A systems analysis of scheduled air transportation networks
Author(s)Swan, William M.
Massachusetts Institute of Technology. Flight Transportation Laboratory
MetadataShow full item record
This work establishes the conditions for airline system design building from submodels of smaller aspects of air transportation. The first three sections develop submodels which then are combined in extensive numerical studies of singles market services. The final section discusses the changes to this problem that occur due to network effects. The first section develops a simple model of the cost of providing scheduled transportation on a link. The cost of aircraft of various capacities are divided into a per-frequency cost and a per-capacity cost for conventional subsonic turbojet designs. This cost structure implies that the more capacity provided in conjunction with a fixed schedule of departures the lower the average cost per seat. It is suggested that such aircraft scale economies create a trend toward monopoly or at least oligopoly services. The second section develops a model for demand. The market for transportation is argued to be the city pair. Demand for scheduled service is expressed in terms of fare, frequency and load factor. Fare, frequency, and load factor are combined into total perceived price for the service. This price depends on the consumer's personal value of time. With only a few competitors in such a market, only a few of the technically possible qualities of service will be offered. The services available will be suited better to some tastes than to others. Distributional effects influence the politics of regulation and have been neglected in the past. In this light it is shown that competitive firms are likely to design their services for the same value of time. Product matching increases costs without improving the distribution of benefits. Chapter 4 develops in detail the statistical model used to estimate denied boarding rates from long run design load factors. The development raises doubts about the viability of competition in this dimension. Chapter 5 develops the optimal service for a single carrier on a single city pair market. Optima defined by maximum traffic at zero loss show the importance of the flexibility in aircraft capacity for long run system design. Both algebraic solutions and extensive numerical studies suggest that optimal designs depend on traffic and distance. Changes in frequency and capacity are large; load factor and fare are more stable. Optima are shallow for U.S. domestic cost structures. The final section brings to the discussion issues associated with networks of services. Most U.S. domestic city pairs have amounts of traffic of only modest size compared to the efficient aircraft capacities. Networks overcome these limitations by sharing vehicles among markets. This is done at the expense of extra departure costs. The network design tradeoff in its simplest form is shown to be between larger aircraft capacities and longer stage lengths. The corresponding routing patterns emphasize stops and connections or direct flights. Network design adds another degree of flexibility to the design of transport services: the number of intermediate stops per passenger trip. This affects both cost and service quality.
Originally presented as the author's Ph. D. thesis, M.I.T. Dept. of Aeronautics and Astronautics, 1979June 1979Includes bibliographical references (p. 231-234)
Cambridge, Mass. : Massachusetts Institute of Technology, Flight Transportation Laboratory, 
FTL report (Massachusetts Institute of Technology. Flight Transportation Laboratory) ; R79-5
Airlines, System analysis, Mathematical models, Cost of operation, United States