Abstract
Artificial neural network (ANN) aimed to simulate the behavior of the nervous system as well as the human brain. Neural network models are mathematical computing systems inspired by the biological neural network in which try to constitute animal brains. ANNs recently extended, presented, and applied by many research scholars in the area of geotechnical engineering. After a comprehensive review of the published studies, there is a shortage of classification of study and research regarding systematic literature review about these approaches. A review of the literature reveals that artificial neural networks is well established in modeling retaining walls deflection, excavation, soil behavior, earth retaining structures, site characterization, pile bearing capacity (both skin friction and end-bearing) prediction, settlement of structures, liquefaction assessment, slope stability, landslide susceptibility mapping, and classification of soils. Therefore, the present study aimed to provide a systematic review of methodologies and applications with recent ANN developments in the subject of geotechnical engineering. Regarding this, a major database of the web of science has been selected. Furthermore, meta-analysis and systematic method which called PRISMA has been used. In this regard, the selected papers were classified according to the technique and method used, the year of publication, the authors, journals and conference names, research objectives, results and findings, and lastly solution and modeling. The outcome of the presented review will contribute to the knowledge of civil and/or geotechnical designers/practitioners in managing information in order to solve most types of geotechnical engineering problems. The methods discussed here help the geotechnical practitioner to be familiar with the limitations and strengths of ANN compared with alternative conventional mathematical modeling methods.
Similar content being viewed by others
Explore related subjects
Discover the latest articles, news and stories from top researchers in related subjects.Avoid common mistakes on your manuscript.
1 Introduction
Because of a large number of complicated problems in most engineering applications, engineers depend on computational intelligence as well as soft computing analysis instead of following huge complicated calculations [1, 2]. In engineering problems, much sophisticated statistical analysis and mathematical modeling are introduced in order to solve engineering problems [3,4,5]. Challenges associated with the reliable engineering design solution and development of technology complicated the geotechnical engineering environment even more [6, 7]. Certainly, investigating the engineering properties of rock and soil masses show uncertain and varied behavior due to their imprecise and complex natures. On the other hand, many other materials in the field of civil engineering (e.g., steel, timber, and concrete) show far more homogeneity and isotropy. The artificial neural networks (ANNs), based on the sophisticated mathematical models and advanced software tools, can help to assess all the reliable choices available with respect to a predefined project outcome [8, 9]. The ANN, however, need to be used along with one optimization algorithm to reduce the rate of error especially in complex problems such as compressed sensing [10]. ANN offers the required tools for geotechnical engineers, working in large consultant companies, to make a fast and in most cases better decisions to improve the quality of their performance and to reduce risks [11]. Numerous researchers have discussed the operation and structure of ANNs (e.g., Wang [12], Choobbasti et al. [13], Gandomi and Alavi [14], Mukhlisin et al. [15], Lian et al. [16], Salsani et al. [17], Bahrami et al. [18], Mert [19], Moayedi and Rezaei [20]). As an alternative and effective approach, which has been proved to have a degree of success and reliability [21], is mainly based on the data alone to define the parameters and structure of the model [8]. The ANN was used in numerous academic subjects and projects, such as risk assessment [22, 23], health and medical [24], image processing [25, 26], mathematics [27,28,29,30], early warnings related to geotechnical problems [31], geosciences and remote sensing [32], business and management [33], civil engineering [14, 34,35,36] and particularly to the geotechnical engineering as the main concern for this study [32, 37].
In recent years some scholars have successfully attempted to generate, extend and present the new utility determining tools and approaches, as well as ANNs methods and techniques into the field of civil engineering. This interesting topic has been reviewed several times by researchers such as Flood and Kartam [38], Flood and Kartam [39], Adeli [3], Lu et al. [40], Lazarevska et al. [34] and Li and Hao [41]. Indeed, the use of ANNs method in the geotechnical engineering problems, as the first multicriteria assessment method, was presented in the early 1990s by Bolt [42]. Different subjects have been studied using the ANNs method such as faults modeling [42], underground openings [43], braced excavation [44], pile integrity testing [45], pile bearing capacity [20, 46,47,48,49,50,51,52,53,54,55,56], predicting geotechnical parameters [1, 57,58,59], modeling tunnel boring machine (TBM) performance [60], kinematic soil pile interaction response parameters [61], slope stability [62]. There are very few research studies that classified and reviewed the ANNs application for these approaches in various areas such as; principles and understanding of NNs in civil engineering [3], shallow foundations [63], pile foundations [64], corrosion monitoring [65]. Various approaches have been suggested regarding the previous findings on the ANN application in geotechnical engineering. However, the conducted surveys were limited to specific subjects such as pile foundation [66], shallow foundation [63] or general subject of geotechnical engineering [8] and did not keep up with the new challenges and changing situation in the field of geotechnical engineering. Thus, the authors think that there is an absence of a systematic review from the recently published studies performed in the highlighted area. Also, the authors believe that there is a great demand for a comprehensive review paper, combining the available methods as well as current studies.
2 Literature review and distribution of the papers
Since the early 1900s, and up to the date of writing this paper, there are more than 4000 research scholar articles in the field of geotechnical engineering which indexed in the web of science (WOS). Distribution of papers published in the considered area, based on the source title, is presented in Table 1. In this regard, when the search narrows to the application of ANNs in the subject of geotechnical engineering only 152 articles, with a very limited number of source title, remained. Distribution of papers, based on the source title, in the use of ANN-based models in geotechnical engineering, is tabulated in Table 2. Figure 1 shows the subject of ANN application in the field of geotechnical engineering based on both total publications by year and sum of times cited by year. It can be seen that the number of publications has increased sharply from one publication in the year 2000 to twenty-two publications in the Year 2017. ANN-based methods have been applied increasingly as an effective methods in most geotechnical engineering subjects, including: tunneling [67, 68], mathematical constitutive modeling [69], underground openings [70, 71], geo-material properties [72, 73], bearing capacity of pile [20, 53, 64]; slope stability [47, 74,75,76,77,78,79]; liquefaction [80, 81], earth retaining structures [82, 83], soil swelling [84, 85], classification of soils [86, 87] and site characterization [88, 89]. Indeed, the fundamentals of modern ANN and hybrid ANN methods were developed (in some cases generated) in the 2000s and 1960s. The research and development of ANN methods increased between 2010 and 2017, but as presented in Fig. 1, it seems that the exponential (here means rapid increase) growth of this process continued. ANN application in pile bearing capacity prediction (both skin friction and end-bearing capacities), modeling soil behavior, earth retaining structures and excavations, site characterization, liquefaction, slope stability, and classification of soils are reviewed in this section. To make readers familiar with the geotechnical interest of research scholar authors listed the journals that mostly publish on the subject of geotechnical engineering (see Table 1). On the other hand, the distribution of papers in the subject of ANN application in geotechnical engineering according to the source title (see Table 2) is illustrated.
2.1 ANN application in modeling soil behavior
AttohOkine and Fekpe [90] employed adaptive NNs to model strength properties of local lateritic soils. They compared the results from generalized adaptive NNs (GANN) with some of the traditional back-propagation NN (BPNN) approaches for modeling the in situ soil strength properties based on raw California bearing ratio (CBR) data. Their results showed the GANN seems to be more effective. Zhu et al. [91] used recurrent NN (RNN) to model shearing characteristics of residual soil. The network was able to determine volumetric strains during shearing courses and abrupt changes in axial. Pal [92] modeled seismic liquefaction potential using a support vector machine (SVM). In this study, the data were collected from several field tests such as standard penetration test (SPT value) and cone penetration test (CPT) and utilized to assess the liquefaction potential using the SVM-based classification approach. The author concluded that the complex relationship between the liquefaction potential and different soil parameters can effectively be presented using the SVM. Pala et al. [93] employed the ANNs to analyses the dynamic soil-structure interaction of buildings. They used the back-propagation (BP) algorithm. The results showed that the solution time is quite fast and the analysis and modeling stages are minimized. The NNs has shown excellent performance for the solution of soil-structure interaction problems.
Nazzal and Tatari [94] and Park et al. [95] used genetic algorithms and ANN, respectively, to propose a practical model and predict the resilient modulus of subgrade soils. They concluded that the ANN-based predictive models work as a simple and reliable mathematical tool. Groholski and Hashash [96] developed of a framework for extracting dynamic soil performance and pore water pressure response from field data collected in downhole array test (after ASTM D7400). To represent pore water pressure generation during cyclic loading they introduced an ANN-based constitutive model. As a result, the successful presentation of the established model is well demonstrated. Distribution of the papers in the subject of ANN application in soil behavior according to publication years, research areas of their publication and source title is listed in Table 3.
