---

Optimization techniques can be efficiently utilized to achieve an optimal shape for arch dams. This optimal design can consider the conditions of the economy and safety simultaneously. The main aim is to present an applicable and practical model and suggest an algorithm for optimization of concrete arch dams to enhance their seismic performance. To achieve this purpose, a preliminary optimization is accomplished using PSO procedure in the first stage. Capabilities of Ansys Parametric Design Language (APDL) are applied for modeling the Dam-Foundation-Reservoir system. In the second stage with training the neural network, Group Method of Data Handling (GMDH) and replacement of Ansys analyst, optimal results have been achieved with the lowest error and less number of iteration respectively. Then a real world double-arch dam is presented to demonstrate the effectiveness and practicality of the PSO-GMDH. The numerical results reveal that the proposed method called PSO-GMDH provides faster rate and high searching accuracy to achieve the optimal shape of arch concrete dams and the modification and optimization of shape have a quite important role in increasing the safety against dynamic design loads.

---

Reinforced concrete (RC) slabs exhibit complexities in their structural behavior under load due to the composite nature of the material and the multitude and variety of factors that affect such behavior. Current methods for determining the load-deflection behavior of reinforced concrete slabs are limited in scope and are mostly dependable on the results of experimental tests. In this study, an alternative approach using Artificial Neural Network (ANN) model is produced to predict the load-deflection behavior of a two-way RC slab. In the study, 30 sets of RC slab specimens of sizes 700mm x 600mm x 75mm were cast, cured for 28days using the sprinkling method of curing and tested for deflection experimentally by applying loads ranging from 10kN to 155kN at intervals of 5kN. ANN model was then developed using the neural network toolbox of ANN in MATLAB version R2015a using back propagation algorithm. About 54% of the RC specimens were used for the training of the network while 23% of the sets were used for validation leaving the remaining 23 % for testing the network. The experimental test results show that the higher the applied load on the slab, the higher the deflection. The result of the ANN model shows a good correlation between the experimental test and the predicted results with training, validation and test correlation coefficients of 0.99692, 0.98921 and 0.99611 respectively. It was also found that ANN model is quite efficient in determining the deflection of 2-way RC slab. The predicted accuracy of performance value for the load-deflection set falls at 96.67% of the experimental load-deflection with a 0.31% minimum error using the Microsoft spreadsheet model. As such the comprehensive spreadsheet tool created to incorporate the optimum neural network. The spreadsheet model uses the Microsoft version 2013 excel tool software and can be used by structural engineers for instantaneous access to the prediction if any aspect of a concrete slab behavior given minimal data to describe the slab and the loading condition.

---

For seismic resistant design of critical structures, a dynamic analysis, based on either response spectrum or time history is frequently required. Due to the lack of recorded data and randomness of earthquake ground motion that might be experienced by the structure under probable future earthquakes, it is usually difficult to obtain recorded data which fit the necessary parameters (e.g. soil type, source mechanism, focal depth, etc.) well. In this paper, a new method for generating artificial earthquake accelerograms from the target earthquake spectrum is suggested based on the use of wavelet analysis and artificial neural networks. This procedure applies the learning capabilities of neural network to expand the knowledge of inverse mapping from the response spectrum to the earthquake accelerogram. At the first step, wavelet analysis is utilized to decompose earthquake accelerogram into several levels, which each of them covers a special range of frequencies. Then for every level, a neural network is trained to learn the relationship between the response spectrum and wavelet coefficients. Finally, the generated accelerogram using inverse discrete wavelet transform is obtained. In order to make earthquake signals compact in the proposed method, the multiplication sample of LPC (Linear predictor coefficients) is used. Some examples are presented to demonstrate the effectiveness of the proposed method.

---

The use of supplementary cementing materials is gradually increasing due to technical, economical, and environmental benefits. Supplementary cementitious materials (SCM) are most commonly used in producing ready mixed concrete (RMC). A quantitative understanding of the efficiency of SCMs as a mineral admixture in concrete is essential for its effective utilisation. The performance and effective utilization of various SCMs can be possible to analyze, using the concept of the efficiency factor (k-value). This study describes the overview of various studies carried out on the efficiency factor of SCMs. Also, it is an effort directed towards a specific understanding of the efficiency of SCMs in concrete. Further it includes an overview of artificial neural network (ANN) for the prediction of the efficiency factor of SCMs in concrete. It is found that The model generated through ANN provided a tool to calculate efficiency factor (k) and capture the effects of different parameters such as, water-binder ratio; cement dosage; percentage replacement of SCMs; and curing age.

