# Application of Artificial Intelligence (AI) for Sustainable Highway and Road System

^{1}

^{2}

^{*}

## Abstract

**:**

^{2}of 90.5% and relative deviation are scattered around zero line. Besides, the mean, median and standard deviations of experimental and the predicted values are very close. In addition, the mean absolute Error, root mean square error and fractional bias values were found to be low, indicating the high performance of the developed model.

## 1. Introduction

^{2}), Pearson correlation coefficient (R), Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and Fractional Bias (FB). Considering the problems that commonly accompany the cost of experimentation as well as construction delay, AI based modeling is found to be satisfactory for predicting the adhesive force of asphalt binders.

## 2. Materials and Methods

#### 2.1. Materials

^{2}/g and excellent thermal and chemical stability. Other than their unique physical properties, their chemical related properties may have also attracted the attention of researchers in various applications. With the accumulation of certain compounds or external atoms, the attributes of SWCNT are significantly altered. It was used in this study with three distinct fractions of SWCNT weight (0.5%, 1.0% and 1.5%). Table 2 shows the inut and output factors as summary.

#### 2.2. Sample Preparation

#### 2.3. Background of AFM

#### 2.4. Description of Proposed Model

#### 2.4.1. Support Vector Regression (SVR)

^{1}, z

^{2}, …, z

^{n}) denotes the input value and y

_{i}∈ Rl represents the output value. Furthermore, v ∈ Rn, c ∈ R, and n represent the weight vector, a mathematical constant number, and the number of the training dataset, respectively. Additionally, ϕ(z) is an irregular function to assign input data to the high-dimensional feature space. To define v and c, the following formula is employed based on the principle of SRM:

#### 2.4.2. Hyperparameter Optimization Using BOA

## 3. Result & Discussion

#### 3.1. Hybrid BOA-SVR Model Development

`ɛ`, $\gamma )$ since the model performance depends heavily on these parameters. K-fold cross-validation was used to prevent overfitting. In this research, 5-fold cross-validation was selected to protect against overfitting because it showed a low RMSE with less computational time. The kernel function type, kernel parameter value, epsilon, and value of the box constraint were tuned using the BOA technique and the predictive accuracies of the models assessed. Figure 2 shows the progress of the SVR hyperparameter optimization, including the optimal point. The scores for the minimum objective observed of 2279.7 was observed at 12 iterations. The level of accuracy was used to determine the optimal model, applying the other parameters shown in Table 3.

#### 3.2. Evaluation of the Hybrid BOA-SVR Model

_{Exp}is the experimental adhesive strength, ${\overline{V}}_{Exp}$ is the mean experimental adhesive strength, V

_{M}is the model predicted adhesive strength, N is the total number of data. The results of all the performance indices for the model are shown in Table 5. High values of Pearson correlation coefficient (>0.95) with statistical a p-value of 0.000 indicating predicted and experimental results were superimposed. All the statistical error parameters (e.g., MAE, RMSE) were observed to be low (see Table 5 and Ref. [32]). Besides, the performance of a model is acceptable if $\left|FB\right|\le 0.5$ and has a value of zero for an ideal model [4,5]. Thus, the results, as shown in Table 3, indicated that the predictive model used in this study were strongly reliable in predictions. In summary, the BOA-SVR predictions are satisfactory when the correlation coefficient and errors estimates are found to be close to 1 and low, respectively. It is clear that the models met the criteria.

## 4. Conclusions

- A hybrid AI model of BOA-SVR is developed for the anticipation of adhesive force of asphalt.
- The mean, median and standard deviation of experimental and predicted adhesive force seems very close. The interquartile ranges of the experimental and predicted results are also closed which are 111.78 and 101.95, respectively.
- The predicted results overlap with those of the laboratory tests, since the R
^{2}and adjusted R^{2}values between the experimental and predicted values are approximately 90.5%. - The developed model shows that the relative deviations are well dispersed around zero line with low deviations. The residual data points also lie around the zero line, which further validates the reliability of the proposed model.
- The values of statistical error parameters (MAE, RMSE) were obtained to be low. Besides, the value of fractional bias ($\left|\mathrm{FB}\right|)$ is found to be 0.0213, which is very close to zero, indicating that the model is reliable and robust.

