Showing papers by "Amir H. Alavi published in 2011"

A robust data mining approach for formulation of geotechnical engineering systems

Abstract: Purpose – The complexity of analysis of geotechnical behavior is due to multivariable dependencies of soil and rock responses. In order to cope with this complex behavior, traditional forms of engineering design solutions are reasonably simplified. Incorporating simplifying assumptions into the development of the traditional models may lead to very large errors. The purpose of this paper is to illustrate capabilities of promising variants of genetic programming (GP), namely linear genetic programming (LGP), gene expression programming (GEP), and multi‐expression programming (MEP) by applying them to the formulation of several complex geotechnical engineering problems.Design/methodology/approach – LGP, GEP, and MEP are new variants of GP that make a clear distinction between the genotype and the phenotype of an individual. Compared with the traditional GP, the LGP, GEP, and MEP techniques are more compatible with computer architectures. This results in a significant speedup in their execution. These method...

Nonlinear Genetic-Based Models for Prediction of Flow Number of Asphalt Mixtures

Abstract: Rutting has been considered the most serious distress in flexible pavements for many years. Flow number is an explanatory index for the evaluation of the rutting potential of asphalt mixtures. In this study, a promising variant of genetic programming, namely, gene expression programming (GEP), is utilized to predict the flow number of dense asphalt-aggregate mixtures. The proposed constitutive models relate the flow number of Marshall specimens to the coarse and fine aggregate contents, percentage of air voids, percentage of voids in mineral aggregate, Marshall stability, and Marshall flow. Different correlations were developed using different combinations of the influencing parameters. The comprehensive experimental database used for the development of the correlations was established on the basis of a series of uniaxial dynamic-creep tests conducted in this study. Relative importance values of various predictor variables were calculated to determine their contributions to the flow number prediction. A multiple-least-squares-regression (MLSR) analysis was performed to benchmark the GEP models. For more verification, a subsequent parametric study was carried out, and the trends of the results were confirmed with the results of previous studies. The results indicate that the proposed correlations are effectively capable of evaluating the flow number of asphalt mixtures. The GEP-based formulas are simple, straightforward, and particularly valuable for providing an analysis tool accessible to practicing engineers.

Permanent deformation analysis of asphalt mixtures using soft computing techniques

Abstract: Research highlights? New prediction models derived by means of multi expression programming (MEP) and multilayer perceptron (MLP) of artificial neural networks give reliable estimates of the flow number of dense asphalt-aggregate mixtures. The MEP-based straightforward formulas are much more practical for the engineering applications compared with the complicated equations provided by MLP. ? The proposed models correlate the flow number of Marshall specimens with the coarse and fine aggregate contents, percentage of bitumen, percentage of voids in mineral aggregate, Marshall stability, and Marshall flow. ? Sensitivity and parametric analyses were performed to verify the validity of the derived models. The obtained results were confirmed with the experimental study results and those of previous studies. ? The proposed MEP and MLP-based models perform superior than the developed regression models. ? The derived design equations can reliably be used as a quick check on solutions developed by more time consuming and in-depth deterministic analyses. This study presents two branches of soft computing techniques, namely multi expression programming (MEP) and multilayer perceptron (MLP) of artificial neural networks for the evaluation of rutting potential of dense asphalt-aggregate mixtures. Constitutive MEP and MLP-based relationships were obtained correlating the flow number of Marshall specimens to the coarse and fine aggregate contents, percentage of bitumen, percentage of voids in mineral aggregate, Marshall stability, and Marshall flow. Different correlations were developed using different combinations of the influencing parameters. The comprehensive experimental database used for the development of the correlations was established upon a series of uniaxial dynamic creep tests conducted in this study. Relative importance values of various predictor variables of the models were calculated to determine the significance of each of the variables to the flow number. A multiple least squares regression (MLSR) analysis was performed to benchmark the MEP and MLP models. For more verification, a subsequent parametric study was also carried out and the trends of the results were confirmed with the experimental study results and those of previous studies. The observed agreement between the predicted and measured flow number values validates the efficiency of the proposed correlations for the assessment of the rutting potential of asphalt mixtures. The MEP-based straightforward formulas are much more practical for the engineering applications compared with the complicated equations provided by MLP.

Nonlinear modeling of shear strength of SFRC beams using linear genetic programming

Abstract: A new nonlinear model was developed to evaluate the shear resistance of steel fiberreinforced concrete beams (SFRCB) using linear genetic programming (LGP). The proposed model relates the shear strength to the geometrical and mechanical properties of SFRCB. The best model was selected after developing and controlling several models with different combinations of the influencing parameters. The models were developed using a comprehensive database containing 213 test results of SFRC beams without stirrups obtained through an extensive literature review. The database includes experimental results for normal and high-strength concrete beams. To verify the applicability of the proposed model, it was employed to estimate the shear strength of a part of test results that were not included in the modeling process. The external validation of the model was further verified using several statistical criteria recommended by researchers. The contributions of the parameters affecting the shear strength were evaluated through a sensitivity analysis. The results indicate that the LGP model gives precise estimates of the shear strength of SFRCB. The prediction performance of the model is significantly better than several solutions found in the literature. The LGP-based design equation is remarkably straightforward and useful for pre-design applications.

