An examination of distributed planning in the world of air traffic control

A Distributed Planning System is a network whose nodes represent distinct processors, each cooperating with a selected set of others to achieve a common set of goals. We address two important issues in this paper: (i) How individual processors should be interconnected so that their capacities are fully utilized and their goals can be accomplished effectively and efficiently. (ii) What kind of planning activity the individual processors should engage in. We have chosen the Air Traffic Control (ATC) environment, an intellectually challenging domain of much practical importance, to study different ideas in distributed planning systems. Here we show the results obtained with one type of possible organizational architecture, Location-Centered Cooperative mode of operation. We describe and demonstrate algorithmically a simulation-based planning process based on this architecture. The work described in this paper is part of our continuing effort in examining ways in which greater responsibilities can be delegated to computers in ATC.