## Author Contributions

## Funding

## Institutional Review Board Statement

## Informed Consent Statement

## Data Availability Statement

## Conflicts of Interest

## References

- Yao, H.; Dai, Q.; You, Z.; Ye, M.; Yap, Y.K. Rheological properties, low-temperature cracking resistance, and optical performance of exfoliated graphite nanoplatelets modified asphalt binder. Constr. Build. Mater.
**2016**, 113, 988–996. [Google Scholar] [CrossRef] - Xiao, F.; Amirkhanian, A.; Amirkhanian, S. Influence on rheological characteristics of asphalt binders containing carbon nanoparticles. J. Mater. Civ. Eng.
**2011**, 23, 423–431. [Google Scholar] [CrossRef] - Hassan, M.; Mamun, A.A.; Hossain, M.; Arifuzzaman, M. Moisture damage modeling in lime and chemically modified asphalt at nanolevel using ensemble computational intelligence. Computat. Intell. Neurosci.
**2018**, 1100–1114. [Google Scholar] [CrossRef] [Green Version] - Arifuzzaman, M.; Gazder, U.; Islam, M.; Mamun, A. Prediction and sensitivity analysis of CNTs-modified asphalt’s adhesion force using a radial basis neural network model. J. Adhes. Sci. Technol.
**2019**, 34, 1100–1114. [Google Scholar] [CrossRef] - Arifuzzaman, M. Advanced ANN prediction of moisture damage in CNT modified asphalt binder. Soft Comput. Civ. Eng.
**2017**, 1, 1–11. [Google Scholar] - Arifuzzaman, M.; Islam, M.S.; Hossain, M.I. Moisture damage evaluation in SBS and lime modified asphalt using AFM and artificial intelligence. Neural Comput. Appl.
**2017**, 28, 125–134. [Google Scholar] [CrossRef] - Galan, J.J.; Silva, L.M.; Pérez, I.; Pasandín, A.R. Mechanical Behavior of Hot-Mix Asphalt Made with Recycled Concrete Aggregates from Construction and Demolition Waste: A Design of Experiments Approach. Sustainability
**2019**, 11, 3730. [Google Scholar] [CrossRef] [Green Version] - Plati, C.; Cliatt, B. A Sustainability Perspective for Unbound Reclaimed Asphalt Pavement (RAP) as a Pavement Base Material. Sustainability
**2019**, 11, 78. [Google Scholar] [CrossRef] [Green Version] - Anwar, M.K.; Shah, S.A.R.; Sadiq, A.N.; Siddiq, M.U.; Ahmad, H.; Nawaz, S.; Javead, A.; Saeed, M.H.; Khan, A.R. Symmetric Performance Analysis for Mechanical Properties of Sustainable Asphalt Materials under Varying Temperature Conditions: An Application of DT and NDT Digital Techniques. Symmetry
**2020**, 12, 433. [Google Scholar] [CrossRef] [Green Version] - Han, S.; Chung, H.-W.; Subgranon, T.; Tia, M. Development of Mix Designs for Minimally Refined Reclaimed Asphalt Pavement (RAP) Concrete for Florida Concrete Test Road. Sustainability
**2018**, 10, 2598. [Google Scholar] [CrossRef] [Green Version] - Arifuzzaman, M.; Tarefder, R.; Islam, M. Behavior of Carbon Nano-Tubes (CNTs) as Modifier to Resist Aging and Moisture Damage in Asphalt. Nanosci. Nanotechnol. Asia
**2020**, 10, 1–7. [Google Scholar] [CrossRef] - Arifuzzaman, M.; Gazder, U.; Alam, M.; Sirin, O.; Mamun, A. Modelling of Asphalt’s Adhesive Behaviour Using Classification and Regression Tree (CART) Analysis. Comput. Intell. Neurosci.
**2019**, 7. [Google Scholar] [CrossRef] [Green Version] - Mamun, A.; Arifuzzaman, M.; Taha, R. Nano scale aging characterization of carbon nanotube modified asphalt binders. In Proceedings of the Advances in Materials and Pavement Prediction; CRC Press: London, UK, 2018; p. 4. [Google Scholar] [CrossRef]
- Espinoza, M.; Campos, N.; Yang, R.; Ozer, H.; Aguiar-Moya, J.P.; Baldi, A.; Loría-Salazar, L.G.; Al-Qadi, I.L. Carbon Footprint Estimation in Road Construction: La Abundancia–Florencia Case Study. Sustainability