2.2 ANN application in pile capacity prediction
Different techniques of ANN, along with experimental experiments, applied in other studies such as Moayedi and Rezaei [20] and Mosallanezhad and Moayedi [53], Nazir et al. [97] and Moayedi [98] to predict pile bearing capacity, pile settlement, pile skin friction and/or pile end-bearing capacity. One of the most basic researches on the pile is provided by Chan et al. [99]. They have released a training dataset using back-propagation neural network to develop a prediction model for the evaluation of the skin friction as well as end-bearing capacity in piles. After comparison between the generated networks, they generate more reliable outputs than a pile driving. In this regard, Ismail and Jeng [100] established a HON-PILE model (high-order neural network model) in order to model the load-settlement behavior of piles. Indeed, a total number of 121 research scholar articles on the applicability of ANNs on the pile bearing capacity (both lateral and axial) were indexed in the WOS (see Table 4). It can be seen that almost 86% of the publications are listed in the category of “engineering geology,” “engineering civil” and “geosciences multidisciplinary.”
2.3 ANN application in earth retaining structures
Research on the earth retaining structures has been always one of the main interests between the geotechnical engineers. Countries such as Peoples R China (20.743%), USA (12.384%), England (6.502%), Japan (5.882%), France (5.623%), Italy (4.664%), Canada (4.025%) provided the most published articles on the subject of retiming structures in the WOS. However, when it comes to the ANN applicability, there are very few studies on the applicability of the neural network on the estimation of retaining structures behaviors. Studies such as Li et al. [101] and Chen and Wang [102] worked on the deformation prediction of the pile-anchor retaining structure. Li et al. [101] investigated on the application of neural network to predict displacement of deep foundation pit retaining structure. Their research indicated that the soft computing method is a useful and valid method for prediction of deformation in the foundation pit. Similarly, Chen and Wang [102] used ANN to predict the deformation characteristics of pile-anchor structure (one of the established retaining system) in deep foundation pit. Distribution of the papers on the applicability of the ANNs in earth retaining structures as well as excavation according to the source title is tabulated in Table 5.
The ANN was also used to formalize and synthesize data derived from FE modeling studies. Up to the date of writing this article, there are 48 research scholars indexed in the WOS on the use of the neural network on excavations. In general, the input parameters used in the provided models were the wall stiffness, the soil layer thickness/excavation width ratio, excavation width, soil unit weight, soil undrained shear strength, the height of excavation, and undrained soil modulus/shear strength ratio [103, 104]. The maximum wall deflection was selected as the only output. For instance in braced excavation and in soft clay Goh et al. [44] established an ANN model to estimate maximum wall deflections (normally in the top of the wall). The results produced high accuracy with coefficients of correlation equivalent to 0.984 and 0.967 for the training and testing datasets, respectively.
Kwon and Wilson [105] used NNs to explore the impact of a deep excavation on other adjacent excavations. They applied NNs to investigate the influence of each parameter and the deformation increase on the deformation variation derived from extensometer measurements. Jan et al. [106] also investigated the use of ANN prediction model in the deep excavation. To collect the required data for training and verification, eighteen different case histories of deep excavations, with a minimum of four and maximum of seven excavation (construction) stages, were selected. The results of simulation show that not only the ANN can determine the maximum deflection of the diaphragm wall but also it can predict the location and the magnitude which the maximum deformations occur. Chua and Goh [107] used Bayesian NNs to determine wall deformation behavior in a deep excavation. It is found that the trained model could be used as a reliable and simple prediction tool. They could calculate the maximum wall deformation. Huang et al. [108] investigated the ANN-based reliability analysis for deep excavation. Chern et al. [109] applied a neural network to predict successfully lateral wall deflection in the top–down excavation. Yu et al. [110] used ANNs in artificial intelligent prediction model in order to calculate shallow settlement adjacent to the excavation of a foundation pit.
2.4 ANN application in site characterization
In all geotechnical engineering problems, site characterization is known as an important step that needs to be considered. It is essential to explore the subsurface before doing any project analysis. Many researchers such as Huang and Siller [111], Yilmaz et al. [112], Garcia-Fernandez and Jimenez [113], Orhan et al. [114], Kim et al. [115], Cao et al. [116], Wang [117], Aladejare and Wang [118] and Roy and Jakka [119] worked on this subject. Several researchers also applied ANNs to improve the estimation of the site characterization. For example, in order to represent the data obtained from borehole Huang and Siller [111] developed a fuzzy set-based model which uses to infer the subsurface profile. Bagtzoglou and Hossain [88] used RBFN for hydrologic inversion. The RBFN was used as a reliable method in the context of site characterization. In this regard, Samui and Sitharam [120] employed a relevance vector machine and least-square (LS) SVM based on corrected SPT data in order to estimate site characterization. Samui and Sitharam [121] modeled site characterization using ANN and Kriging. An extensive number of data (2700 field SPT values) were collected from SPTs in 3D subsurface of Bangalore, India.
2.5 ANN application in liquefaction
Soil liquefaction defined as a phenomenon whereby a partially or fully saturated soil (in most cases sands) substantially loses stiffness and strength in response to a specifically applied stress. In this regard, the applied stresses usually are induced by earthquake shaking causing the soil to behave like a liquid (with no shear strength). The liquefaction often leads to extensive damage and very high defamation to most infrastructures. The reason behind such large deformation is that the soil will lose its basic shear strength due to an increase in the pore pressure. Indeed, the soil liquefaction is introduced as one of the multicriteria tasks to assess in geotechnical earthquake engineering. Many experts stated that the assessment of soil liquefaction, due to a lot of variables, is the most complicated phenomena in geotechnical engineering [122,123,124,125,126]. Table 6 presents the distribution of the papers on the applicability of the ANNs in liquefaction evaluation. As one of the earliest researches on the use of ANN in liquefaction assessment, Goh [127] used ANN to solve the complex relationship between different soil parameters and seismic loading applied in order to explore liquefaction potential. The network model was trained using 13 case recorded real-world earthquakes. The study included eight input variables (SPT value, the mean grain size, the fines content, the earthquake magnitude, the equivalent dynamic shear stress, the total and effective stress, and the maximum horizontal acceleration at ground surface) and only one output variable. The output was assigned a no liquefaction (binary value of 0) and, for sites with extensive or moderate liquefaction potential (value of 1). The results gained by the proposed neural network model were compared with the conventional mathematical method that further developed by Seed et al. [128]. In comparison with the success rate of 84% from the method presented by Seed et al. [128], the study revealed that the ANN model gave reliable predictions in 95% of cases. Goh [129], Juang et al. [130], Liu et al. [131] and Chern and Lee [81] used ANN to evaluate liquefaction resistance based on raw CPT data. The results from the neural network showed a minimum success rate of 94%, which is acceptable in comparison with previous evaluation method presented by Shibata and Teparaksa [132] with a success rate of 84%. Wang and Rahman [133] presented a neural network-based model for liquefaction phenomena caused by horizontal ground displacement. Young-Su and Byung-Tak [134] used ANNs to predict liquefaction resistance of sands. Hsu et al. [135] applied ANN to model liquefaction resistance. In their study, a total of 217 sets of shear wave velocity data, 31 from Taiwan after 1999 and 186 from the previous reports and studies, were collected and synthesized. Zhang et al. [136] evaluated soil liquefaction based on multivariate adaptive regression splines and capacity energy concept.
On the other hand, many others also investigated the use of SVM method in landslide assessment. For instance, Pal [92] employed SVM-based modeling to assess liquefaction potential induced by seismic loading. The data were collected from actual SPT and CPT field data. In fact, SVMs can provide better performance and required few user-defined parameters in comparison with the ANN approach. Similarly, Goh and Goh [137] explored the use of SVM in geotechnical engineering with the main focus on seismic liquefaction data. They trained and tested the SVM model based on a relatively large data set comprising 226 field records of CPT measurements and liquefaction performance. The results of classification showed that the overall success rate for the entire data set is 98%.
2.6 ANN application in slope stability
The slope stability analysis has been always a big challenge for geotechnical engineers. This is of course because of a wide variety of variables affecting the slope stability. Indeed, for most civil engineers working with software (i.e., include many details and variables) is not usually acceptable. Researchers such as Lu and Rosenbaum [138], Li and Liu [139], Liu et al. [140], Zhang et al. [36], Aghajani et al. [141], Rahul et al. [142], Gordan et al. [143], Kostic et al. [11] and Li et al. [144] studied on the subject of ANN application in slope stability. In their approaches, the input parameters were horizontal profile, gradient, location, height, vertical profile, soil texture, geological origin, the direction of slopes, depth of weathering, vegetation, maximum precipitation hour, and maximum daily precipitation. The slope failure potential was taken as the main output. Table 7 presents the distribution of the papers on the applicability of the ANNs in slope stability based on the source title.
