---

The main contribution of the present paper is to train efficient neural networks for seismic design of double layer grids subject to multiple-earthquake loading. As the seismic analysis and design of such large scale structures require high computational efforts, employing neural network techniques substantially decreases the computational burden. Square-on-square double layer grids with the variable length of span and height are considered. Back-propagation (BP), radial basis function (RBF) and generalized regression (GR) neural networks are trained for efficiently prediction of the seismic design of the structures. The numerical results demonstrate the superiority of the GR over the BP and RBF neural networks.

---

Highway construction projects are one of the most important construction projects in the world. Therefore predicting the time of these kinds of projects is important. Basically highway projects are including few activities which are repeating along the horizontal direction. One of the best methods for scheduling these types of projects is linear scheduling method. The repetitive nature of the highway activities is a good reason for schedulers to use linear scheduling methods in order to estimate the time of the project. One of the most important factors in linear projects is considering the effect of the activities productivity on scheduling. The first part of the research has been proposed to quantify the main equation of the identified factors for predicting the daily production rates of the embankment activity. The second part is scheduling the highway construction projects by developing the LSMvpr method based on the application of the embankment activity productivity equation. The purpose of the research is to develop the LSMvpr method for scheduling the highway construction projects by considering the concept of activity productivity in the shape of an equation varying by independent variables changes. By the use of multiple regression analysis the coefficients of affecting factors have been calculated in order to gain a production rate equation for predicting the embankment activity productivity. A software package has been presented for scheduling a highway construction project by coding in MATLAB. The offered software used for validating the model for scheduling the highway construction projects.

---

Slope stability is one of the most complex and essential issues for civil and geotechnical engineers, mainly due to life and high economical losses resulting from these failures. In this paper, a new approach is presented for estimating the Safety Factor (SF) for circular failure slope using hybrid support vector regression (SVR) and Ant Colony Optimization (ACO). The ACO is combined with the SVR for determining the optimal value of its user-defined parameters. The optimization implementation by the ACO significantly improves the generalization ability of the SVR. In this research, the input data for the SF estimation consists of the values of geometrical and geotechnical input parameters. As an output, the model estimates the SF that can be modeled as a function approximation problem. A data set that includes 46 data points is applied in current study, while 32 data points are used for constructing the model, and the remainder data points (14 data points) are used for assessment of the degree of accuracy and robustness. The results obtained show that the hybrid SVR with ACO model can be used successfully for estimation of the SF.

---

In the present study, two different data-driven models, artificial neural network (ANN) and multiple linear regression (MLR) models, have been developed to predict the 28 days compressive strength of concrete. Seven different parameters namely 3/4 mm sand, 3/8 mm sand, cement content, gravel, maximums size of aggregate, fineness modulus, and water-cement ratio were considered as input variables. For each set of these input variables, the 28 days compressive strength of concrete were determined. A total number of 140 input-target pairs were gathered, divided into 70%, 15%, and 15% for training, validation, and testing steps in artificial neural network model, respectively, and divided into 85% and 15% for training and testing steps in multiple linear regression model, respectively. Comparing the testing steps of both of the models, it can be concluded that the artificial neural network model is more capable in predicting the compressive strength of concrete in compare to multiple linear regression model. In other words, multiple linear regression model is better to be used for preliminary mix design of concrete, and artificial neural network model is recommended in the mix design optimization and in the case of higher accuracy requirements.

---

In this paper, two different data driven models, genetic programming (GP) and multivariate adoptive regression splines (MARS), have been adopted to create the models for prediction of bridge risk score. Input parameters of bridge risks consists of safe risk rating (SRR), functional risk rating (FRR), sustainability risk rating (SUR), environmental risk rating (ERR) and target output. The total dataset contains 66 bridges data in which 70% of dataset is taken as training and the remaining 30% is considered for testing dataset. The accuracy of the models are determined from the coefficient of determination (R²). If the R² the testing model is close to the R² value of the training model, that particular model is to be consider as robust model. The modeling mechanisms and performance is quite different for both the methods hence comparative study is carried out. Thus concluded robust models performance based on the R² value, is checked with mathematical statistical equations.  In this study both models were performed, examined and compared the results with mathematical methods successfully. From this work, it is found that both the proposed methods have good capability in predestining the results. Finally, the results reveals that genetic Programming is marginally outperforms over the MARS technique.

