An operational framework for the semantics of agent communication languages

Abstract

From an engineering perspective, the agent abstraction can be suitably exploited for tackling cooperation of heterogeneous systems, facilitating semantic interoperability of independently developed software components. But, in order to support the sound design of infrastructures for agent systems, adequate models have to be studied as to grasp the important aspects of cooperation at the desired level of abstraction. In this paper we focus on the semantics of Agent Communication Languages (ACLs), and study the approach based on the idea of describing an agent as a grey-box software component, representing its behaviour by means of transition systems. This framework provides an operational description of ACLs, considering the single-step evolution and interactive capability of an agent, which contrasts the classical frameworks based on intentional descriptions, which rely on the concept of agent mental state. Some examples are provided to show the flavours of the proposed model to describe various semantic aspects of communicative acts.