Most practical applications prove that the estimation of slope stability analysis using ANN is achievable. In this regard, a well-trained neural network learning system reveals an extremely fast convergence, a better generalization and a high degree of accuracy for the slope stability problems. Lu and Rosenbaum [138] employed Grey and ANNs systems for the prediction of slope stability. Li and Liu [139] used AI forecast procedures for the slope stability. Liu et al. [140] applied a fast robust NN model called Extreme Learning Machine (ELM) to find a solution for the prediction of slope stability problems. After comparing several ANN techniques, the results prove that, in most of the common slope stability analysis, the ELM act as a helpful way to the genetic algorithm and the GRNN-based models. Gordan et al. [143] combined Particle Swarm Optimization (PSO) and neural network to predict slope stability induced by seismic loading. Kostic et al. [11] developed a model for prediction slope stability based on the ANNs. In this regard, they employed multilayer feed-forward network. The obtained results indicated a high level of statistical reliability.
2.7 ANN application in landslides assessment
The use of ANN in landslide susceptibility mapping is indeed well established. Perhaps the most well-known applicability of the ANN is in the subject of landslides [16, 35]. Indeed, the use of ANNs method in the landslide hazard mapping problems, as one of the multicriteria evaluation method, was introduced in the late 1990s by Yamagami et al. [145], Cai et al. [146] and Kobayashi et al. [147]. Different subjects have been studied using the ANNs method in landslide such as risk assessment [148], susceptibility analysis [149, 150], prediction [151, 152], earthquake-induced/triggering [153, 154], susceptibility mapping by geographical information system [155, 156]. Figure 2 shows the applicability of the NN in the subject of landslide based on both total publications by year and sum of times cited by year (Fig. 2). It can be seen that the number of publications has increased sharply from 10 publications in the year 2005 to 85 publications in the year 2016. Similarly, the number of citations per year raised to 3200 citations per year. This is showing that the subject is still one of the main interests of the researchers. Table 8 presents the distribution of the papers on the applicability of the ANNs in landslide assessment.
3 Research methodology
For the presented review study a new technique called PRISMA method, proposed by Moher et al. [157] was used. It is important to note that the PRISMA statement has two different parts; (1) systematic reviews and (2) meta-analysis. This method is well described in Shamseer et al. [158] and Mardani et al. [159]. Systematic reviews found out about research topics in order to provide summaries from the objectives and what has been written in the literature [157]. Generally, the systematic review part tries to present a full overview of research scholars performed on a specific subject (here it was applicability of ANNs in geotechnical engineering) until the present date. However, when it comes to the meta-analysis part, it offers main findings of statistical approach from previously published works. The main objective of the PRISMA method is to assist practitioners and researchers in finding a complete, simple, and clear literature review [158, 160]. There are many good examples of previous studies which used the PRISMA method in various fields. In overall they intend to present a comprehensive review of the most recent published articles as a literature review [158, 160, 161]. In our review study, we considered three main steps (1) search in articles indexed in WOS, (2) selection of the eligible published articles, and (3) extraction of datasets and summarizing the data [159].
4 Literature search
The database of WOS was used to provide a systematic review of applications and methodologies of ANN-based models in the subject of geotechnical engineering. The most recent published papers were found based on the searching several keywords such as bearing capacity, pile, retaining structures, excavation, site characterization, liquefaction, liquefaction susceptibility, slope stability, landslides and different ANN-based model approaches use in the field of geotechnical engineering. We have chosen those articles from the literature which were published between 2010 and 2018. In this regard, and according to our strategy search, a total of 734 scholarly papers were extracted. In the next step, and after a double check, we duplicated articles with repeated information. As a result, 108 papers were selected (see Fig. 3). Then, to remove the duplicated articles from the selected list, we eliminated 22 records due to duplicates. In the end, screened papers were selected based on different structures of the papers such as titles, keywords, topics, abstracts, and studies that were unrelated to the topic of this review were removed.
4.1 ANN
The ANN is known as a tool to model the multicriteria and complex systems involved in approximation problems. The theoretical background of the ANN is comprehensively discussed by Hill et al. [162], and Wang [12]. Tables 9 and 10 are listed the distribution of papers in the subject of ANN applicability in the field of geotechnical engineering based on the research areas and source title, respectively. A typical structure of most neural network-based models consists of a number of nodes (or processing elements, PEs), that are typically arranged in form of several layers: one or more hidden layers, an output layer and an input layer [8]. A general flowchart for the ANN models is shown in Fig. 4.
4.2 FFNN
FFNN is a simple NN model used for modeling numerous nonlinear phenomena [164, 165]. As established by Hornik [164] it is a standard multilayer FFNN which are able to give a prediction value, in a very specific and satisfying sense, to any measurable function. One interesting capability of FFNN method is to pre-mapping the input data before sending it to the hidden layer for further processing. The selected data set for the training the network is firstly multiplied by Wh (a specific weight matrix). The results will later be added to a bias vector (bh). As in the final step, a transfer function will be applied (e.g., during the data processing in the hidden layer). In fact, the process of network training is a modification of the introduced weight matrixes and bias vectors. This is because the outcome of the trained network needs to be minimized (the distance between both training data and network results). The increase in the number of neurons and layers, for instance, to get a better result, is not desirable. This is because such an increase can lead to a more complex network which later produces problems in both convergence and training of the networks. In this regard, the FFNN method is well described in the literature. For instance, several good examples of using the FFNN technique in the subject of geotechnical engineering are in Han et al. [166], Uncuoglu et al. [167], Lian et al. [168] and Protopapadakis et al. [169]. The FFNN method has been very popular among the researchers. Figure 5 shows the recent use of FFNN method in the scholar papers between 1999 and 2017. In this regard, the total publications by year increased to about 550 publications (indexed in the web of science only) and the summation of times cited by year.
4.3 RBFN
RBFN is another widely used MLP network. The RBFN is also well recommended because of its less time consuming during the training of the networks. The structure of an RBFN is like a single layer feed-forward network as shown in Fig. 6. The structure of the RBFN and its application in the subject of geotechnical engineering is well described in Mustafa et al. [170], Shu and Gong [171], Kang et al. [172] and Moayedi and Hayati [74]. The only difference is that in this function (e.g., in all the hidden layers) a radial basis (also called radbas) function is used in comparison with the FFNN function which it was a sigmoid function. The RBFN can simply be defined as Eq. (1).
The RBFN method has been employed successfully in various research areas. Distribution of papers in the subject of RBFN applicability based on the research areas is tabulated in Table 11. However, the use of RBFN in the field of geotechnical engineering is still remained unknown. In recent years, there are very few studies use RBFN in the field of geotechnical engineering; modeling free-surface seepage flow [173], reliability analysis [174], predicting rock mass deformation modulus [175], three-dimensional simulations of tensile cracks [176], reliability analysis of geotechnical engineering [177], groutability prediction of permeation grouting [178, 179].
4.4 GRNN
GRNN includes four separate layers: the first layer is the input layer where the data will be introduced to the network and prepared for the training. The second layer is the pattern layer which follows a specific pattern function. Thirdly, it is the summation layer and finally, the results are generated from the output layer. The use of GRNN has been widely used in the subject of civil engineering and more particularly in the field of geotechnical engineering. Many researchers such as Ibric et al. [180], Pal and Deswal [181], Jiang et al. [182], Goorani and Hamidi [183] and Rajesh and Choudhury [184] used the GRNN through their studies. In this regard, the structure and application of GRNN are also well discussed in Cigizoglu and Alp [185] and Li et al. [186]. In training the neural network, linear activation and RBFN are used in both output and hidden layers. Each pattern layer unit is associated with the two different neurons in the summation layer (called D and S summation neurons). D summation neuron is used to calculate un-weighted outputs of pattern neurons, while S summation neuron computes the sum of weighted responses of the pattern layer [187].
The term D is Gaussian function and it is defined in Eq. (3):
where n the training pattern’s number, m the number of elements applied in the input vector, yi the weight connection (connection between the ith neuron in the pattern layer and the neuron in the S summation), D the defined Gaussian function, xk and xik are the jth elements of x and xi, respectively.
A search in the topic of GRNN shows a total number of 1771 articles in different research areas. Distribution of research scholar papers in the subject of GRNN applicability based on the research areas is listed in Table 12. However, similar to the RBFN, the applicability of the GRNN method in the field of civil engineering is still considered a new topic. There are only a few studies that use the GRNN in the field of geotechnical engineering; compressive strength analysis of reinforced soil [188], slope stability inference [36, 140], lateral load bearing capacity modeling of piles [189], determination of ultimate bearing capacity of concrete driven piles in sand [190], expansive soil characterization [191] and three-dimensional site characterization [89].
4.5 Adaptive neuro-fuzzy inference system (ANFIS)
ANFIS is one of the strong learning systems for prediction of complex functions. ANFIS was first proposed by Jang and Sun [192]. Among fuzzy inference systems, it is one of the most commonly used training systems. In fact, ANFIS uses a Takagi–Sugeno fuzzy inference system (FIS). The structure and procedure of the ANFIS are presented and discussed by Jang [193]. General use of ANFIS in geotechnical engineering is well described in other studies such as Cabalar et al. [194]. However, the most common use of ANFIS is in the subject of landslide susceptibility mapping [195, 196], through landslide risk management [197, 198], rock-cutting trencher [199], constitutive modeling [200], prediction of uniaxial strength of rocks [201], liquefaction prediction [202], swelling potential [203], and permeability estimation [204].
