MODELLING AGENTS and COMMUNICATION ---------------------------------- When modelling Air Campaign Planning (ACP), it is not sufficient to represent the activities which take place, their order, and the resources which they use. Multi-perspective modelling frameworks such as CommonKADS and the Information Systems Architecture remind us that it is also important to define WHO performs these activities, and WHERE it is necessary to communicate information between these agents. In order to present this information, we have chosen to use a Role Activity Diagram (RAD). The RAD shows which agents perform particular processes, and shows all information which must be communicated between agents as labelled links. In the diagram shown, each senior officer and each planning cell is treated as a separate agent (yellow rectangle). The diagram shows all the activities which are performed (white squares) and all the interactions which take place (labelled links) between the issuing of a JFC's initial instructions to a JFACC and the production of an Air Tasking Order (ATO). Wherever an interaction takes place, information (often in the form of a document) passes from one agent to another. The links between agents are labelled to show what information is being passed between agents. For example, the Recommended Defense Posture is produced by the Air Strategy planning cell and is communicated to the JFACC. If a piece of information (e.g. the recommended JIPTL) is passed to several agents, then the activity which produces that piece of information is labelled with an X. Information is also communicated witin an agent, particularly if that agent is a planning cell. The extent to which this occurs can be determined by examining the inputs and outputs of activities. For example, the broad weaponeering assessment performed by the Intel Plans cell requires the potential target list, previously produced by the same planning cell, as input.