Planning in the World of the Air Traffic Controller

An enroute air traffic control (ATC) simulation has provided the basis for research into the marriage of discrete simulation and artificial intelligence techniques. A program which simulates, using real world data, the movement of aircraft in an ATC environment forms a robot's world model. Using a production system to respond to events in the simulated world, the robot is able to look ahead and form a plan of instructions which guarantees safe, expedient aircraft transit. A distinction is made between the real world, where pilots can make mistakes, change their minds, etc., and an idealized plan-ahead world which the robot uses; the overall simulation alternates between updating the real world and planning in the idealized one to investigate the robot's ability to plan in the face of uncertainty.