4.6 Imperialist competitive algorithm (ICA)
The imperialist competitive algorithm function (also called ICA) is a global method of search population-based that was firstly proposed by Atashpaz-Gargari and Lucas [205] an followed by many other researchers such as Ahmadi et al. [206], Marto et al. [207], Mosallanezhad and Moayedi [53] and Moayedi and Armaghani [50]. The ICA has been used in many optimization problems. This is because it involves a procedure similar to many other evolutionary algorithms such as those used by Thangavelautham et al. [208], Manouchehrian et al. [209], Lian et al. [210] and Gandomi and Kashani [211]. The imperialist competitive algorithm begins with a candidate solution (or initial population), which, along with the imperialist competitive algorithm itself, consists of many countries [205]. In this step, all countries are separated into two main categories (shown in Fig. 7): (1) some of the best countries which called imperialists) and (2) the remaining countries which called colonies. In order to make an empire, first, the colonies required to be distributed through the best countries (called here imperialists or stronger countries). The distribution of the colonies is according to the relative strength of the countries in which the stronger countries could get a higher number of colonies. This competition will continue as the empires intend to expand their territories and control over more colonies. At the end of the competition algorithm (as mentioned by title imperialist competitive algorithm), the stronger empires expanded their power by taking control of weaker colonies. The process is like variables with higher relevancy could impact more on the output layer. Once a predefined stopping criterion is satisfied, the process will stop. A more detailed description of the designed steps in the imperialist competitive algorithm alone is discussed the literature by researchers such as Ghorbani and Jokar [212], and Al Dossary and Nasrabadi [213]. An overview of the imperialist competitive algorithm is depicted in Fig. 8.
5 Articles eligibility
For the purpose of eligibility, in this step of review, we independently reviewed each of the full text. However, the selected articles were shortlisted in the last step of the review. The shortlisted articles were selected according to the main objectives of this review paper on the subject of applicability of soft computing and ANN-based techniques in the field of geotechnical engineering. In this regard manuscript that applied ANN (e.g., any source of neural network modeling approach) in the subject of geotechnical engineering were chosen. Noteworthy, editorial notes, book chapters, master and doctoral dissertations, unpublished working papers, a non-English language written papers and textbooks were not involved. In addition, several previous studies have employed other techniques such as geostatistical, limit equilibrium, FE etc.; therefore, in this step also we did not include those studies. At the end of article eligibility decision, we selected 196 articles which used directly the ANN models in the geotechnical engineering. These articles could meet the considered selection criteria.
6 Summarizing and data extraction
In this step of review, some required data was collected and finally 196 articles were summarized and reviewed. In flowing, all selected papers were categorized into different application areas including; engineering, geology, computer science, water resources, environmental sciences, ecology, mechanics, materials science, geochemistry geophysics, mathematics, construction building technology, meteorology atmospheric sciences, mining mineral processing, instruments instrumentation, oceanography, science technology other topics, transportation and operations research management science (see Table 13). In addition, articles were reviewed and summarized according to various criteria such as journals and conferences names, the year of publication, authors, the method and technique used, research objectives, solution and modeling, and results and findings.
7 Conclusion
This review paper discussed the applications and theory with several ANN recent developments in the subject of geotechnical engineering. The ANN is introduced as a reliable tool for complex problems. The authors believe the ANN-based model, as a user-friendly and time-saving, is a good alternative to the FEM and conventional mathematical modeling. This is because, normally, problems in the field of geotechnical engineering evolve with many variables which make it hard to be modeled using conventional mathematical methods. Since the ANN-based methods are able to (1) rank the variables and alternatives, (2) evaluate stronger and weaker criteria, and (3) performing the comparative analysis, recently, ANN research interest increased largely in the geotechnical problems. In contrast, it is important to note that the neural networks have begun to replace by deep structured learning. Throughout this review paper, the use of ANN in the subject of geotechnical engineering were categorized into twenty-five research areas. Apart from the research areas, articles were classified according to the authors, source titles (either journal or conferences names), the year of publication, research areas, the used technique, solution and modeling, and outcomes.
References
Lee SJ, Lee SR, Kim YS (2003) An approach to estimate unsaturated shear strength using artificial neural network and hyperbolic formulation. Comput Geotech 30(6):489–503
Pujitha AK, Sivaswamy J (2018) Solution to overcome the sparsity issue of annotated data in medical domain. CAAI Trans Intell Technol 3(3):153–160
Adeli H (2001) Neural networks in civil engineering: 1989–2000. Comput-Aided Civi Infrastruct Eng 16(2):126–142
Panwar P, Michael P (2018) Empirical modelling of hydraulic pumps and motors based upon the Latin hypercube sampling method. Int J Hydromechatron 1(3):272–292
Gao W, Wang W, Dimitrov D, Wang Y (2018) Nano properties analysis via fourth multiplicative ABC indicator calculating. Arab J Chem 11(6):793–801
Zhang RL, Lowndes IS (2010) The application of a coupled artificial neural network and fault tree analysis model to predict coal and gas outbursts. Int J Coal Geol 84(2):141–152
Moayedi H, Huat B, Thamer A, Torabihaghighi A, Asadi A (2010) Analysis of longitudinal cracks in crest of Doroodzan Dam. Electron J Geotech Eng, USA (15D):337–347
Shahin MA, Jaksa MB, Maier HR (2001) Artificial neural network applications in geotechnical engineering. Aust Geomech 36(1):49–62
Johnson JL (2018) Design of experiments and progressively sequenced regression are combined to achieve minimum data sample size. Int J Hydromechatron 1(3):308–331
Zhou Y, Sun Q, Liu J (2018) Robust optimisation algorithm for the measurement matrix in compressed sensing. CAAI Trans Intell Technol 3(3):133–139
Kostic S, Vasovic N, Todorovic K, Samcovic A (2016) Application of artificial neural networks for slope stability analysis in geotechnical practice. In: 2016 13th Symposium on neural networks and applications (neural) pp 89–94
Wang S-C (2003) Artificial neural network, interdisciplinary computing in java programming. Springer, Berlin, pp 81–100
Choobbasti AJ, Farrokhzad F, Barari A (2009) Prediction of slope stability using artificial neural network (case study: Noabad, Mazandaran, Iran). Arab J Geosci 2(4):311–319
Gandomi AH, Alavi AH (2012) A new multi-gene genetic programming approach to non-linear system modeling. Part II: geotechnical and earthquake engineering problems. Neural Comput Appl 21(1):189–201
Mukhlisin M, El-Shafie A, Taha MR (2012) Regularized versus non-regularized neural network model for prediction of saturated soil-water content on weathered granite soil formation. Neural Comput Appl 21(3):543–553
Lian C, Zeng ZG, Yao W, Tang HM (2014) Ensemble of extreme learning machine for landslide displacement prediction based on time series analysis. Neural Comput Appl 24(1):99–107
Salsani A, Daneshian J, Shariati S, Yazdani-Chamzini A, Taheri M (2014) Predicting roadheader performance by using artificial neural network. Neural Comput Appl 24(7–8):1823–1831
Bahrami A, Monjezi M, Goshtasbi K, Ghazvinian A (2011) Prediction of rock fragmentation due to blasting using artificial neural network. Eng Comput 27(2):177–181
Mert E (2014) An artificial neural network approach to assess the weathering properties of sancaktepe granite. Geotech Geol Eng 32(4):1109–1121
