genetic algorithm

In this study, the stability of the combined PSO-GA algorithm in estimating magnetic model parameters is evaluated and compared with two other algorithms: Particle Swarm Optimization (PSO) and Genetic Algorithm (GA). The PSO algorithm is employed to enhance the action vector, while the GA algorithm is applied to rectify the decision vectors. This amalgamation of algorithms, utilizing genetic operators, enhances exploration and exploitation capabilities. The optimization algorithm exhibits the potential to be applied in the exploration and estimation of mineral reservoirs, offering a rapid method for simulating magnetic anomalies based on ideal geological models. Moreover, in geophysical investigations, it is customary to employ modeling techniques using standard geometric shapes like spheres, cylinders, vertical prisms, dikes, and similar forms to assess magnetic anomaly attributes. Furthermore, in current research, the algorithm's performance is examined by introducing Gaussian white noise to synthetic data, demonstrating its effectiveness even when faced with noise levels as high as 25%. Additionally, authentic airborne magnetic data from the Basiran region in South Khorasan province are applied to validate the model, confirming its consistency with geological findings.

Identification and detection of unexploded ordnance is of particular importance due to life and environmental problems, and due to the presence of many fragments left from the explosion of ordnance, their separation is very important. In this article, in order to separate the unexploded ordnance, spherical geometrical parameters are obtained in two conductive and non-conductive states by using electromagnetic induction data. Electromagnetic responses in the frequency domain and in the induction range, in two modes: the eddy current response caused by the stimulation of the conductive body by the magnetic field and the channel current response due to the disturbance created in the radiating electric field in terms of the sphere response, considering a quantity called the coefficient Polarization is considered and the objective function is determined with the obtained equations. To generate the data, the rotation of the coaxial transmitter and receiver coils, as well as the frequency change of the radiation fields, were used simultaneously to obtain the buried anomaly parameters. In this research, genetic algorithm method with Roulette cycle was used to estimate the parameters. In the genetic algorithm, each randomly generated response is called a chromosome, and a set of chromosomes is called a population. Chromosomes are composed of genes and their values are binary numbers. The value of each chromosome is measured by a function called fit, which measures the usefulness and appropriateness of the solution for the given problem. In this work, the induced voltage was obtained by simultaneously sweeping the frequency (between 18 and 20 kHz with 21 samples) and the angle of the turns with the horizon (between 0 and π in 61 samples) and a real response of 1281 points. In the genetic algorithm, the voltages induced in each response (chromosome) are compared with the voltages obtained in the actual response and the fitting function is defined by calculating the inverse of the mean square error. Chromosome has a better response that has a smaller mean squared error (the inverse of a larger mean squared error). With this process, half of the chromosomes are selected and used in the next generation. In the next generation, the selected chromosomes are combined with each other and produce new chromosomes called children. In one generation, a number of chromosomes undergo mutations in their genes, in which each chromosome has up to 3 It will have random mutations in its genes. The chromosome that is maintained in the population for the next generation is selected by Darwin's law of evolution. A chromosome with a higher fitness value has a higher probability of being selected again in the next generation. After several generations, the chromosome value converges to a specific value that is the best solution for the problem. To validate the method, electromagnetic induction data with 5% noise was used. The obtained results show that the used method has an acceptable ability to estimate the parameters.

Recently, due to the trend of decreasing rainfall and increasing groundwater pumping rate, there have been concerns about the risks caused by the decrease in the volume of aquifer reserves and the drop in the groundwater level, and as a consequence the land subsidence. Also, in Ajabshir plain due to a gradual decrease in the water level, it is necessary to estimate the subsidence and investigate the subsidence potential to prevent its harmful risks in the future. For this purpose, using the ALPRIFT framework, which includes seven layers of parameters affecting subsidence, the subsidence potential map was zoned. The subsidence potential index was obtained in low and moderate ranges. In the next step, using Sentinel-1 satellite images, the subsidence during the years 2020-2021 was estimated to be 2.4 cm, which had a significant correlation with the groundwater level of the water year 2020-2021 and subsidence potential. In addition, artificial intelligence optimization methods including fuzzy logic (Sugeno) and genetic algorithm were used in order to fix the defects of applying expert opinions and increase the correlation between subsidence (Insar) and ALPRIFT, among these models, Sugeno's fuzzy method provided the best correlation between the two subsidence maps and ALPRIFT. The correlation between subsidence with ALPRIFT, ALPRIFT-GA and ALPRIFT-SFL was obtained as 0.46, 0.62 and 0.72 respectively.

