Efficient operation is essential for a commercial passenger airline to be profitable in a very competitive industry. Optimisation has become key tool for planning and operations in the airline industry because small improvements can make a great difference in profit. A large airline flies about 4,000 flight segments daily. Therefore if the net revenue can be increased by the tiny amount of $100 per flight per day, annual profit will increase by more than one hundred million dollars per year.
Optimisation is used primarily for scheduling and pricing (yield management). Here we focus on three aspects of scheduling:
* the assignment of airplane types to flights, the fleeting problem
* the routing of individual aircraft, the rotation problem
* the assignment of crews to flights, the crew scheduling problem
The hub-and-spoke network structure used by the airlines permits a huge number of options in scheduling planes and crews. The only logical way to evaluate these options systematically is through an optimisation model.
For the most part, the three components are considered separately in the indicated order, because a fully integrated model would be impossible to handle. However, the fleeting and rotation solutions can impose severe restrictions on crew planning and therefore a major challenge is to capture key aspects of crew scheduling in the previously solved airplane scheduling models.
Each of the three component models is difficult to solve because they are integer programs (NP-hard Problems) with a large number of variables and constraints. The fleet assignment problem can have 50,000 variables and roughly the same number of constraints. The crew-scheduling problem can have millions of variables and hundreds of constraints. The objective function used by the airlines is profit maximisation or cost minimisation. Unfortunately these objectives do not consider robustness of the solutions, which is very important since equipment breakdowns and bad weather cause disruptions in the planned schedule nearly every day.
The incorporation of robustness into these models presents another challenge that is related to the real-time rescheduling that must be done when flights are delayed or canceled. So far the industry has done very little with optimisation in real time scheduling and that is the major challenge of the future.

George L. Nemhauser was born in New York City and was educated at the Bronx High School of Science, City College of New York (B.Ch.E. 1958) and Northwestern University (M.S. 1959, Ph.D. 1961).
He joined the faculty of the Johns Hopkins University as Assistant Professor of Operations Research and Industrial Engineering in 1961. In 1970, he was appointed Professor of Operations Research and Industrial Engineering at Cornell University and Leon Welch Professor in 1984. He served as School Director during the period 1977 - 1983. He came to Georgia Tech's School of Industrial and Systems Engineering in 1985 as the A. Russell Chandler Professor and was appointed Institute Professor in 1991. He is also research director of The Logistics Institute. He has held visiting faculty positions at the University of Leeds, U.K. and the University of Louvain, Belgium. At Louvain he worked at the Center for Operations Research and Econometrics and was Research Director for two years.
His principal research interests are in the area of discrete optimization. He is the author of 3 books and more than 100 papers. He has supervised more than 30 doctoral dissertations. His current interests are in solving large-scale mixed-integer programming problems and he is actively working on several applications, especially scheduling problems in the airline industry. He is one of the developers of MINTO, a software system for solving mixed-integer programs.
His honors and awards include membership in the National Academy of Engineering, Kimball medal and Lanchester prize (twice) and Morse lecturer of ORSA. He received awards for outstanding teaching at Johns Hopkins.
He has served ORSA as Council Member, President and Editor of Operations Research. He is the founding and current Editor of Operations Research Letters. He is co-editor of Handbooks of Operations Research and Management Science. He is the Past Chairman of the Mathematical Programming Society.
He has served various government agencies including NSF, NIST and NRC.