Abstract

Agents are autonomous system components that communicate using message passing. This paper presents a higher-order agent calculus and its implementation in the lazy functional programming language EBG. The calculus is given a semantics using a translation to the -calculus that encodes higherorder functions and normal order evaluation. 1 Introduction There is increasing demand for distributed software that is able to work collaboratively. The components of such a system are often referred to as agents [Jen98]. Each agent works autonomously but is able to communicate with other agents by passing messages containing data of arbitrary complexity. An agent is similar to a conventional program in that most of its computation is sequential and can be described using existing models. Agents differ from conventional computation with respect to inter-agent communication mechanisms. Messages may be sent asynchronously and are buffered in a queue by the receiver until they are processed. An...

Citations