Hydrocarbon reservoir modeling is an important tool in characterizing reservoir and forecasting its performance. A realistic reservoir model can be generated by combining different datasets in an optimal manner. Seismic data has always been of interest due to its extensive areal coverage and high lateral resolution compared to well-based data. Using time-lapse (4D) seismic data, the reservoir dynamic modeling is improved. Integrating 4D seismic data into dynamic reservoir modeling process requires that the static reservoir model to be coherent with the pre-production 2D/3D seismic data. To do so, we use seismic matching loop approach, which is based on geostatistical techniques and optimization algorithms to update static reservoir models. A key point in this regard is that seismic data or information is used twice. Acoustic data are first used as a constraint when generating reservoir models based on a geostatistical simulation algorithm. Then, they are compared to the acoustic responses computed for the simulated reservoir models. The implemented optimization algorithm in this study is a hybrid of particle swarm optimization (PSO) and genetic algorithm (GA). The obtained results on a synthetic case study show that the proposed method, compared to the conventional geostatistical methods, results in generation of static reservoir models that are qualitatively and quantitatively more consistent with pre-production seismic data.

Gravity surveying is applied for studying geological structures, for example, basement topography underneath the sediment loads. In potential areas for hydrocarbon and groundwater resources, depth of basement can be estimated using different optimization methods, including stochastic global optimization algorithms. These methods include many functions call of the forward function, so usual forward approaches that discrete the sediment volume into a set of right rectangular prisms need too much computational time. This can be controversial issue while implementing three-dimensional stochastic inversion. In this study, 3D Cauchy-type integral as a fast forward function is applied to accelerate the gravity inversion for 3D determination of the depth to basement. Cai and Zhdanov (2015a,b) introduced this effective approach for potential fields modeling. This method in modeling the sediment-basement interface not only replaces all prisms of conventional volume approach with a gridded basement, but also uses simple mathematical terms in comparison with customary prismatic methods which include trigonometric and logarithmic expressions. Synthetic forward modeling of both of our realistic basin models assesses the validity of the forward operator. Evaluation time for one of the model basins based on the Cauchy-type integral in comparison with the prismatic method which was carried out by two different techniques of forward modeling, is 15 and 50 order lower. Implementing genetic algorithm on the gravity data, the depth of the basement was recovered. The misfit of our data achieved by the algorithm with initial population equal to 10 times of total number of parameters and carrying 700 generations, was lower than 2 mGal. Optimal values were obtained as 80% and 20% for crossover and mutation, respectively. In addition, due to the non-uniqueness of the gravity problem, the genetic algorithm uses a smoothing constraint. By fixing the optimal parameters of genetic algorithm, the optimization process is repeated to find the optimal value for the smoothing factor yielding the most accurate model based on the RMS of the reconstructed model. Results show that a smoothing factor between 0.005-0.015, reconstructs stable solutions. Besides, applying a Gaussian filter, a smoothing filter with the kernel size equal to 11×11 to the calculated depths, achieves more stable evaluations. Noisy synthetic and noise-free gravity data were inverted for one symmetric basin and the algorithm has been able to successfully reconstruct the basement. The case study area is the Aman-Abad alluvial plain (Iran) which its main parts are located in the Sanandaj-Sirjan zone in the Zagros Mountains of Iran. The suitable parameters of the genetic algorithm are found by synthetic tests to invert real gravity data to image the interface of the impermeable layer groundwater. The most common polynomial regression, i.e., degree 1 is applied to calculate residual gravity anomaly. Reconstructed depths from residual gravity anomaly match properly with gravity anomaly trend. Deep parts of the basement (as impermeable surface) have been estimated about 150m which it looks promising for groundwater resources. According to the previous gravity studies, the calculated maximum thickness of sediment is lower than 200 m and the well data specified depth of the basement is 140 m.