Moayedi H, Rezaei A (2017) An artificial neural network approach for under reamed piles subjected to uplift forces in dry sand. Neural Comput Appl 28:1–10
Shu SX, Gong WH (2016) An artificial neural network-based response surface method for reliability analyses of c-phi slopes with spatially variable soil. China Ocean Eng 30(1):113–122
Dong C, Dong XC, Gehman J, Lefsrud L (2017) Using BP neural networks to prioritize risk management approaches for China’s unconventional shale gas industry. Sustainability 9(6):18
Adams MD, Kanaroglou PS (2016) Mapping real-time air pollution health risk for environmental management: combining mobile and stationary air pollution monitoring with neural network models. J Environ Manag 168:133–141
Lisboa PJG (2002) A review of evidence of health benefit from artificial neural networks in medical intervention. Neural Netw 15(1):11–39
Egmont-Petersen M, de Ridder D, Handels H (2002) Image processing with neural networks—a review. Pattern Recognit 35(10):2279–2301
Ayyildiz M, Cetinkaya K (2017) Predictive modeling of geometric shapes of different objects using image processing and an artificial neural network. Proc Inst Mech Eng Part E-J Process Mech Eng 231(6):1206–1216
Gao W, Dimitrov D, Abdo H (2018) Tight independent set neighborhood union condition for fractional critical deleted graphs and ID deleted graphs. Discrete Contin Dyn Syst-S 12(4&5):711–721
Gao W, Guirao JLG, Basavanagoud B, Wu J (2018) Partial multi-dividing ontology learning algorithm. Inf Sci 467:35–58
Gao W, Guirao JLG, Abdel-Aty M, Xi W (2019) An independent set degree condition for fractional critical deleted graphs. Discret Contin Dyn Syst-S 12(4&5):877–886
Gao W, Wu H, Siddiqui MK, Baig AQ (2018) Study of biological networks using graph theory. Saudi J Biol Sci 25(6):1212–1219
Chou J-S, Thedja JPP (2016) Metaheuristic optimization within machine learning-based classification system for early warnings related to geotechnical problems. Autom Constr 68:65–80
Lary DJ, Alavi AH, Gandomi AH, Walker AL (2016) Machine learning in geosciences and remote sensing. Geosci Front 7(1):3–10
Wong BK, Bodnovich TA, Selvi Y (1997) Neural network applications in business: a review and analysis of the literature (1988–1995). Decis Support Syst 19(4):301–320
Lazarevska M, Knezevic M, Cvetkovska M, Trombeva-Gavriloska A (2014) Application of artificial neural networks in civil engineering. Teh Vjesn 21(6):1353–1359
Chen JJ, Zeng ZG, Jiang P, Tang HM (2016) Application of multi-gene genetic programming based on separable functional network for landslide displacement prediction. Neural Comput Appl 27(6):1771–1784
Zhang ZF, Liu ZB, Zheng LF, Zhang Y (2014) Development of an adaptive relevance vector machine approach for slope stability inference. Neural Comput Appl 25(7–8):2025–2035
Chou JS, Thedja JPP (2016) Metaheuristic optimization within machine learning-based classification system for early warnings related to geotechnical problems. Autom Constr 68:65–80
Flood I, Kartam N (1994) Neural networks in civil engineering.1. principles and understanding. J Comput Civ Eng 8(2):131–148
Flood I, Kartam N (1994) Neural networks in civil engineering.2. systems and application. J Comput Civ Eng 8(2):149–162
Lu PZ, Chen SY, Zheng YJ (2012) Artificial intelligence in civil engineering. Math Probl Eng 145974:1–22
Li J, Hao H (2016) A review of recent research advances on structural health monitoring in Western Australia. Struct Monit Maint 3(1):33–49
Bolt G (1991) Fault models for artificial neural networks. IEEE, Piscataway
Lee C, Sterling R (1992) Identifying probable failure modes for underground openings using a neural network. Int J Rock Mech Min Sci 29(1):49–67
Goh ATC, Wong KS, Broms BB (1995) Estimation of lateral wall movements in braced excavations using neural networks. Can Geotech J 32(6):1059–1064
Watson JN, Fairfield CA, Wan C, Sibbald A (1995) The use of artificial neural networks in pile integrity testing. Civil Comp Press, Edinburgh
Lee IM, Lee JH (1996) Prediction of pile bearing capacity using artificial neural networks. Comput Geotech 18(3):189–200
Niroumand H, Kassim KA, Nazir R, Faizi K, Adhami B, Moayedi H, Loon W (2012) Slope stability and sheet pile and contiguous bored pile walls. Electron J Geotech Eng 17:19–27
Moayedi H, Nazir R, Mosallanezhad M (2015) Determination of reliable stress and strain distributions along bored piles. Soil Mech Found Eng 51(6):285–291
Nazir R, Moayedi H, Mosallanezhad M, Tourtiz A (2015) Appraisal of reliable skin friction variation in a bored pile. Proc Inst Civ Eng-Geotech Eng 168(1):75–86
Moayedi H, Armaghani DJ (2017) Optimizing an ANN model with ICA for estimating bearing capacity of driven pile in cohesionless soil. Eng Comput 34(2):347–356
Moayedi H, Mosallanezhad M (2017) Uplift resistance of belled and multi-belled piles in loose sand. Measurement 109:346–353
Moayedi H, Mosallanezhad M, Nazir R (2017) Evaluation of maintained load test (MLT) and pile driving analyzer (PDA) in measuring bearing capacity of driven reinforced concrete piles. Soil Mech Found Eng 54(3):150–154
Mosallanezhad M, Moayedi H (2017) Developing hybrid artificial neural network model for predicting uplift resistance of screw piles. Arab J Geosci 10(22):10
Nazir R, Moayedi H, Subramaniam P, Gue S-S (2017) Application and design of transition piled embankment with surcharged prefabricated vertical drain intersection over soft ground. Arab J Sci Eng 43:1–10
Moayedi H, Hayati S (2018) Applicability of a CPT-based neural network solution in predicting load-settlement responses of bored pile. Int J Geomech 18(6):06018009
Moayedi H, Hayati S (2018) Artificial intelligence design charts for predicting friction capacity of driven pile in clay. Neural Comput Appl. https://doi.org/10.1007/s00521-018-3555-5
Asadi A, Moayedi H, Huat BB, Boroujeni FZ, Parsaie A, Sojoudi S (2011) Prediction of zeta potential for tropical peat in the presence of different cations using artificial neural networks. Int J Electrochem Sci 6(4):1146–1158
Asadi A, Moayedi H, Huat BBK, Parsaie A, Taha MR (2011) Artificial neural networks approach for electrochemical resistivity of highly organic soil. Int J Electrochem Sci 6(4):1135–1145
Asadi A, Shariatmadari N, Moayedi H, Huat BB (2011) Effect of MSW leachate on soil consistency under influence of electrochemical forces induced by soil particles. Int J Electrochem Sci 6(7):2344–2351
Benardos AG, Kaliampakos DC (2004) Modelling TBM performance with artificial neural networks. Tunn Undergr Space Technol 19(6):597–605
Ahmad I, El Naggar M, Khan AN (2007) Artificial neural network application to estimate kinematic soil pile interaction response parameters. Soil Dyn Earthq Eng 27(9):892–905
Chakraborty A, Goswami D (2017) Prediction of slope stability using multiple linear regression (MLR) and artificial neural network (ANN). Arab J Geosci 10(17):11
Shahin MA (2015) A review of artificial intelligence applications in shallow foundations. Int J Geotech Eng 9(1):49–60
Fatehnia M, Amirinia G (2018) A review of genetic programming and artificial neural network applications in pile foundations. Int J Geo-Eng 9(1):20
Mabbutt S, Picton P, Shaw P, Black S (2012) Review of artificial neural networks (ANN) applied to corrosion monitoring. In: Ball A, Mishra R, Gu F, Rao BKN (eds) 25th international congress on condition monitoring and diagnostic engineering. Iop Publishing Ltd., Bristol
Shahin MA (2016) State-of-the-art review of some artificial intelligence applications in pile foundations. Geosci Front 7(1):33–44
Lai JX, Qiu JL, Feng ZH, Chen JX, Fan HB (2016) Prediction of soil deformation in tunnelling using artificial neural networks. Comput Intell Neurosci 16:33
Alimoradi A, Moradzadeh A, Naderi R, Salehi MZ, Etemadi A (2008) Prediction of geological hazardous zones in front of a tunnel face using TSP-203 and artificial neural networks. Tunn Undergr Space Technol 23(6):711–717
Alavi AH, Gandomi AH (2011) A robust data mining approach for formulation of geotechnical engineering systems. Eng Comput 28(3–4):242–274
Zhang WG, Goh ATC (2016) Predictive models of ultimate and serviceability performances for underground twin caverns. Geomech Eng 10(2):175–188
Zhang WG, Goh ATC (2015) Regression models for estimating ultimate and serviceability limit states of underground rock caverns. Eng Geol 188:68–76
