The Oasis Air Traac Management System

This paper describes research into automated support for air tra c controllers performing tactical air tra c management. It describes the OASIS system developed to help alleviate air tra c congestion. The system achieves this by maximizing runway utilization, achieved through arranging landing aircraft into an optimal order, assigning them a landing time, and then monitoring the progress of each individual aircraft in real time. OASIS is agent-oriented: its major components are independent agents, each solving a part of the overall problem. The system's exible behavior results in part from this co-operative problem solving approach, and in part from the multiple levels of feedback employed between agents in the system and between the system and its environment. The highly dynamic nature of the air tra c control environment places very stringent real-time constraints on the system's reasoning processes, and requires the system to exhibit both goal-directed and reactive behavior. OASIS computes the landing sequence using an any-time algorithm. OASIS is currently implemented using the Procedural Reasoning System (PRS), a real-time reasoning system capable of reasoning about and performing complex tasks in a robust and exible manner.