---

In the present study, the multivariate adaptive regression splines (MARS) technique is employed to estimate the drying shrinkage of concrete. To this purpose, a very big database (RILEM Data Bank) from different experimental studies is used. Several effective parameters such as the age of onset of shrinkage measurement, age at start of drying, the ratio of the volume of the sample on its drying surface, relative humidity, cement content, the ratio between water and cement contents, the ratio of sand on total aggregate, average compressive strength at 28 days, and modulus of elasticity at 28 days are included in the developing process of MARS model. The performance of MARS model is compared with several codes of practice including ACI, B3, CEB MC90-99, and GL2000. The results confirmed the superior capability of developed MARS model over existing design codes. Furthermore, the robustness of the developed model is also verified through sensitivity and parametric analyses.

---

In the current study two methods are evaluated for predicting the compressive strength of concrete containing metakaolin. Adaptive neuro-fuzzy inference system (ANFIS) model and stepwise regression (SR) model are developed as a reliable modeling method for simulating and predicting the compressive strength of concrete containing metakaolin at the different ages. The required data in training and testing state obtained from a reliable data base. Then, a comparison has been made between proposed ANFIS model and SR model to have an idea about the predictive power of these methods.

---

Lightweight concrete (LWC) is a kind of concrete that made of lightweight aggregates or gas bubbles. These aggregates could be natural or artificial, and expanded polystyrene (EPS) lightweight concrete is the most interesting lightweight concrete and has good mechanical properties. Bulk density of this kind of concrete is between 300-2000 kg/m³. In this paper flexural strength of EPS is modeled using four regression models, nine neural network models and four adaptive Network-based Fuzzy Interface System model (ANFIS). Among these models, ANFIS model with Bell-shaped membership function has the best results and can predict the flexural strength of EPS lightweight concrete more accurately.
 

---

key factor in the successful application of a tunnel boring machine (TBM) in tunneling is the ability to develop accurate penetration rate estimates for determining project schedule and costs. Thus establishing a relationship between rock properties and TBM penetration rate can be very helpful in estimation of this vital parameter. However, this parameter cannot be simply predicted since there are nonlinear and unknown relationships between rock properties and TBM penetration rate. Relevance vector regression (RVR) is one of the robust artificial intelligence algorithms proved to be very successful in recognition of relationships between input and output parameters. The aim of this paper is to show the application of RVR in prediction of TBM performance. The model was applied to available data given in open source literatures. In this model, uniaxial compressive strengths of the rock (UCS), the distance between planes of weakness in the rock mass (DPW) and rock quality designation (RQD) were utilized as the input parameters, while the measured TBM penetration rates was the output parameter. The performances of the proposed predictive model was examined according to two performance indices, i.e., coefficient of determination (R²) and mean square error (MSE). The obtained results of this study indicated that the RVR is a reliable method to predict penetration rate with a higher degree of accuracy.

---

The evaluation of seismic slope performance during earthquakes is important, because the failure of slope (such as an earth dam, natural slope, or constructed earth embankment) can result in significant financial losses and human. It is important, therefore, to be able to forecast such displacements induced by earthquake. However, the traditional forecasting methods, such as empirical formulae, are inaccurate because most of them do not take into consideration all the relevant factors. In this paper, new intelligence method, namely relevance vector regression (RVR) optimized by dolphin echolocation (DE) and grey wolf optimizer (GWO) algorithms is introduced to forecast the earthquake induced displacements (EID) of slopes. The DE and GWO algorithms is combined with the RVR for determining the optimal value of its user-defined paramee RVR. The performances of the proposed predictive models were examined according to two performance indices, i.e., coefficient of determination (R²) and mean square error (MSE). The obtained results of this study indicated that the RVR-GWO model is a reliable method to forecast EID with a higher degree of accuracy (MSE= 0.0160 and R²= 0.9955).

---

During project planning, the prediction of TBM performance is a key factor for selection of tunneling methods and preparation of project schedules. During the construction, TBM performance need to be evaluated based on the encountered rock mass conditions. In this paper, the model based on a relevance vector regression (RVR) optimized by dolphin echolocation algorithm (DEA) for prediction of specific rock mass boreability index (SRMBI) is proposed. The DEA is combined with the RVR for determining the optimal value of its user-defined parameters. The optimized RVR by DEA was employed to available data given in the open source literature. In this model, rock mass uniaxial compressive strength, brittleness index (B_i), volumetric joint account (J_v), and joint orientation (J_o) were used as the input, while the SRMBI was the output parameter. The performances of the suggested predictive model were tested according to two performance indices, i.e., mean square error and determination coefficient. The results show that the RVR- DEA model can be successfully utilized for estimation of the SRMBI in mechanical tunneling.