---

The prediction of the ultimate bearing capacity of the pile under axial load is one of the important issues for many researches in the field of geotechnical engineering. In recent years, the use of computational intelligence techniques such as different methods of artificial neural network has been developed in terms of physical and numerical modeling aspects. In this study, a database of 100 prefabricated steel and concrete piles is available from existing publications to solve issues related to pile’s bearing capacity analysis. Three different artificial neural network algorithms were developed for comparing and verifying the obtained results at analyzing the bearing capacity of pile in soils. During the modeling process, the geometric properties of different piles, soil materials properties, friction angle and flap numbers (hammer blows) were selected as input parameters to the selected network and the output from the network was considered as the bearing capacity of the pile. Finally, the performance of radial base function type neural networks was compared with model tree method and predictive neural networks based on different learning algorithms such as Levenberg-Marquardt and Bayesian Regulation Back Propagation Algorithms. It was observed that the radial base neural network in some cases achieved better results from accuracy based on common statistical parameters such as correlation coefficient, mean absolute error percentage and root mean square error as compared to other stated methods and it showed the acceptable performance in modeling and predicting the desired output close to the target's results.

---

In this study, the complex behavior of steel encased reinforced concrete (SRC) composite beam–columns in biaxial bending is predicted by multilayer perceptron neural network. For this purpose, the previously proposed nonlinear analysis model, mixed beam-column formulation, is verified with biaxial bending test results. Then a large set of benchmark frames is provided and P-M_x-M_y triaxial interaction curve is obtained for them. The specifications of these frames and their analytical results are defined as inputs and targets of artificial neural network and a relatively accurate estimation model of the nonlinear behavior of these beam-columns is presented. In the end, the results of neural network are compared to some analytical examples of biaxial bending to determine the accuracy of the model.

---

Concrete is the second most consumed material after water and the most widely used construction material in the world. The compressive strength of concrete is one of its most important mechanical properties, which highly depends on its mix design. The present study uses the intelligent methods with instance-based learning ability to predict the compressive strength of concrete. To achieve this objective, first, a set of data pertaining to concrete mix designs containing fly ash was collected. Then, mix design parameters were used as the inputs of the artificial neural network (ANN), support vector machine (SVM), and adaptive neuro-fuzzy inference system (ANFIS) developed for predicting the compressive strength. In all these models, prediction accuracy largely depends on the parameters of the learning model. Hence, the particle swarm optimization (PSO) algorithm, as a powerful population-based algorithm for solving continuous and discrete optimization problems, was used to determine the optimal values of algorithm parameters. The hybrid models were trained and tested with 426 experimental data and their results were compared by statistical criteria. Comparing the results of the developed models with the real values showed that the ANFIS-PSO hybrid model has the best performance and accuracy among the assessed 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.
 

---

Direct drilling method and the use of microtremor studies are among the most commonly used available methods utilized to estimate dynamic parameters for a site. One of the most important parameters is the dominant period of the site whose estimation plays a pivotal role in seismic hazard mitigation. The conventional models obtained are not capable of estimating the parameters that govern the seismic response of a site. Therefore, Artificial Neural Networks (ANNs) are reliable and practical estimation methods that can be used to analyze comprehensive measurements such as dominant period of a site, and improve the data. In this paper, the performance of ANNs has been investigated on calculation of the dominant period for a site. Three different models, namely BP, RBF and ANFIS, have been compared to determine the best model that provides the most accurate estimation for the dominant period. The input parameters have been chosen to be alluvial layer thickness, grain size, specific gravity, effective stress, shear wave velocity, standard penetration number, Atterberg limits. Each of the three models has been trained and tested for these input parameters and a unique output which is the dominant period of the site. The results showed that ANNs successfully model complex relationships between soil parameters and seismic parameters of the site, and provide a robust tool to accurately estimate the dominant period of a site. The accurate estimations can be then used for engineering applications including damage assessment and structural health monitoring. In addition, The obtained emulator of RBF model shows the least model error in estimation of dominant period and has been found to be superior to the other evaluated methods.