Asr AA, Javadi A (2016) Air losses in compressed air tunnelling: a prediction model. Proc Inst Civ Eng-Eng Comput Mech 169(3):140–147
Latifi N, Vahedifard F, Ghazanfari E, Horpibulsuk S, Marto A, Williams J (2018) Sustainable improvement of clays using low-carbon nontraditional additive. Int J Geomech 18(3):10
Moayedi H, Hayati S (2018) Modelling and optimization of ultimate bearing capacity of strip footing near a slope by soft computing methods. Appl Soft Comput 66:208–219
Moayedi H, Huat B, Kazemian S, Asadi A (2010) Optimization of shear behavior of reinforcement through the reinforced slope. Electron J Geotech Eng
Moayedi H, Huat BB, Asadi A (2010) Strain absorption optimization of reinforcement in geosynthetic reinforced slope-experimental and FEM modeling. Electron J Geotech Eng, USA 15
Nazir R, Ghareh S, Mosallanezhad M, Moayedi H (2016) The influence of rainfall intensity on soil loss mass from cellular confined slopes. Measurement 81:13–25
Nazir R, Moayedi H (2014) Soil mass loss reduction during rainfalls by reinforcing the slopes with the surficial confinement. World Academy of Science, Engineering and Technology. Int J Geol Environ Eng 8(6):381–384
Raftari M, Kassim KA, Rashid ASA, Moayedi H (2013) Settlement of shallow foundations near reinforced slopes. Electron J Geotech Eng 18:797–808
Shahri AA (2016) Assessment and prediction of liquefaction potential using different artificial neural network models: a case study. Geotech Geol Eng 34(3):807–815
Chern SG, Lee CY (2009) CPT-based simplified liquefaction assessment by using fuzzy-neural network. J Mar Sci Technol-Taiwan 17(4):326–331
Calabrese A, Lai CG (2013) Fragility functions of blockwork wharves using artificial neural networks. Soil Dyn Earthq Eng 52:88–102
Moayedi H, Huat BB, Moayedi F, Asadi A, Parsaie A (2011) Effect of sodium silicate on unconfined compressive strength of soft clay. Electron J Geotech Eng 16:289–295
Garg A, Garg A, Tai K, Barontini S, Stokes A (2014) A computational intelligence-based genetic programming approach for the simulation of soil water retention curves. Transp Porous Media 103(3):497–513
Erzin Y (2007) Artificial neural networks approach for swell pressure versus soil suction behaviour. Can Geotech J 44(10):1215–1223
Latifi N, Marto A, Eisazadeh A (2016) Experimental investigations on behaviour of strip footing placed on chemically stabilised backfills and flexible retaining walls. Arab J Sci Eng 41(10):4115–4126
Latifi N, Rashid ASA, Siddiqua S, Abd Majid MZ (2016) Strength measurement and textural characteristics of tropical residual soil stabilised with liquid polymer. Measurement 91:46–54
Bagtzoglou AC, Hossain F (2009) Radial basis function neural network for hydrologic inversion: an appraisal with classical and spatio-temporal geostatistical techniques in the context of site characterization. Stoch Environ Res Risk Assess 23(7):933–945
Juang CH, Jiang T, Christopher RA (2001) Three-dimensional site characterisation: neural network approach. Geotechnique 51(9):799–809
AttohOkine NO, Fekpe ESK (1996) Strength characteristics modeling of lateritic soils using adaptive neural networks. Constr Build Mater 10(8):577–582
Zhu JH, Zaman MM, Anderson SA (1998) Modelling of shearing behaviour of a residual soil with recurrent neural network. Int J Numer Anal Methods Geomech 22(8):671–687
Pal M (2006) Support vector machines-based modelling of seismic liquefaction potential. Int J Numer Anal Methods Geomech 30(10):983–996
Pala M, Caglar N, Elmas M, Cevik A, Saribiyik M (2008) Dynamic soil-structure interaction analysis of buildings by neural networks. Constr Build Mater 22(3):330–342
Nazzal MD, Tatari O (2013) Evaluating the use of neural networks and genetic algorithms for prediction of subgrade resilient modulus. Int J Pavement Eng 14(4):364–373
Park HI, Kweon GC, Lee SR (2009) Prediction of resilient modulus of granular subgrade soils and subbase materials using artificial neural network. Road Mater Pavement Des 10(3):647–665
Groholski DR, Hashash YMA (2013) Development of an inverse analysis framework for extracting dynamic soil behavior and pore pressure response from downhole array measurements. Int J Numer Anal Methods Geomech 37(12):1867–1890
Nazir R, Moayedi H, Pratikso A, Mosallanezhad M (2014) The uplift load capacity of an enlarged base pier embedded in dry sand. Arab J Geosci 8:1–12
Moayedi H (2019) Optimization of ANFIS with GA and PSO estimating α in driven shafts. Eng Comput 35:1–12
Chan WT, Chow YK, Liu LF (1995) Neural-network—an alternative to pile driving formulas. Comput Geotech 17(2):135–156
Ismail A, Jeng DS (2011) Modelling load-settlement behaviour of piles using high-order neural network (HON-PILE model). Eng Appl Artif Intell 24(5):813–821
Li YZ, Yao QF, Qin LK (2008) The application of neural network to deep foundation pit retaining structure displacement prediction. World Acad Union-World Acad Press, Liverpool
Chen YH, Wang YW (2012) The analysis on the deformation predition of pile-anchor retaining structure in deep foundation pit in Kunming. In: Zhou XG, Chu MJ, Liu JM, Qu SY, Fan HT (eds) Progress in Structure, Pts 1-4. Trans Tech Publications Ltd., Stafa-Zurich, pp 1222–1225
Tiryaki B (2008) Predicting intact rock strength for mechanical excavation using multivariate statistics, artificial neural networks, and regression trees. Eng Geol 99(1–2):51–60
Cao JW, Huang WH, Zhao T, Wang JZ, Wang RR (2017) An enhance excavation equipments classification algorithm based on acoustic spectrum dynamic feature. Multidimens Syst Signal Process 28(3):921–943
Kwon S, Wilson JW (1998) Investigation of the influence of an excavation on adjacent excavations, using neural networks. J S Afr Inst Min Metall 98(3):147–156
Jan JC, Hung SL, Chi SY, Chern JC (2002) Neural network forecast model in deep excavation. J Comput Civ Eng 16(1):59–65
Chua CG, Goh ATC (2005) Estimating wall deflections in deep excavations using Bayesian neural networks. Tunn Undergr Space Technol 20(4):400–409
Huang FK, Wang GS (2007) ANN-based reliability analysis for deep excavation. IEEE, New York
Chern S, Tsai JH, Chien LK, Huang CY (2009) Predicting lateral wall deflection in top–down excavation by neural network. Int J Offshore Polar Eng 19(2):151–157
Yu J, Chen HM, Yu J, Chen HM (2009) Artificial neural network’s application in intelligent prediction of surface settlement induced by foundation pit excavation. Ieee Computer Soc, Los Alamitos
Huang YT, Siller TJ (1997) Fuzzy representation and reasoning in geotechnical site characterization. Comput Geotech 21(1):65–86
Yilmaz O, Eser M, Berilgen M (2009) Applications of engineering seismology for site characterization. J. Earth Sci 20(3):546–554
Garcia-Fernandez M, Jimenez MJ (2012) Site characterization in the Vega Baja, SE Spain, using ambient-noise H/V analysis. Bull Earthq Eng 10(4):1163–1191
Orhan A, Turkoz M, Tosun H (2013) Preliminary hazard assessment and site characterization of MeAYelik campus area. EskiAYehir-Turk Nat Hazards Earth Syst Sci 13(1):75–84
Kim AR, Cho GC, Kwon TH (2014) Site characterization and geotechnical aspects on geological storage of CO2 in Korea. Geosci J 18(2):167–179
Cao ZJ, Wang Y, Li DQ (2016) Quantification of prior knowledge in geotechnical site characterization. Eng Geol 203:107–116
Wang JP (2016) Site characterization with multiple measurement profiles from different tests: a Bayesian approach. Soils Found 56(4):712–718
Aladejare AE, Wang Y (2017) Sources of uncertainty in site characterization and their impact on geotechnical reliability-based design. ASCE-ASME J Risk Uncertain Eng Syst Part A-Civ Eng 3(4):12
Roy N, Jakka RS (2017) Near-field effects on site characterization using MASW technique. Soil Dyn Earthq Eng 97:289–303
Samui P, Sitharam TG (2010) Site characterization model using least-square support vector machine and relevance vector machine based on corrected SPT data (N-c). Int J Numer Anal Methods Geomech 34(7):755–770
Samui P, Sitharam TG (2010) Site characterization model using artificial neural network and kriging. Int J Geomech 10(5):171–180
Dwivedi VK, Dubey RK, Thockhom S, Pancholi V, Chopra S, Rastogi BK (2017) Assessment of liquefaction potential of soil in Ahmedabad region. West India J Indian Geophys Union 21(2):116–123
Monkul MM, Gultekin C, Gulver M, Akin O, Eseller-Bayat E (2015) Estimation of liquefaction potential from dry and saturated sandy soils under drained constant volume cyclic simple shear loading. Soil Dyn Earthq Eng 75:27–36
