ABSTRACT
Artificial Neural Networks (ANNs) are computational tools that operate similar to biological neurons in human brain. They are mainly used for processing information, where inherent and complex patterns are available. They are capable of capturing the nonlinear relationship between the inputs and outputs of a dataset reliably (Haykin 1999). In the field of transportation geotechnics, ANNs have been usually implemented as powerful regression tools. For example, they were successfully used for pavement layer backcalculation by Ceylan et al. (2005), Meier & Rix (1995) and Pekcan et al. (2006). Previous studies also showed that ANNs were powerful and quite effective as forward analysis tools to replace or mimic advanced pavement analyses (Ceylan et al. 1999; Meier et al. 1997).