**2019**, 11, 2276. [Google Scholar] [CrossRef] [Green Version] - Cantisani, G.; Di Mascio, P.; Moretti, L. Comparative Life Cycle Assessment of Lighting Systems and Road Pavements in an Italian Twin-Tube Road Tunnel. Sustainability
**2018**, 10, 4165. [Google Scholar] [CrossRef] [Green Version] - Kleizienė, R.; Šernas, O.; Vaitkus, A.; Simanavičienė, R. Asphalt Pavement Acoustic Performance Model. Sustainability
**2019**, 11, 2938. [Google Scholar] [CrossRef] [Green Version] - Amirkhanian, A.N.; Amirkhanian, S.N. Long-term ageing influence on rheological characteristics of asphalt binders containing carbon nanoparticles. Int. J. Pavement Eng.
**2011**, 12, 533–541. [Google Scholar] - Rath, P.; Love, J.E.; Buttlar, W.G.; Reis, H. Performance Analysis of Asphalt Mixtures Modified with Ground Tire Rubber Modifiers and Recycled Materials. Sustainability
**2019**, 11, 1792. [Google Scholar] [CrossRef] [Green Version] - White, G. A Synthesis on the Effects of Two Commercial Recycled Plastics on the Properties of Bitumen and Asphalt. Sustainability
**2020**, 12, 8594. [Google Scholar] [CrossRef] - Rodrigues, C.; Capitão, S.; Picado-Santos, L.; Almeida, A. Full Recycling of Asphalt Concrete with Waste Cooking Oil as Rejuvenator and LDPE from Urban Waste as Binder Modifier. Sustainability
**2020**, 12, 8222. [Google Scholar] [CrossRef] - Omranian, S.R.; Hamzah, M.O.; Pipintakos, G.; Vuye, C.; Hasan, M.R.M. Effects of Short-Term Aging on the Compactibility and Volumetric Properties of Asphalt Mixtures Using the Response Surface Method. Sustainability
**2020**, 12, 6181. [Google Scholar] [CrossRef] - Liu, H.; Fu, L.; Jiao, Y.; Tao, J.; Wang, X. Short-Term Aging Effect on Properties of Sustainable Pavement Asphalts Modified by Waste Rubber and Diatomite. Sustainability
**2017**, 9, 996. [Google Scholar] - Ge, D.; Zhou, X.; Chen, S.; Jin, D.; You, Z. Laboratory Evaluation of the Residue of Rubber-Modified Emulsified Asphalt. Sustainability
**2020**, 12, 8383. [Google Scholar] [CrossRef] - Zhang, H.-L.; Su, M.-M.; Zhao, S.-F.; Zhang, Y.-P.; Zhang, Z.-P. High and low temperature properties of nano-particles/polymer modified asphalt. Constr. Build. Mater.
**2016**, 114, 323–332. [Google Scholar] [CrossRef] - de Melo, J.V.S.; Trichês, G. Evaluation of properties and fatigue life estimation of asphalt mixture modified by organophilic nanoclay. Constr. Build. Mater.
**2017**, 140, 364–373. [Google Scholar] [CrossRef] - Shahriari, B.; Swersky, K.; Wang, Z.; Adams, R.P.; De Freitas, N. Taking the human out of the loop: A review of Bayesian optimization. Proc. IEEE
**2016**, 4, 148–175. [Google Scholar] [CrossRef] [Green Version] - Mockus, J. Bayesian Approach to Global Optimization, 1st ed.; Springer: Dordrecht, The Netherlands, 1989; p. 270. [Google Scholar] [CrossRef]
- Snoek, J.; Larochelle, H.; Adams, R.P. Practical Bayesian optimization of machine learning algorithms. Adv. Neural Inf. Process. Syst.
**2012**, 4, 2951–2959. [Google Scholar] - Kumar, A.; Luo, J.; Bennett, G.F. Statistical evaluation of Lower Flammability Distance(LFD) using four hazardous release models. Process Saf. Prog.
**1993**, 12, 1–11. [Google Scholar] [CrossRef] - Sultana, N.; Hossain, S.M.; Alam, M.; Islam, M.; Al Abtah, M. Soft computing approaches for comparative prediction of the mechanical properties of jute fiber reinforced concrete. Adv. Eng. Softw.
**2020**, 149, 102887. [Google Scholar] [CrossRef] - Arifuzzaman, M.; Hassan, R. Moisture damage prediction of polymer modified asphalt binder using support vector regression. J. Comput. Theor. Nanosci.
**2014**, 11, 2221–2227. [Google Scholar] [CrossRef] - Kumar, A.; Bellam, N.K.; Sud, A. Performance of an industrial source complex model: Predicting long-term concentrations in an urban area. Environ. Prog.
**1999**, 18, 93–100. [Google Scholar] [CrossRef] - Shafabakhsh, G.; Mirabdolazimi, S.; Sadeghnejad, M. Evaluation the effect of nano-tio2 on the rutting and fatigue behavior of asphalt mixtures. Constr. Build. Mater.
**2014**, 54, 566–571. [Google Scholar] [CrossRef]