Shahri AA, Behzadafshar K, Rajablou R (2013) Verification of a new method for evaluation of liquefaction potential analysis. Arab J Geosci 6(3):881–892
Kayen R, Moss RES, Thompson EM, Seed RB, Cetin KO, Kiureghian AD, Tanaka Y, Tokimatsu K (2013) Shear-wave velocity-based probabilistic and deterministic assessment of seismic soil liquefaction potential. J Geotech Geoenviron Eng 139(3):407–419
Arango I, Lewis MR, Kramer C (2000) Updated liquefaction potential analysis eliminates foundation retrofitting of two critical structures. Soil Dyn Earthq Eng 20(1–4):17–25
Goh A (1994) Seismic liquefaction potential assessed by neural networks. J Geotech Eng 120(9):1467–1480
Seed HB, Tokimatsu K, Harder LF, Chung RM (1985) Influence of SPT procedures in soil liquefaction resistance evaluations. J Geotech Eng-ASCE 111(12):1425–1445
Goh ATC (1994) Nonlinear modelling in geotechnical engineering using neural networks. Trans Inst Eng, Aust Civ Eng 36(4):293–297
Juang CH, Chen CJX, Tien YM (1999) Appraising cone penetration test based liquefaction resistance evaluation methods: artificial neural network approach. Can Geotech J 36(3):443–454
Liu BY, Ye LY, Xiao ML, Miao S (2006) Artificial neural network methodology for soil liquefaction evaluation using CPT values. In: Huang DS, Li K, Irwin GW (eds) Intelligent computing, part I: international conference on intelligent computing, Icic 2006, part I. Springer, Berlin, pp 329–336
Shibata T, Teparaksa W (1988) Evaluation of liquefaction potentials of soils using cone penetration tests. Soils Found 28(2):49–60
Wang J, Rahman MS (1999) A neural network model for liquefaction-induced horizontal ground displacement. Soil Dyn Earthq Eng 18(8):555–568
Young-Su K, Byung-Tak K (2006) Use of artificial neural networks in the prediction of liquefaction resistance of sands. J Geotech Geoenviron Eng 132(11):1502–1504
Hsu SC, Yang MD, Chen MC, Lin JY (2011) Neural network modeling of liquefaction resistance from shear wave velocity. In: Zhou M (ed) 2011 3rd World congress in applied computing, computer science, and computer engineering. Information Engineering Research Inst, Newark, p 155
Zhang WG, Goh ATC, Zhang YM, Chen YM, Xiao Y (2015) Assessment of soil liquefaction based on capacity energy concept and multivariate adaptive regression splines. Eng Geol 188:29–37
Goh ATC, Goh SH (2007) Support vector machines: their use in geotechnical engineering as illustrated using seismic liquefaction data. Comput Geotech 34(5):410–421
Lu P, Rosenbaum MS (2003) Artificial neural networks and Grey Systems for the prediction of slope stability. Nat Hazards 30(3):383–398
Li SJ, Liu YX (2004) Intelligent forecast procedures for slope stability with evolutionary artificial neural network. In: Yin FL, Wang J, Guo CG (eds) Advances in neural networks–Isnn 2004, Pt 2. Springer, Berlin, pp 792–798
Liu ZB, Shao JF, Xu WY, Chen HJ, Zhang Y (2014) An extreme learning machine approach for slope stability evaluation and prediction. Nat Hazards 73(2):787–804
Aghajani HF, Salehzadeh H, Shahnazari H (2015) Application of artificial neural network for calculating anisotropic friction angle of sands and effect on slope stability. J Cent South Univ 22(5):1878–1891
Rahul A, Khandelwal M, Rai R, Shrivastva BK (2015) Evaluation of dump slope stability of a coal mine using artificial neural network. Geomech Geophys Geo-Energy Geo-Resour 1(3–4):69–77
Gordan B, Armaghani DJ, Hajihassani M, Monjezi M (2016) Prediction of seismic slope stability through combination of particle swarm optimization and neural network. Eng Comput 32(1):85–97
Li AJ, Khoo S, Lyamin AV, Wang Y (2016) Rock slope stability analyses using extreme learning neural network and terminal steepest descent algorithm. Autom Constr 65:42–50
Yamagami T, Jiang JC, Ueta Y (1997) Back calculation of strength parameters for landslide control works using neural networks. A a Balkema Publishers, Leiden
Cai DS, Wang GY, Hu TS (1998) A neural network method of landslide prediction of the Geheyan reservoir area of Qingjiang. A a Balkema Publishers, Leiden
Kobayashi T, Furuta H, Hirokane M, Tanaka S, Tatekawa I (1998) Data mining and analysis for landslide risk using neural networks. A a Balkema Publishers, Leiden
Dahigamuwa T, Yu QY, Gunaratne M (2016) Feasibility study of land cover classification based on normalized difference vegetation index for landslide risk assessment. Geosciences 6(4):14
Pradhan B, Lee S (2010) Regional landslide susceptibility analysis using back-propagation neural network model at Cameron highland, Malaysia. Landslides 7(1):13–30
Murillo-Garcia FG, Alcantara-Ayala I (2015) Landslide susceptibility analysis and mapping using statistical multivariate techniques: Pahuatlan, Puebla, Mexico. In: Wu W (ed) Recent advances in modeling landslides and Debris flows. Springer, Berlin, pp 179–194
Souza FT, Ebecken NFF (2004) A data mining approach to landslide prediction. In: Zanasi A, Ebecken NFF, Brebbia CA (eds) Data mining V: data mining, text mining and their business applications. Wit Press, Southampton, pp 423–432
Wu AL, Zeng ZG, Fu CJ (2014) Data mining paradigm based on functional networks with applications in landslide prediction. In: Proceedings of the 2014 international joint conference on neural networks. IEEE, New York, pp 2826–2830
Li Y, Chen G, Tang C, Zhou G, Zheng L (2012) Rainfall and earthquake-induced landslide susceptibility assessment using GIS and artificial neural network. Nat Hazards Earth Syst Sci 12(8):2719–2729
Xu C, Shen LL, Wang GL (2016) Soft computing in assessment of earthquake-triggered landslide susceptibility. Environ Earth Sci 75(9):17
Wang WD, Xie CM, Du XG (2009) Landslides susceptibility mapping based on geographical information system, GuiZhou, south–west China. Environ Geol 58(1):33–43
Ilia I, Koumantakis I, Rozos D, Koukis G, Tsangaratos P (2015) A geographical information system (GIS) based probabilistic certainty factor approach in assessing landslide susceptibility: the case study of Kimi, Euboea, Greece. Springer, Cham
Moher D, Liberati A, Tetzlaff J, Altman DG, Prisma Group (2009) Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 6(7):e1000097
Shamseer L, Moher D, Clarke M, Ghersi D, Liberati A, Petticrew M, Shekelle P, Stewart LA, Prisma PG (2015) Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015: elaboration and explanation. BMJ-Br Med J 349:25
Mardani A, Nilashi M, Zakuan N, Loganathan N, Soheilirad S, Saman MZM, Ibrahim O (2017) A systematic review and meta-analysis of SWARA and WASPAS methods: theory and applications with recent fuzzy developments. Appl Soft Comput 57:265–292
Welch V, Petticrew M, Tugwell P, Moher D, O’Neill J, Waters E, White H (2012) PRISMA-equity 2012 extension: reporting guidelines for systematic reviews with a focus on health equity. Plos Med 9(10):7
Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, Ioannidis JPA, Clarke M, Devereaux PJ, Kleijnen J, Moher D (2009) The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. Plos Med 6(7):28
Hill T, Marquez L, O’Connor M, Remus W (1994) Artificial neural network models for forecasting and decision making. Int J Forecast 10(1):5–15
Shafaei SM, Nourmohamadi-Moghadami A, Kamgar S (2016) Development of artificial intelligence based systems for prediction of hydration characteristics of wheat. Comput Electron Agric 128:34–45
Hornik K (1991) Approximation capabilities of multilayer feedforward networks. Neural Netw 4(2):251–257
Hornik K, Stinchcombe M, White H (1989) Multilayer feedforward networks are universal approximators. Neural Netw 2(5):359–366
Han J, Moraga C, Sinne S (1996) Optimization of feedforward neural networks. Eng Appl Artif Intell 9(2):109–119
Uncuoglu E, Laman M, Saglamer A, Kara HB (2008) Prediction of lateral effective stresses in sand using artificial neural network. Soils Found 48(2):141–153
Lian C, Zeng ZG, Yao W, Tang HM (2013) Displacement prediction model of landslide based on a modified ensemble empirical mode decomposition and extreme learning machine. Nat Hazards 66(2):759–771
Protopapadakis E, Schauer M, Pierri E, Doulamis AD, Stavroulakis GE, Bohrnsen JU, Langer S (2016) A genetically optimized neural classifier applied to numerical pile integrity tests considering concrete piles. Comput Struct 162:68–79
Mustafa MR, Rezaur RB, Rahardjo H, Isa MH (2012) Prediction of pore-water pressure using radial basis function neural network. Eng Geol 135:40–47
