Paper 11 -- Prediction of Operational Failure States at Airports by Means of Stochastic Transition Matrices

The paper presents a new methodology to predict non-nominal system states at airports with the help of stochastic transition matrices. These matrices are formed by constant monitoring of operational parameters at airports which reveal the statistical linkage between failure events. Once transition matrices have "learned" from these observations, they can be used to predict failure states from any given starting condition. Moreover, they can also be applied in reverse direction for diagnostic analyses. This methodology is a basis for future decision support tools and clearly aims to ensure a more efficient handling of critical events at airports. The concept of probabilistic failure prediction and diagnosis with transition matrices is introduced and outlined in the paper. Moreover, first results of its application to a model airport system are shown as well as its potential future role in Total Airport Management (TAM) concepts.