In a geophysical inversion process, the observed data is transformed to the meaningful properties of earth. For inversion of petrophysical properties, we need a rock physics model that links the petrophysical properties to the elastic properties of rock. The more accurate the model, the more reliable the results. There are a variety of procedures in which can invert petrophysical properties of earth through seismic data. Those procedures include experimental and empirical methods (In these methods, seismic data is assumed to be a function of some special petrophysical features of a zone), statistical methods and theoretical methods such as Biot model. Such theoretical method predicts elastic properties of rocks such as velocities and quality factors as functions of physical properties of rock and fluid. Here, we use BISQ  (Biot-Squirt flow) model for inversion of petrophysical properties of reservoir rock. The BISQ describes seismic wave propagation in a fluid saturated poroelastic medium. This model consists of both Biot and squirt flow models and its accuracy is confirmed by several researchers versus other models. The model is developed for several anisotropic media too. Biot model relates the attenuation of seismic wave to parallel motion of fluid in a solid frame; while, the squirt flow model relates it to local motion of fluid. It is proven that both mechanisms exist during seismic wave propagation and BISQ model is correct for both of them, simultaneously. In a petrophysical properties inversion process, there are two vital elements. First one, as described before, is using a rock physical model and the other one is mathematical method by which we solve an optimization problem that minimizes misfit between observed and predicted data. Here, we choose PNGA (Parallel Niche Genetic Algorithm) because of nonlinearity of BISQ model. Moreover, PNGA as an evolutionary algorithm, has capability of dealing with multi-objective optimization problems. We apply the mentioned method on both synthetic and real data. The inversion results show acceptable correlation with the used quantities in generation of synthetic and well logs.

Nderground land subsidence due to extraction of groundwater resources is one of the events that have been observed in abundance in Iran. If it is not properly managed, this phenomenon can cause irreparable damage to the areas affected by this phenomenon. Population growth and crop technology have also led to more use of groundwater resources in some parts of Iran.
The rate of land subsidence due to excessive groundwater abstraction in Tehran plain has increased over recent years. Under such conditions, its harmful effects on agricultural land, buildings, vital arteries, and deformations of the earth's surface develop in the form of hollows, cracks and fissure in the earth. In this research, using the ALPRIFT model, which is a simple weighing model, consists of seven layers that are effective on subsidence (aquifer media, land use, pumping of groundwater, recharge, impact of aquifer thickness, fault distance, groundwater level decline). The level of vulnerability of plain subsidence was obtained by combining layers in the Arc GIS for a period of one year (2015-2016) and the InSAR satellite radar map was used to validate this model. Despite the acceptability of the results, the genetic algorithm was used to improve the results and optimize the weights. The results showed that the optimization of the weights applied by the genetic algorithm, with increasing correlation coefficient from 0.44 to 0.75, between the subsidence index and the subsidence in the plain, is more capable of evaluating subsidence

In order to implement sustainable groundwater resources management, it is necessary to model the behavior of groundwater level. Groundwater is a nonlinear and complex system which Data-driven models can be modeled this system without approximation and simplification. This study evaluates the performances of data-driven models (Support Vector Regression (SVR) and Artificial Neural Network (ANN)) for forecasting groundwater levels of Hashtgerd plain at 1, 2 and 3 months ahead. One of the most important steps in constructing data-driven models is determining the optimal combination of input variables. In this paper, the optimal composition was determined using the hybrid of gamma test and genetic algorithm, also the optimal length of input composition in teaching period was determined using the M-test method. The results of this study showed that the models for 1 month ahead performed better than for 2 and 3 months ahead forecasts and the accuracy of the models decreased as months ahead increased. The Developed Discrepancy Ratio (DDR) was applied to evaluate the efficiency of models. The results showed that the ANFIS model had the best performance for one month ahead groundwater level forecasting. Also, the results demonstrated that the ANFIS and MLP models were able to provide better performance in groundwater levels forecasting for a shorter-time period (e.g., 1 month ahead). The performance of the SVR model for a longer-time period (e.g., 3 month ahead) was superior to the MLP and ANFIS models.