Genetic-based modeling of uplift capacity of suction caissons

Abstract: In this study, classical tree-based genetic programming (TGP) and its recent variants, namely linear genetic programming (LGP) and gene expression programming (GEP) are utilized to develop new prediction equations for the uplift capacity of suction caissons. The uplift capacity is formulated in terms of several inflecting variables. An experimental database obtained from the literature is employed to develop the models. Further, a conventional statistical analysis is performed to benchmark the proposed models. Sensitivity and parametric analyses are conducted to verify the results. TGP, LGP and GEP are found to be effective methods for evaluating the horizontal, vertical and inclined uplift capacity of suction caissons. The TGP, LGP and GEP models reach a prediction performance better than or comparable with the models found in the literature.

Nonlinear Neural-Based Modeling of Soil Cohesion Intercept

Abstract: A new model was derived to estimate undrained cohesion intercept (c) of soil using Multilayer Perceptron (MLP) of artificial neural networks. The proposed model relates c to the basic soil physical properties including coarse and fine-grained contents, grains size characteristics, liquid limit, moisture content, and soil dry density. The experimental database used for developing the model was established upon a series of unconsolidated-undrained triaxial tests conducted in this study. A Nonlinear Least Squares Regression (NLSR) analysis was performed to benchmark the proposed model. The contributions of the parameters affecting c were evaluated through a sensitivity analysis. The results indicate that the developed model is effectively capable of estimating the c values for a number of soil samples. The MLP model provides a significantly better prediction performance than the regression model.

New Ground-Motion Prediction Equations Using Multi Expression Programing

Abstract: High-precision attenuation models were derived to estimate peak ground acceleration (PGA), velocity (PGV), and displacement (PGD) using a new variant of genetic programming, namely multi expression programming (MEP). The models were established based on an extensive database of ground-motion recordings released by Pacific Earthquake Engineering Research Center (PEER). For more validity verification, the models were employed to predict the ground-motion parameters of the Iranian plateau earthquakes. The results indicate that the MEP attenuation models are capable of effectively estimating the peak ground-motion parameters. The proposed models are able to reach a prediction performance comparable with the attenuation relationships found in the literature.

Applications of Computational Intelligence in Behavior Simulation of Concrete Materials

Abstract: The application of Computational Intelligence (CI) to structural engineering design problems is relatively new. This chapter presents the use of the CI techniques, and specifically Genetic Programming (GP) and Artificial Neural Network (ANN) techniques, in behavior modeling of concrete materials. We first introduce two main branches of GP, namely Tree-based Genetic Programming (TGP) and Linear Genetic Programming (LGP), and two variants of ANNs, called Multi Layer Perceptron (MLP) and Radial Basis Function (RBF). The simulation capabilities of these techniques are further demonstrated by applying them to two conventional concrete material cases. The first case is simulation of concrete compressive strength using mix properties and the second problem is prediction of elastic modulus of concrete using its compressive strength.

Formulation of uplift capacity of suction caissons using multi expression programming

Abstract: Suction caissons have increasingly been used as foundations and anchors for deepwater offshore structures in the last decade. The increased use of suction caissons defines a serious need to develop more authentic methods for simulating their behavior. Reliable assessment of uplift capacity of caissons in cohesive soils is a critical issue facing design engineers. This paper proposes a new approach for the formulation of the uplift capacity of suction caissons using a promising variant of Genetic Programming (GP), namely Multi Expression Programming (MEP). The proposed model is developed based on experimental results obtained from the literature. The derived MEP-based formula takes into account the effect of aspect ratio of caisson, shear strength of clayey soil, point of application and angle of inclination of loading, soil permeability and loading rate. A subsequent parametric analysis is carried out and the trends of the results are confirmed via previous studies. The results indicate that the MEP formulation can predict the uplift capacity of suction caissons with an acceptable level of accuracy. The proposed formula provides a prediction performance better than or comparable with the models found in the literature. The MEP-based simplified formulation is particularly valuable for providing an analysis tool accessible to practicing engineers.

Nonlinear genetic-based simulation of soil shear strength parameters

Abstract: New nonlinear solutions were developed to estimate the soil shear strength parameters utilizing linear genetic programming (LGP). The soil cohesion intercept (c) and angle of shearing resistance (φ) were formulated in terms of the basic soil physical properties. The best models were selected after developing and controlling several models with different combinations of influencing parameters. Comprehensive experimental database used for developing the models was established upon a series of unconsolidated, undrained, and unsaturated triaxial tests conducted in this study. Further, sensitivity and parametric analyses were carried out. c and φ were found to be mostly influenced by the soil unit weight and liquid limit. In order to benchmark the proposed models, a multiple least squares regression (MLSR) analysis was performed. The validity of the models was proved on portions of laboratory results that were not included in the modelling process. The developed models are able to effectively learn the complex relationship between the soil strength parameters and their contributing factors. The LGP models provide a significantly better prediction performance than the regression models.

Towards the prediction of business failure via computational intelligence techniques

Abstract: This is a pioneer study that presents two branches of computational intelligence techniques, namely linear genetic programming (LGP) and radial basis function (RBF) neural network to build models for bankruptcy prediction. The main goal is to classify samples of 140 bankrupt and non-bankrupt Iranian corporations by means LGP and RBF. Another important contribution of this paper is to identify the effective predictive financial ratios based on an extensive bankruptcy prediction literature review and a sequential feature selection analysis. In order to benchmark the proposed models, a log–log regression analysis is further performed. A comparative study on the classification accuracy of the LGP, RBF and regression-based models is conducted. The results indicate that the proposed models effectively let estimate any enterprise in the aspect of bankruptcy. The LGP models have a significantly better prediction performance in comparison with the RBF and regression models.