Shu SX, Gong WH (2015) Radial basis function neural network-based method for slope stability analysis under two-dimensional random field. Rock Soil Mech 36(4):1205–1210
Kang F, Li JJ, Xu Q (2017) System reliability analysis of slopes using multilayer perceptron and radial basis function networks. Int J Numer Anal Methods Geomech 41(18):1962–1978
Zhang W, Dai BB, Liu Z, Zhou CY (2017) Modeling free-surface seepage flow in complicated fractured rock mass using a coupled RPIM-FEM method. Transp Porous Media 117(3):443–463
Samui P, Kurup P, Dhivya S, Jagan J (2016) Reliability analysis of quick sand condition. Geotech Geol Eng 34(2):579–584
Asadizadeh M, Hossaini MF (2016) Predicting rock mass deformation modulus by artificial intelligence approach based on dilatometer tests. Arab J Geosci 9(2):15
Peng C, Wu W, Zhang BY (2015) Three-dimensional simulations of tensile cracks in geomaterials by coupling meshless and finite element method. Int J Numer Anal Methods Geomech 39(2):135–154
Wang Q, Lin J, Ji J, Fang H (2014) Reliability analysis of geotechnical engineering problems based on an RBF metamodeling technique. Crc Press-Taylor & Francis Group, Boca Raton
Liao KW, Fan JC, Huang CL (2011) An artificial neural network for groutability prediction of permeation grouting with microfine cement grouts. Comput Geotech 38(8):978–986
Liao KW, Huang CL (2011) Estimation of groutability of permeation grouting with microfine cement grouts using RBFNN. In: Liu D, Zhang H, Polycarpou M, Alippi C, He H (eds) Advances in neural networks—Isnn 2011, Pt Iii. Springer, Berlin, p 475
Ibric S, Jovanovic M, Djuric Z, Parojcic J, Solomun L, Lucic B (2007) Generalized regression neural networks in prediction of drug stability. J Pharm Pharmacol 59(5):745–750
Pal M, Deswal S (2008) Modeling pile capacity using support vector machines and generalized regression neural network. J Geotech Geoenviron Eng 134(7):1021–1024
Jiang P, Zeng ZG, Chen JJ, Huang TW (2014) Generalized regression neural networks with K-fold cross-validation for displacement of landslide forecasting. In: Zeng Z, Li Y, King I (eds) Advances in Neural Networks–—Isnn 2014. Springer, Berlin, pp 533–541
Goorani M, Hamidi A (2015) A generalized plasticity constitutive model for sand–gravel mixtures. Int J Civ Eng 13(2B):133–145
Rajesh BG, Choudhury D (2017) Generalized seismic active thrust on a retaining wall with submerged backfill using a modified pseudodynamic method. Int J Geomech 17(3):10
Cigizoglu HK, Alp M (2006) Generalized regression neural network in modelling river sediment yield. Adv Eng Softw 37(2):63–68
Li HZ, Guo S, Li CJ, Sun JQ (2013) A hybrid annual power load forecasting model based on generalized regression neural network with fruit fly optimization algorithm. Knowl-Based Syst 37:378–387
Kumar CS, Arumugam V, Sengottuvelusamy R, Srinivasan S, Dhakal H (2017) Failure strength prediction of glass/epoxy composite laminates from acoustic emission parameters using artificial neural network. Appl Acoust 115:32–41
Vardhan H, Bordoloi S, Garg A, Garg A, Sreedeep S (2017) Compressive strength analysis of soil reinforced with fiber extracted from water hyacinth. Eng Comput 34(2):330–342
Ahangar-Asr A, Javadi AA, Johari A, Chen Y (2014) Lateral load bearing capacity modelling of piles in cohesive soils in undrained conditions: an intelligent evolutionary approach. Appl Soft Comput 24:822–828
Samui P (2012) Determination of ultimate capacity of driven piles in cohesionless soil: a multivariate adaptive regression spline approach. Int J Numer Anal Methods Geomech 36(11):1434–1439
Samui P, Das SK, Sitharam TG (2009) Application of soft computing techniques to expansive soil characterization. In: Gopalakrishnan K, Ceylan H, Okine NOA (eds) Intelligent and soft computing in infrastructure systems engineering: recent advances. Springer, Berlin, pp 305–323
Jang JSR, Sun CT (1995) Neuro-fuzzy modeling and control. Proc IEEE 83(3):378–406
Jang SR (1993) ANFIS: adaptive-network-based fuzzy inference system. IEEE Trans Syst, Man, Cybern 23(3):665–685
Cabalar AF, Cevik A, Gokceoglu C (2012) Some applications of adaptive neuro-fuzzy inference system (ANFIS) in geotechnical engineering. Comput Geotech 40:14–33
Balamurugan G, Ramesh V, Touthang M (2016) Landslide susceptibility zonation mapping using frequency ratio and fuzzy gamma operator models in part of NH-39, Manipur. India Nat Hazards 84(1):465–488
Ramesh V, Anbazhagan S (2015) Landslide susceptibility mapping along Kolli hills Ghat road section (India) using frequency ratio, relative effect and fuzzy logic models. Environ Earth Sci 73(12):8009–8021
Bui DT, Pradhan B, Revhaug I, Nguyen DB, Pham HV, Bui QN (2015) A novel hybrid evidential belief function-based fuzzy logic model in spatial prediction of rainfall-induced shallow landslides in the Lang Son city area (Vietnam). Geomat Nat Hazards Risk 6(3):243–271
Vasu NN, Lee SR, Pradhan AMS, Kim YT, Kang SH, Lee DH (2016) A new approach to temporal modelling for landslide hazard assessment using an extreme rainfall induced-landslide index. Eng Geol 215:36–49
denHartog MH, Babuska R, Deketh HJR, Grima MA, Verhoef PNW, Verbruggen HB (1997) Knowledge-based fuzzy model for performance prediction of a rock-cutting trencher. Int J Approx Reason 16(1):43–66
Ghaboussi J, Sidarta DE (1998) New nested adaptive neural networks (NANN) for constitutive modeling. Comput Geotech 22(1):29–52
Grima MA, Babuska R (1999) Fuzzy model for the prediction of unconfined compressive strength of rock samples. Int J Rock Mech Min Sci 36(3):339–349
Baykasoglu A, Cevik A, Ozbakir L, Kulluk S (2009) Generating prediction rules for liquefaction through data mining. Expert Syst Appl 36(10):12491–12499
Kayadelen C, Taskiran T, Gunaydin O, Fener M (2009) Adaptive neuro-fuzzy modeling for the swelling potential of compacted soils. Environ Earth Sci 59(1):109–115
Sezer A, Goktepe BA, Altun S (2010) Adaptive neuro-fuzzy approach for sand permeability estimation. Environ Eng Manag J 9(2):231–238
Atashpaz-Gargari E, Lucas C (2007) Imperialist competitive algorithm: an algorithm for optimization inspired by imperialistic competition. IEEE, Piscataway
Ahmadi MA, Ebadi M, Shokrollahi A, Majidi SMJ (2013) Evolving artificial neural network and imperialist competitive algorithm for prediction oil flow rate of the reservoir. Appl Soft Comput 13(2):1085–1098
Marto A, Hajihassani M, Armaghani DJ, Mohamad ET, Makhtar AM (2014) A novel approach for blast-induced flyrock prediction based on imperialist competitive algorithm and artificial neural network. Sci World J 2014:1–11
Thangavelautham J, Smith A, El Samid NA, Ho A, Boucher D, Richard J, D’Eleuterio GMT (2008) Multirobot lunar excavation and ISRU using artificial-neural-tissue controllers. In: ElGenk MS (ed) Space technology and applications international forum staif 2008. Amer Inst Physics, Melville, p 229
Manouchehrian A, Gholamnejad J, Sharifzadeh M (2014) Development of a model for analysis of slope stability for circular mode failure using genetic algorithm. Environ Earth Sci 71(3):1267–1277
Lian C, Zeng ZG, Yao W, Tang HM, Chen CLP (2016) Landslide displacement prediction with uncertainty based on neural networks with random hidden weights. IEEE Trans Neural Netw Learn Syst 27(12):2683–2695
Gandomi AH, Kashani AR (2018) Automating pseudo-static analysis of concrete cantilever retaining wall using evolutionary algorithms. Measurement 115:104–124
Ghorbani A, Jokar MRA (2016) A hybrid imperialist competitive-simulated annealing algorithm for a multisource multi-product location-routing-inventory problem. Comput Ind Eng 101:116–127
Al Dossary MA, Nasrabadi H (2016) Well placement optimization using imperialist competitive algorithm. J Pet Sci Eng 147:237–248
Author information
Authors and Affiliations
Ethics declarations
Conflict of interest
The authors declare that there is no conflict of interest in presenting this manuscript.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Corresponding author at Ton Duc Thang University, Ho Chi Minh City, Vietnam.
Rights and permissions
About this article
Cite this article
Moayedi, H., Mosallanezhad, M., Rashid, A.S.A. et al. A systematic review and meta-analysis of artificial neural network application in geotechnical engineering: theory and applications. Neural Comput & Applic 32, 495–518 (2020). https://doi.org/10.1007/s00521-019-04109-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00521-019-04109-9