Permeability is a property of the reservoir rock, which deals with the flow of fluid from the reservoir and is an important factor in oil and gas production. This parameter is measured via coring and core laboratory analysis, which is an expensive and time-consuming process and also is not a feasible approach for every oil and gas field. Nowadays, the permeability can also be calculated using the data of petrophysical logs by means of statistical and intelligent techniques. The present study uses four wells drilled in Kangan and Dalan formations within South Pars gas field to predict permeability using fuzzy logic. Out of totally eight features extracted from each well, four more effective features were selected using correlation-based feature selection tools. Then, regression, multi-layer perceptron, RBF neural network, Local Linear Model Trees (LOLIMOT), type-1 and type -2 fuzzy systems were utilized for permeability prediction. The results indicated that due to the uncertainty in the petrophysical and permeability parameters, type-2 Fuzzy systems cover better the uncertainties. The aforementioned method predicts the best number of rules using the GSA-GA (Gravitational Search Algorithm - Genetic algorithm) combined algorithms. Fuzzy membership functions were also improved using the K-means clustering algorithms. These improvements led to increased accuracy of the predicted permeability with a coefficient of 0.9768, and a decrease in the root mean square error to 0.1602.

Imperialist Competitive Algorithm (ICA) is a new socio-politically motivated global search strategy that has recently been introduced for dealing with different optimization tasks. The estimation of sodium adsorption ratio (SAR) and chloride content as example of groundwater quality parameters are much more time-consuming and expensive than water salinity measurement.For using this method, a numerical code was designed, and ICA algorithm calculated weight coefficients of the proposed function for the input data. The validation of the simulation with the ICA model showed that MSE in testing sample for SAR and chloride were 0.0134 and 0.0098, respectively. Also, R2 of validity for SAR and chloride were 0.93 and 0.952, respectively. This study was done based on the sample collected from the groundwater of Bostanabad plain, and could be generalized to other areas. The results of this study indicated high speed, convergence and power of the proposed method compared to genetic algorithm (GA) in estimation of sodium adsorption ratio (SAR) and chloride.

It is essential to dtermine petrophysical properties of reservoir rock in exploration and poroduction processes.There are several methods in which predicts such properties and these methode are evolving every day. BISQ model is a novel model that it’s accuracy in determination of seismic properties of reservoir rock is confirmed by different researchers. This model predicts seismic properties such as p-wave and s-wave velocity and quality factor using some petrophysical and elastic properties of reservoir rock. Because of a high order of non-linearity in the BISQ model, inversion of associated petrophysical properties using conventional linear methods is not feasible. So in this paper we used a global optimization method to detemine petrophysical properties of reservoir rock. Inversion process includes two major step. First one is a physical or petrophysical model in which predicts data. And second one is a mathmatical method in which estimate model parameter through an objective function. Selection of mathmatical method depend on physical or petrophysical modle that pridicts data. In this case the petrophysical model is non-linear so we should select a robust mathmatical method that has capibility of determination global optimum point in a highly non-linear space. We use a hybrid genetic algorithm(HGA) in which employs some self adaptive concepts to enhance search ability , diversity , and quality of answers in comparsion with conventional GAs. Here we first introduce inverse methode then the methode is applied on a synthtic and a real data and the results are analyzed.

Archie’s equation, which is the most fundamental equation for water saturation calculation,consists of three factors: Cementation factor (m), saturation exponent (n) and tortuosity (a). Cementation factor is a function of the shape of pores. Hence, the study of pore type is important in determining the Archie’s coefficients. In order to achieve more precise and reliable results in Archie’s coefficient determination and then water saturation accurately, the rocks must be rated based on texture and porosity type, the coefficients should be constrained for each class. In this paper, fractal method is used to rate the resistivity log data and calculate Archie’s coefficient in an exploration well of a hydrocarbon reservoir in southwest of Iran. The results show three different zones based on porosity type and texture of the rocks. Then the Genetic algorithm method is used to calculate the Archie’s coefficients in each of the zones separately. The results show that this method is able to consider the complex behavior of each of the coefficients in the calculations.

