Publications / 2016 Proceedings of the 33rd ISARC, Auburn, USA
One of the main causes of pavement rutting is the repetitive action of traffic loads which results in the accumulation of permanent deformations. As a result, it is important to understand the characteristics of the permanent deformation behavior of asphalt mixes under repeated loading and to build the accurate mix model before they are placed in roadways. This study proposed a hybrid computational intelligence system named SOS-LSSVM for modelling the permanent pavement deformation behavior of asphalt mixtures. The SOS-LSSVM fuses Least Squares Support Vector Machine (LSSVM) and Symbiotic Organisms Search (SOS). LSSVM is employed for establishing the relationship model between the flow number, which is obtained from the laboratory test, and the parameters of the asphalt mix design. SOS is used to find the best LSSVM tuning parameters. A total 118 historical cases were used to establish the intelligence prediction model. Obtained results validate the ability of SOS-LSSVM to model the pavement rutting behavior of asphalt mixture with a relatively high accuracy measured by four error indicators. Therefore, the proposed computational intelligence systems can offer a high benefit for road designers and engineers in decision-making processes.