Predicting Pavement Condition Index from International Roughness Index using Gene Expression Programming

Public agencies use pavement management systems to make objective decisions and maintain pavements above the minimum acceptable performance conditions at minimal costs. To achieve this goal, pavement condition is monitored, so its deterioration could be accurately predicted, in order to decide on any required maintenance and rehabilitation. The Pavement Condition Index (PCI) is a composite index used to assess the condition of flexible pavements. The International Roughness Index (IRI) is a smoothness or quality of ride indicator, which is the cumulative vertical movements or vibrations divided by the profile length. Collecting IRI is straightforward and much more affordable than collecting pavement distress data. Predicting the PCI for pavement management assessments without evaluating the extent and severity of the distresses saves costs and person-hours. In this paper, gene expression programming (GEP) was adopted for the first time to predict the PCI from the IRI, using data that was half compiled from the existing literature and the other half was measured and collected in the field by the authors. The PCI values predicted by the GEP model were compared to the estimated PCI values from applying models published in previous studies. The GEP model outperformed all the other available models in the literature, with a maximum R-2 of 82% for the complete dataset.

Automated summary

Public agencies use pavement management systems to monitor pavement condition, accurately predict its deterioration, maintain pavements, and save costs.