A Genetic Algorithm–based Approach to Predict Pavement Maintenance Strategies: Iraqi Expressway NO.1 Case Study

Abstract

ABSTRACT:The present paper articulates the applicability of genetic algorithms (GAs) as an optimization tool capable of supporting decision-makers (DMs) to make the right decisions throughout the selection of an optimal pavement maintenance strategy and to predict future pavement condition. GAs efficiently take advantage of historical information to locate search points with improved performance. In this regard, pavement condition index (PCI) for the in-service pavement of the selected case study (Expressway No.1 (R4/A) in Iraq) is estimated based on ASTM D6433-11 and using MicroPAVER 6.5.2 software. Moreover, the related field measurements of the in-service pavement distresses are carried out and classified. To predict the optimal maintenance strategy for the pavement segments within the selected pavement portion case study, a GA optimization technique is implemented as an application of stochastic approach using EVOLVER 6.3.1 software to evaluate the pavement performance based on PCI. For the required validation process of the predicted PCI results obtained by the GA technique, predicted PCI results obtained by experts' opinions based on the design questionnaire were estimated and applied. The statistical validation analyses showed that the predicted PCI values obtained via EVOLVER 6 Genetic Algorithm software seem to be close to those obtained via the analyses of the experts' questionnaires. Based on the research outcomes, it is concluded that one can recognize using the presented procedure throughout the implementation of stochastic approach in the form of GA to predict the optimal pavement maintenance strategy for in-service pavement.