knowledge transfer


Econometric and fuzzy models for the forecast of demand in the airport of Rhodes

Submitted by Josefa Z. Hernandez 27.11.2001, IBA B


A civil aviation authority requires demand forecasts at many levels for planning and other purposes. At the strategic level, forecasts are required for long-term planning over time periods of 10-20 years. More detailed forecast, but over a similar time span, are required for mayor investment projects such as the expansions of terminals and runways. By contrast, the medium-range analysis, 3-10 years ahead, is needed for the annual planning of the aviation authority. The most difficult part is the selection of the relevant causal variables to be taken into account in forecasting and the specification of the type of functional relationship existing between the dependent and independent variables.


At the Section of Transportation, in the Democritus Thrace University (Greece), a fuzzy regression analysis for the forecasting of the airport demand has been used. Instead of probability functions of other forecasting methods, the function of the relationship between the variables can be seen as possibility functions. The fuzzy linear regression model becomes a possibilistic one, that can be used in the context of possibilistic theory to provide a new methodology for capturing the vague and incomplete knowledge by means of possibility distributions.

In fuzzy linear regression models, the difference between data and estimated values is assumed to form an ambiguity that is due the systemís structure. Moreover, the proposed model seeks to bring the ambiguity of the relationship back to the system coefficients and offers one way to construct an accurate relationship which enters directly in the model through fuzzy coefficients.

Status and results

This fuzzy method of forecasting airport demand has taken into account the airport of Rhodes as a case study. Moreover it was used conventional regression methods, but this approach, ignores significant changes that may occur in import explanatory variables that drive demand, however the limits of the fuzzy regression depend on the unpredictable events which affect demand. In the case of the airport of Rhodes, such an unpredictable event was de War in the Persian Gulf in 1991, that contained the upward growth of demand at the airport.

The accuracy of prediction proves to be satisfactory. However, it is never possible to fully predict human behaviour.

Adaptivity and portability

Thanks to the fuzzy techniques the vague and the incomplete knowledge can be captured.

More information

V.A. Profillidis / Journal of Air Transport Management 6 (2000) 95-100

technical reports & case studies

executive summaries

scientific papers




search eunite with