---

In recent years, soft computing and artificial intelligence techniques such as artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS) have been effectively used in various civil engineering applications. This study aims to examine the potential of ANN and ANFIS for modeling the compressive strength of concrete containing expanded perlite powder (EPP). For doing this, a total of forty-five EPP incorporated concrete mixtures were produced and tested for compressive strength at different curing ages of 3, 7, 28, 42 and 90 days. Two different ANN models were developed and the suitable and stable ANN architecture for each model was considered by calculating various statistical parameters. For comparative purposes, two ANFIS models with different membership functions were also trained. According to the results, it can be concluded that the proposed ANN models relatively give a good degree of accuracy in predicting the compressive strength of concrete made with EPP, higher than that of observed from ANFIS models.

---

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.

---

Reliable and accurate streamflow forecasting plays a crucial role in water resources systems (WRS) especially in dams operation and watershed management. However, due to the high uncertainty associated WRS components and nonlinear nature of streamflow generations, the realistic streamflow forecasts is still one of the most challenging issue in WRS. This paper aimed to forecast one-month ahead streamflow of Karun river (Iran) by coupling an artificial neural network (ANN) with an improved binary version of gravitational search algorithm (IBGSA), named ANN- IBGSA. To this end, the best lag number for each predictor at Poleshaloo station was firstly selected by auto-correlation function (ACF). The ANN-IBGSA was used to minimize the sum of RMSE and R2 and to identify the optimal predictors. Finally, to characterize the hydro-climatic uncertainties associated with the selected predictors, an
implicit approach of Monte-Carlo simulation (MCS) was applied. The ACF plots indicated a significant correlation up to a lag of two months for the input predictors. The ANN-IBGSA identified the Tmean (t-1), Q(t-1) and Q(t) as the best predictors. Findings demonstrated that the ANN-IBGSA forecasts were considerably better than those previously carried out by researchers in 2013. The average improvement values were 9.91%, 11.85% and 9.13% for RMSE, R2 and MAE, respectively. The Monte-Carlo simulations demonstrated that all of forecasted values lie within the 95% confidence intervals.
 

---

Depending on the tectonic activities, most buildings subject to multiple earthquakes, while a single design earthquake is suggested in most seismic design codes. Perhaps, the lack of easy assessment to second shock information and sometimes use of inappropriate methods in estimating these features cause successive earthquakes mainly were ignored in the analysis procedure. In order to overcome to above deficiencies, the learning abilities of artificial neural networks (ANNs) are used in two steps to evaluate the seismic capacity of steel frames consisting moment-resisting frames, ordinary concentrically, and buckling restrained brace (BRB) under critical consecutive earthquakes. For this purpose, peak ground acceleration of second shock (PGA_a) is estimated based on the first shock features in the first step. Next, second ANNs estimate the decreased capacity of the damaged structure for LS and CP performance level according to the proposed PGA_a from the previous step and some seismic and structural features. The results indicate that ANNs are trained to generalize the unseen information very well and reflect good precision in predicting target results in both steps. Finally, the effect of different parameters and repeated shocks is investigated on the seismic performance of mentioned frames. The results show the proper performance of BRB frames in the case of real and repeated earthquakes.
 

---

The main aim of this study, is to evaluate the seismic reliability of steel concentrically braced frame (SCBF) structures optimally designed in the context of performance-based design. The Monte Carlo simulation (MCS) method and neural network (NN) techniques were utilized to conduct the reliability analysis of the optimally designed SCBFs. Multi-layer perceptron (MLP) trained by back propagation technique was used to evaluate the required structural responses and then the total exceedence probability associated with the seismic performance levels was estimated by the MCS method. Three numerical examples of 5-, 10-, and 15-story SCBFs with fixed and optimal topology of braces are presented and their probability of failure was evaluated considering the resistance characteristics and the seismic loading of the structures. The numerical results indicate that the SCBFs with optimal topology of braces were more reliable than those with fixed topology of braces.  
 

