Predicting the dynamic modulus of asphalt mixture using machine learning techniques: An application of multi biogeography-based programming Ali Behnood ⇑ , Emadaldin Mohammadi Golafshani ⇑ Lyles School of Civil Engineering Purdue University, 550 Stadium Mall, West Lafayette, IN 47907-2051, USA Department of Civil Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran highlights The dynamic modulus ( E j j) of asphalt mixture characterizes its performance at various climate and load conditions. A new E j j predictive model is proposed using biogeography-based optimization (BBO) algorithm. The proposed model provides a high accuracy in prediction of E j j. The proposed model outperforms the previously developed models. Temperature and frequency are the most significant factors affecting the E j j. article info Article history: Received 13 April 2020 Received in revised form 2 August 2020 Accepted 14 September 2020 Keywords: Asphalt Pavement Dynamic modulus Biogeography-based programming Automatic regression abstract The dynamic modulus E j j of asphalt mixtures can be used to characterize the behavior of asphalt pave- ments at a wide range of traffic and climate conditions. The use of E j j predictive models instead of direct laboratory-based measurements can provide several advantages as they do not need trained personnel and expensive equipment. In this study, biogeography-based programming (BBP) was used to develop E j j predictive models with improved accuracy compared to previously developed models. For this pur- pose, two models with different architectures were developed using a dataset containing information on 4022 asphalt mixture samples. Another dataset including the records of 90 asphalt mixtures was used for testing the developed models and comparing their performance with some of the most commonly used models for the prediction of E j j. The results showed that both architectures provided E j j predictive mod- els with excellent accuracy. Moreover, the developed models were found to outperform the Witczak model, Hirsch model, and ANN model. The first BBP model included only four variables: temperature (T), frequency (F), voids in mineral aggregate (VMA), and low-temperature PG (PG L ). The second BBP model included eight variables: T, F, VMA, PG L , high temperature PG (PG H ), asphalt content (AC), volume of effective bitumen content (V beff ), and recycled asphalt pavement (RAP) content. A parametric study and a sensitivity analysis indicated that T and F were the most influential factors affecting the values of E j j. Ó 2020 Elsevier Ltd. All rights reserved. 1. Introduction Asphalt pavement is the most widely used type of pavement in the construction of highways and expressways around the world. An appropriate design guide can extend the service life of asphalt pavements and reduce maintenance costs. The Mechanistic- Empirical Pavement Design Guide (MEPDG) provides an effective tool to use traffic- and climate-related inputs to design the pave- ment. The dynamic modulus ( E j j) of asphalt mixture, which char- acterizes the performance of asphalt mixture under various loading and environmental conditions, is a key input used in the MEPDG. Accurate measurement of the E j j can be done through laboratory-based tests at various simulated temperatures and loading frequency. However, to characterize the performance of asphalt mixture under different traffic and climate conditions, a considerable number of specimens are required, which is a costly and time-taking procedure and requires trained personnel. Alternatively, E j j can be estimated using predictive equations https://doi.org/10.1016/j.conbuildmat.2020.120983 0950-0618/Ó 2020 Elsevier Ltd. All rights reserved. ⇑ Corresponding authors. E-mail addresses: abehnood@purdue.edu (A. Behnood), Golafshani@srbiau.ac.ir (E. Mohammadi Golafshani). Construction and Building Materials 266 (2021) 120983 Contents lists available at ScienceDirect Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat