when the autonomy of the agents increases or the number of exchanged messages for a communicative goal is limited. In this paper we conjugate ideas drawn from situation semantics theory, human communication, and the multi-agent systems (MAS) field to reduce the impact of vagueness and uncertainty present in the communication. The main advances are achieved with the help of context information, collaboration and reinforcement learning using an agent communication language: the Semantic Agent Programming Language (S-APL).