Abstract: Aiming at the lack of airport pavement original data and its stochastic variability, a new method that integrates Gray model and Markov chain is presented and is described in details by instance. Firstly, a GM (1, 1) model is built to conduct short-term PCI prediction and to expand the dataset. Secondly, least-square method is applied based on the expanded dataset to derive Markov state transition probability matrix. Using this matrix, middle and long term pavement performance is predicted by Markov chain. Thirdly, the precision sensitivity is analyzed,proving the high precision of this integrated model with big variability and sample data. This method can be used as an effective tool for airport pavement performance prediction.

Key words: airport pavement, performance prediction, Gray model, Markov prediction, pavement management