Abstract

Aircraft landing problem (ALP) is considered as a scheduling problem where aircrafts are sequenced and allocated with appropriate time slots. In this thesis ALP problem is investigated where several constraints such as aircraft’s landing time windows, minimum separation time and position shifting constraints are taken into consideration. Existing approaches such as optimized solution based methods and heuristic methods to tackle different aspects of the problem are reviewed, and a static mathematical model is studied. The mathematical model is solved and verified using random data generated from simulation. The data are generated based on Pierre Elliott Trudeau International Airport (YUL) in Montreal, Quebec, Canada as well as from relevant data base library. AnyLogictm software was used to simulate aircraft landing operations in a runway environment. An agent based simulation was designed to include the dynamic event of aircrafts arrivals to the runway system.
In the agent based system, an iterative bidding framework is used to generate flight landing schedule in a decentralized environment. In the decentralized environment, we consider each flight as a self-interest agent competing with other flights to get the most appropriate landing time. The efficiency of the decentralized approach is also studied. The results of the decentralized approach are compared with the centralized ALP solution. The results show that the agent based solution approach is able to generate reasonable landing comparing to optimal aircraft landing schedule from the centralized ALP model.