Flight delays have always been one of the main problems in air transportation. The purpose of this paper is to apply a critique of the current system for measuring delays and suggest a new index that aggregates departure and arrival delays in a unique indicator. This index evaluates the delays in a continuous and gradated mode instead of imposing limits for a flight to be considered as on schedule or delayed. To construct this index we use data envelopment analysis (DEA) with weights restrictions. The results show that the model with only three variables is the most appropriate. They also show that, for the dataset analyzed, the average delays of the two main Brazilian airlines are quite similar.
Data Envelopment Analysis; Air Transportation; delays
