Abstract
The use of intelligent software agents is a modelling paradigm that is gaining increasing attention in the applications of distributed systems. This paper identifies essential characteristics of agents and shows how they can be mapped into a coloured Petri net representation so that the coordination of activities both within agents and between interacting agents can be visualised and analysed. The detailed structure and behaviour of an individual agent in terms of coloured Petri nets is presented, as well as a description of how such agents interact. A key notion is that the essential functional components of an agent are explicitly represented by means of coloured Petri net constructs in this representation.