An earthquake with the magnitude of 5.9 ML shocked the Qeshm island located in the Persian Gulf on 27 November, 2005 at 13:53:22 local time. The earthquake occurred due to the reactivation of a NE-SW fault with a major reverse mechanism accompanied by a minor strike-slip component. Another earthquake of 5.5 MW occurred on the same day at ca. 20:00 local time. The interesting feature of this earthquake is that the calculated mechanism for its strongest aftershock, which occurred ca. 6 hours after the main-shock, was a strike-slip mechanism that is completely different the pure reverse mechanism for the main-shock. This study uses inversion of InSAR observation of earth surface displacement field boundary values to solve parameters of these 2 earthquakes. The results show activation of southern part of the Qeshm fault caused by the first earthquake along 7 km of its length. This event induced the second earthquake by activation of another strike-slip fault which is parallel to the Gavarzin anticline. Estimated slip was 96 cm for the first earthquake and 9 cm for the second one. Based on the estimated parameters of the these 2 earthquakes, the maximum displacement induced by the first earthquake was 6.7 cm in west, 4.6 cm in south and 16.4 cm in vertical directions on the earth surface. The maximum displacement of the second earthquake in west, south and vertical directions were 1.3, 1.6 and 1.4 cm respectively.

Summary: Linear and nonlinear time-history analyses are necessary tools for many fields of structural/earthquake engineering including vulnerability analyses and response evaluations. In this regard، one of the most important requirements is the selection of proper acceleration time histories as the analysis input. These records can be obta from natural earthquake records or can be generated synthetically and artificially. The first option is usually preferred to the others since it can provide true information about strong ground shaking characteristics to reflect the source، path، and site effects. These characteristics include several parameters such as earthquake magnitude، epicenter distance، duration، Arias intensity، spectral intensity، soil type and peak ground responses. However، well-known seismic design codes offer their linear design spectra to which the earthquake accelerogram is scaled; instead of direct matching the seismologicalparameters. The number of ground motions recorded at a given site is not generally sufficient; thus، it is required to select other earthquakes from real-world catalogues based on their similarity to the site-specific record (s). However، it is a challenging task to deserve such a similarity due to the existence of several criteria affecting the earthquake characteristics. The present work deals with the problem of selecting a number of appropriate ground motions among an available catalogue in order to best match a given site-specific earthquake record. As an unsupervised solution in this research، clustering techniques were concerned. A set of data in a multi-dimensional space were classified into some clusters in such a way that the objects in every cluster have the most similarity with each other and the least similarity with any object in the other clusters. Here، both deterministic and non-deterministic approaches in such a clustering problem were employed and compared. The method of K-means was selected as a deterministic algorithm due to its popularity and computational efficiency among clustering techniques. A set of input clusters was required to be introduced for this algorithm before its run. However، this method itself was deterministic. Such an initial set is usually chosen by trial-error or randomly. Depending on the initial input clusters، the K-means algorithm reveals different local optima instead of global optimum as the clustering result. In order to overcome such a challenge، a non-deterministic approach was studied as well. In this regard، a Genetic Algorithm، GA، was utilized as a well-trusted metaheuristic among non-deterministic algorithms with the capability of overpassing localresults and approaching to the true global optimum. Proper genetic encoding and operators were developed for the current optimal clustering problem. Consequently، a variety of earthquake characteristics were taken into account in the utilized chromosome including magnitude، epicenteral distance، duration، Arias intensity، spectral intensity، soil type، peak ground responses and a specialized error measure. This measure was defined based on how compatible was the response spectrum of the record with the codified design spectrum as the target. In the present study، average silhouette widths and related profile plot were used to evaluate effectiveness of the proposed clustering processes. Quality of clustering was thus measured taking into account both compactness and separation of the clusters. Consequently، the clustering result corresponding to the maximum amount of the average silhouette bandwidth was announced as the solution. Input-records required for the timehistory analyses، were then selected from the cluster which includeed the target siteassociated earthquake. Using different numbers of clusters، the proposed genetic search stood well superior to the method of K-means in achieving the best clustering for the present problem.

Several functions have been proposed in order to estimate the rate of erosion of catchment areas. Since the rate of sediment production and transport are governed by many catchment characteristics such as topography, soil erodibility, land use and etc., direct estimation of these parameters is very difficult, expensive and time consuming. Therefore, the best way to calculate these parameters is using parametric models and optimization techniques. These models use optimization methods to estimate production and transport parameters using observed values of river discharge and sediment load in a given period of time. The results of erosion parameter estimation could be used to predict the volume of the sediment load to design the hydraulic structures such as dam reservoirs. In this research, Samani model is performed on the catchment of Kor and Sivand Rivers, and the production and transport parameters of these catchments are estimated using Genetic Algorithm technique. The model was evaluated for 1 to 3-reservoir cases. The model is optimized in two-reservoir for Kor River and 3-reservoir model for Sivand River.