**Figure 2.**The progress of Bayesian optimization for tuning hyperparameters of SVR. At 12 iterations, the minimum objective of 2279.7 was observed.

Property | Method | Values |
---|---|---|

PG Grade | ASTM D6373 | 66–22 |

Viscosity (centipoise) | ASTM D4402 | 500 |

Specific gravity | ASTM D-70 | 1.02 |

Input Factors | Output | ||
---|---|---|---|

Factor 1 | Factor 2 | Factor 3 | |

Binder Condition (Fresh, aged and wet) | Binder types (Base, SB4, SB5, SBS4 and SBS5) | Percentage of CNT | Adhesion force |

SVR Model Hyperparameter | Values |
---|---|

Epsilon | 8.2622 |

Box Constraint | 138.47 |

Kernel function | Gaussian |

Kernel Scale | 0.98815 |

Parameters | Expt. | SVR |
---|---|---|

Observation | 405 | 405 |

Mean | 181.29 | 179.35 |

Median | 154.19 | 156.47 |

Std. Deviation | 82.39 | 70.80 |

Minimum | 41.96 | 60.68 |

Maximum | 466.93 | 396.84 |

CoefVar | 45.45 | 39.47 |

SEMean | 4.09 | 3.52 |

Interquartile range | 111.78 | 101.95 |

Criterion | SVR |
---|---|

R | 0.951 (p-value 0.000) |

MAE | 14.2602 |

RMSE | 26.5176 |

FB | 0.0213 |

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Arifuzzaman, M.; Aniq Gul, M.; Khan, K.; Hossain, S.M.Z.
Application of Artificial Intelligence (AI) for Sustainable Highway and Road System. *Symmetry* **2021**, *13*, 60.
https://doi.org/10.3390/sym13010060

**AMA Style**

Arifuzzaman M, Aniq Gul M, Khan K, Hossain SMZ.
Application of Artificial Intelligence (AI) for Sustainable Highway and Road System. *Symmetry*. 2021; 13(1):60.
https://doi.org/10.3390/sym13010060

**Chicago/Turabian Style**

Arifuzzaman, Md, Muhammad Aniq Gul, Kaffayatullah Khan, and S. M. Zakir Hossain.
2021. "Application of Artificial Intelligence (AI) for Sustainable Highway and Road System" *Symmetry* 13, no. 1: 60.
https://doi.org/10.3390/sym13010060