---

Despite the advantages of the plastic limit analysis of structures, this robust method suffers from some drawbacks such as intense computational cost. Through two recent decades, metaheuristic algorithms have improved the performance of plastic limit analysis, especially in structural problems. Additionally, graph theoretical algorithms have decreased the computational time of the process impressively. However, the iterative procedure and its relative computational memory and time have remained a challenge, up to now. In this paper, a metaheuristic-based artificial neural network (ANN), which is categorized as a supervised machine learning technique, has been employed to determine the collapse load factors of two-dimensional frames in an absolutely fast manner. The numerical examples indicate that the proposed method's performance and accuracy are satisfactory.
 

---

Cement, water, fine aggregates, and coarse aggregates are combined to produce concrete, which is the most common substance after water and has a distinctly compressive strength, the most important quality indicator. Hardened concrete's compressive strength is one of its most important properties. The compressive strength of concrete allows us to determine a wide range of concrete properties based on this characteristic, including tensile strength, shear strength, specific weight, durability, erosion resistance, sulfate resistance, and others. Increasing concrete's compressive strength solely by modifying aggregate characteristics and without affecting water and cement content is a challenge in the direction of concrete production. Artificial neural networks (ANNs) can be used to reduce laboratory work and predict concrete's compressive strength. Metaheuristic algorithms can be applied to ANN in an efficient and targeted manner, since they are intelligent systems capable of solving a wide range of problems. This study proposes new samples using the Taguchi method and tests them in the laboratory. Following the training of an ANN with the obtained results, the highest compressive strength is calculated using the EVPS and SA-EVPS algorithms.
 

---

In this study, experimental and computational approaches are used in order to develop and optimize self-compacting concrete mixes (Artificial neural network, EVPS metaheuristic algorithm, Taguchi method). Initially, ten basic mix designs were tested, and an artificial neural network was trained to predict the properties of these mixes. The network was then used to generate ten optimized mixes using the EVPS algorithm. Three mixes with the highest compressive strength were selected, and additional tests were conducted using the Taguchi approach. Inputting these results, along with the initial mix designs, into a second trained neural network, 10 new mix designs were tested using the network. Two of these mixes did not meet the requirements for self-compacting concrete, specifically in the U-box test. However, the predicted compressive strength results showed excellent agreement with low error percentages compared to the laboratory results, which indicates the effectiveness of the artificial neural network in predicting concrete properties, thus indicating that self-compacting concrete properties can be predicted with reasonable accuracy. The paper emphasizes the reliability and cost-effectiveness of artificial neural networks in predicting concrete properties. The study highlights the importance of providing diverse and abundant training data to improve the accuracy of predictions. The results demonstrate that neural networks can serve as valuable tools for predicting concrete characteristics, saving time and resources in the process. Overall, the research provides insights into the development of self-compacting concrete mixes and highlights the effectiveness of computational approaches in optimizing concrete performance.
 

---

The main objective of this paper is to optimize the size and layout of planar truss structures simultaneously. To deal with this challenging type of truss optimization problem, the center of mass optimization (CMO) metaheuristic algorithm is utilized, and an extensive parametric study is conducted to find the best setting of internal parameters of the algorithm. The CMO metaheuristic is based on the physical concept of the center of mass in space. The effectiveness of the CMO metaheuristic is demonstrated through the presentation of three benchmark truss layout optimization problems. The numerical results indicate that the CMO is competitive with other metaheuristics and, in some cases, outperforms them.
 

---

The main objective of this study is to predict the maximum inter-story drift ratios of steel moment-resisting frame (MRF) structures at different seismic performance levels using feed-forward back-propagation (FFBP) neural network models. FFBP neural network models with varying numbers of hidden layer neurons (5, 10, 15, 20, and 50) were trained to predict the maximum inter-story drift ratios of 5- and 10-story steel MRF structures. The numerical simulations indicate that FFBP neural network models with ten hidden layer neurons better predict the inter-story drift ratios at seismic performance levels for both 5- and 10-story steel MRFs compared to other neural network models.