In order to obtain more accurate results from application of the method of artificial neural networks, instead of selection of the best network determined by trial and error process, we suitably combine the results of several networks that is called committee machine, to reduce the error, and thus, increasing the accuracy of the output results. In this research, ensemble combination of single artificial neural networks has been used in order to estimate the effective porosity of Kangan gas reservoir rock in South Pars hydrocarbon field. To achieve this goal, well logging data of 4 wells in the area at the depth interval corresponding to Kangan formation were used. Acoustic, density, gamma ray and neutron porosity well log data were assigned as the input of the networks while the effective porosity data were considered as the output of the networks. Back- propagation single neural networks having different structures were trained using regularization method and their results were assessed. Then, the networks with the best results, i.e. contained minimum mean of squares of errors in the test step, were selected for making ensemble combinations. To determine the weighting coefficients of the networks in the linear ensemble combinations, we applied three methods of simple averaging, Hashem’s optimal linear combination and non-analytical optimal linear combination employing genetic algorithm, and their results were compared. The best ensemble combination, in which we had the maximum reduction in mean of squares of errors of the test step compared to the best single neural network, was an optimal linear four-network combination obtained by using genetic algorithm optimization method. This best ensemble combination, compared to the best single neural network, reduced the mean of squares of errors in the training and test steps 3.6% and 11.2%, respectively.

Ground vibration is an inevitable consequence of blasting operations in open pit miningand civil engineering projects. A large amount of energy is released in every blastingexercise in the form of wave propagation, which can cause serious damage to thesurrounding environment. Safe design of blasting for underground structures andtunneling are nowadays carried out mainly based on the permissible peak particle velocity(PPV), which has become a major criterion in setting out the appropriate safety standardsenvisaged to prevent structural damage near such operations. Many published studieshave produced correlations for PPV prediction using experimental data The GotvandOlya Dam is one of the major civil engineering projects in the last thirty years in Iran, constructed as the last dam on the Karoon River with the aim of increasingenvironmentally friendly electrical energy of the country, seasonal flood control in theprovince, and a supply of agricultural water for the strategic farming grounds ofKhoozestan. The geological formations of Gotvand include the two main formations ofBakhtiyari and Aghajari. The existence of discontinuities in these formations as well asjoint sets and their low inherent strength caused a number of challenges in blasting designoperations. This dam is 30 km north west of Shoshtar in the Khoozestan province, and 12km away from Gotvand. In this study, 16 records of 3 components produced from 4 blastswere obtained using 4 seismographs of the type PG-2002 at Intac and surge tanks of theGotvand Olya Dam in order to investigate their effects on underground structures. Thethree-component seismometers were of the GS-11D type and the records were analyzedusing DADISP software. The explosion materials used were ANFO and dynamite, andthe maximum weight of the explosions per delay in the Intac and surge tank were in therange of 32.0 to 343.3 kg. The positions of the seismometers were in the range of 56.2 to 145.5 meters away from the centre of blast blocks. The permissible peak particle velocity was taken from the Korean Institute of Geology, Mining and Materials. Using the scale distance, an exponential equation based on the cubic root of the charge weight per delay and coefficient correlation of 85% was proposed for predicting the PPV. In thisrelationship, the geological coefficients of the Gotvand area were estimated to be 20.339and -3.08. Using a genetic algorithm, an improved coefficient correlation between theexperimental and predicted PPV of 89% was obtained. The results from the geneticalgorithm applied for the coefficients of a, b, and n in the equation of PPV=b(d/wn)a were -9.286849, 9.769703 and 0.745959, respectively. Considering this proposed relationship,the minimum distance of 56.2 m to the centre of the blast block and the permissible peak particle velocity of 254 mm/s for a 10 day old concrete, the maximum charge weight per delay was found to be 254.4 kg.