Estimating the geographic distribution of originating air travel demand using a bi-level optimization model

Abstract

The historical air travel demand plays an important role in the analysis of, for example, major changes in airport use, airport and airline marketing, and airport planning. We provide a bi-level optimization model as a relatively quick and less expensive alternative to survey method to estimate the originating air travel demand and its geographic distribution at an airport. The lower-level model estimates the geographic distribution of originating air travel demand and model coefficients. The upper-level model estimates the airport access distance threshold which is used to model traveler’s airport choice behaviors in the lower-level model. We adopt an Evolutionary Algorithm (EA) to solve the bi-level optimization model. A General Reduced Gradient solution algorithm is used within the EA to solve the lower-level model. We present a real-world case study in which we apply the model to estimate the originating air travel demand and its geographic distribution on the contiguous United States. The model estimates are generally close to the statistics from the American Travel Survey. The comparisons of the model estimates with the statistics from the Washington-Baltimore regional air passenger survey show mixed results. Possible reasons for the estimation errors are identified.