Language Constructs for Multi-agent Programming

In this paper we are concerned with proposing, analyzing and implementing simple, yet flexible, constructs for multi-agent programming. In particular, we wish to extend programming languages based on the BDI style of logical agent model with two such constructs, namely constraints and content/context sets. These two aspects provide sufficient expressive power to allow us to represent, simply and with semantic clarity, a wide range of organisational structures for multi-agent systems. We not only introduce this approach, but provide its formal semantics, through modification of an operational semantics based on the core of AGENTSPEAK, 3APL and METATEM. In addition, we provide illustrative examples by simulating both constraints and content/context sets within the Jason interpreter for AGENTSPEAK. In summary, we advocate the use of these simple constructs in many logic-based BDI languages, by appealing to their applicability, simplicity and clear semantics.