---

Mechanical excavators are widely utilized in civil/mining engineering projects. There are several types of mechanical excavators, such as an impact hammer, tunnel boring machine (TBM) and roadheader. Among these, roadheaders have some advantages (such as, initial investment cost, elimination of blast vibration, minimal ground disturbances and reduced ventilation requirements). The poor performance estimation of the roadheaders can lead to costly contractual claims. Relevance vector regression (RVR) is one of the robust artificial intelligence algorithms proved to be very successful in recognition of relationships between input and output parameters. The aim of this paper is to show the application of RVR in prediction of roadheader performance. The estimation abilities offered using RVR was presented by using field data of achieved from tunnels for Istanbul’s sewerage system, Turkey. In this model, Schmidt hammer rebound values and rock quality designation (RQD) were utilized as the input parameters, while net cutting rates was the output parameter. As statistical indices, coefficient of determination (R²) and mean square error (MSE) were used to evaluate the efficiency of the RVR model. According to the obtained results, it was observed that RVR model can effectively be implemented for roadheader performance prediction.

---

The artificial neural network is such a model of biological neural networks containing some of their characteristics and being a member of intelligent dynamic systems. The purpose of applying ANN in civil engineering is their efficiency in some problems that do not have a specific solution or their solution would be very time-consuming. In this study, four different neural networks including FeedForward BackPropagation (FFBP), Radial Basis Function (RBF), Extended Radial Basis Function (ERBF), and Generalized Regression Neural Network (GRNN) have been efficiently trained to analyze large-scale space structures specifically double-layer barrel vaults focusing on their maximum element stresses. To investigate the efficiency of the neural networks, an example has been done and their corresponding results have been compared with their exact amounts obtained by the numerical solution.

---

This article presents numerical studies on semi-active seismic response control of structures equipped with Magneto-Rheological (MR) dampers. A multi-layer artificial neural network (ANN) was employed to mitigate the influence of time delay, This ANN was trained using data from the El-Centro earthquake. The inputs of ANN are the seismic responses of the structure in the current step, and the outputs are the MR damper voltages in the current step. The required training data for the neural controller is generated using genetic algorithm (GA). Using the El-Centro earthquake data, GA calculates the optimal damper force at each time step. The optimal voltage is obtained using the inverse model of the Bouc-Wen based on the predicted force and the corresponding velocity of the MR damper. This data is stored and used to train a multi-layer perceptron neural network. The ANN is then employed as a controller in the structure. To evaluate the efficiency of the proposed method, three- story, seven- story and twenty-story structures with a different number of MR dampers were subjected to the Kobe, Northridge, and Hachinohe earthquakes. The maximum reduction in structural drifts in the three-story structure are 13.05%, 39.90%, 15.89%, and 8.21%, for the El-Centro, Hachinohe, Kobe, and Northridge earthquakes, respectively. As the control structure is using a pre-trained neural network, the computation load in the event of an earthquake is extremely low. Additionally, as the ANN is trained on seismic pre-step data to predict the damper's current voltage, the influence of time lag is also minimized.

---

Predicting the bearing capability (q_rs) of geogrid-reinforced stone columns poses a significant challenge due to variations in soil and rock parameters across different locations. The behavior of soil and rock in one region cannot be generalized to other regions. Therefore, accurately predicting q_rs requires a complex and stable nonlinear equation that accounts for the complexity of rock engineering problems. This paper utilizes the Rock Engineering System (RES) method to address this issue and construct a predictive model.To develop the model, experimental data consisting of 219 data points from various locations were utilized. The input parameters considered in the model included the ratio between geogrid reinforced layers diameter and footing diameter (d/D), the ratio of stone column length to diameter (L/dsc), the q_rs of unreinforced soft clay (qu), the thickness ratio of Geosynthetic Reinforced Stone Column (GRSB) and USB to base diameter (t/D), and the settlement ratio to footing diameter (s/D). Following the implementation of the RES-based method, a comparison was made with other models, namely linear, power, exponential, polynomial, and multiple logarithmic regression methods. Statistical indicators such as root mean square error (RMSE), mean square error (MSE), and coefficient of determination (R2) were employed to assess the accuracy of the models. The results of this study demonstrated that the RES-based method outperforms other regression methods in terms of accuracy and efficiency.