Adaptive Genetic Algorithm Research Articles

PROVIDE A MODEL BASED SENTIMENT ANALYSIS SYSTEM FOR SALES PREDICTION IN MARKETING ACCORDING TO THE AGA-LSTM NEURAL NETWORK MODEL

Data is today's most powerful tool; valuable facts and information can be determined by analyzing them using appropriate techniques and algorithms. Also, the rapid increase in access to Internet technology to a large mass of people worldwide has increased the importance of analyzing data generated on the web much more than before. The preceding discussion of this research is sales forecasting in marketing, which is very important in this topic. Marketing is a tool through which people's standard of living is developed, which is done before and after the sale. This research presents a model based on a dynamic analysis system for forecasting marketing sales based on the AGA-LSTM neural network model. It is challenging to recognize emotions in natural language, even for humans, and automatic recognition makes it more complicated. This research presents a hybrid deep-learning model for accurate sentiment prediction in real-time multimodal data. In the proposed method, the work process is such that after extracting emotional data from social networks, they are pre-processed and prepared for pattern discovery. The data is evaluated in the adaptive genetic algorithm, and the pattern is discovered in the designed neural network, and this pattern is discovered after discovery. The cornerstone of sales policies is improved. The adaptive genetic algorithm was used to optimize the parameters of the LSTM model, and the model can predict the types of goods and the total volume of online retail sales. In the simulation of the proposed method, in 3000 rounds of training, an accuracy of 76 has been achieved, which is an improvement of 11% compared to the primary method.

Using GA-ANFIS machine learning model for forecasting the load bearing capacity of driven piles

This paper is aimed to apply hybrid machine learning model namely GA-ANFIS, which is a combination of Adaptive Neuro-Fuzzy Inference System (ANFIS) and Genetic Algorithm (GA), for the prediction of total bearing capability of driven piles. A database of 95 Pile Driving Analyzer (PDA) tests carried out at the win power project in Hoa Binh province, Vietnam was used to develop hybrid model. The database was split into 70:30 ratio for training (70%) and validating (30%) model. Accuracy of the model was evaluated using statistical standard indicators: Coefficient of determination (R2), Mean Absolute Error (MAE), and Root mean squared error (RMSE). Results indicated that the GA-ANFIS model has a good performance in correct prediction of the total bearding capability of driven piles on both training (R2 = 0.976) and testing (R2 =0.925) datasets. Therefore, the GA-ANFIS hybrid model is a promising tool for quick and accurate prediction of the total bearing capability of driven piles for the consideration in design and construction of the structures.

Optimization of China's freight transportation structure based on adaptive genetic algorithm under the background of carbon peak.

Transportation is the key carbon emission source after energy supply and industrial production. Under the vision of carbon peak and carbon neutralization, the pressure of reducing carbon emissions in transportation will be greater in the future. This paper constructs a model that takes transportation carbon emission as the main target and freight transportation utility value as the auxiliary target. The constructed model satisfies the constraints of freight turnover in the whole society, freight economic and social benefits, and the ecological constraints of the freight system. With MATLAB, the freight turnover of roadways, railways, and waterways (excluding ocean transportation) in 2030 is solved by using the adaptive genetic algorithm. The results indicate that (I) compared with the current freight structure of China, the roadway freight sharing rate in 2030 will decrease by 8.07%, and the railway freight sharing rate and the waterway freight sharing rate (excluding ocean transportation) will increase by 0.93% and 7.13%, respectively. (II) After optimization, the energy consumption and carbon emission are reduced by 42,471,500 tons (10.3%) and 91,379,400 tons (10.2%) of standard coal, respectively. (III) The adaptive genetic algorithm outperforms the traditional genetic algorithm in terms of convergence speed and accuracy. (IV) As the weight coefficient of carbon emission increases, the utility value of freight transportation consistently decreases, and the sensitivity increases. Meanwhile, as the carbon emission weight coefficient increases, carbon emission keeps decreasing, and the sensitivity decreases.

Two combination methods of piecewise unsaturated tri-stable stochastic resonance system and bearing fault detection under different noise

In this paper, a piecewise unsaturated tri-stable stochastic resonance (PUTSR) system is proposed to address the issue of output saturation. Firstly, the unsaturated characteristics of the PUTSR system are verified, followed by a discussion on the system performance of PUTSR under two different combinations of coupled and cascaded systems. The two-dimensional coupled piecewise unsaturated tri-stable stochastic resonance (TCPUTSR) system and cascaded piecewise unsaturated tri-stable stochastic resonance (CPUTSR) system are proposed. By implementing the adiabatic approximation theory, the steady-state probability density (SPD) of TCPUTSR system and spectral amplification (SA) coefficient of TCPUTSR and CPUTSR systems are derived respectively. The impact of different system parameters on the SPD and SA are also investigated. Finally, PUTSR, TCPUTSR and CPUTSR systems are applied to detection on the fault signal of bearing. Gaussian white noise and color noise are used to simulate actual noise environments in the experiments, while the adaptive genetic algorithm (AGA) is used to optimize the parameters and measure the system performance via signal-to-noise ratio improvement (SNRI). Results indicate that both coupled and cascaded systems demonstrate better detection capabilities than individual systems, with the CPUTSR system offering better enhancement and detection of bearing fault signals. Under Gaussian white noise, the SNRI of the CPUTSR system is 1.449 dB ~ 3.198 dB higher than that of the PUTSR and TCPUTSR systems, while under color noise it is 1.566 dB ~ 2.849 dB higher. This paper mainly compares the spectral amplification capability and fault signal detection capability of the same system under two different combinations of coupled and cascaded. The theoretical and numerical simulations further verify that the designed cascaded system outperforms the coupled system and proves the great superiority of CPUTSR system. These findings provide crucial theoretical support and practical application prospects for engineering purposes.

A Novel Rolling Bearing Fault Diagnosis Method Based on MFO-Optimized VMD and DE-OSELM

Rolling bearings are critical in maintaining smooth operation of rotating machinery and considerably influence its reliability. The signals collected from rolling bearings in field conditions are often subjected to noise, creating a challenge to extract weaker fault features. This paper proposes a rolling bearing fault diagnosis method that addresses the above-mentioned problem through the moth-flame optimization algorithm optimized variational mode decomposition (MFO-optimized VMD) and an ensemble differential evolution online sequential extreme learning machine (DE-OSELM). By using the dynamic adaptive weight factor and genetic algorithm cross operator, the optimization accuracy and global optimization ability of the moth-flame optimization (MFO) are improved, and the two basic parameters of VMD decomposition level and quadratic penalty factor are adaptive selected. Since the vibration characteristics of the signal cannot be fully interpreted by a single index, The effective weighted correlation sparsity index (EWCS) is utilized to extract the relevant intrinsic mode functions (IMF) of VMD decomposition and extract their energies as features. In order to improve the classification accuracy, The energy feature set is subsequently inputted into DE-OSELM for training and classification purposes, and the proposed method is assessed via a sample set with four different health states of actual rolling bearings. Our proposed method results are compared with other diagnosis methods, proving its feasibility to diagnose rolling bearing faults with higher classification accuracy.

Integrating real-time manufacturing data into a novel serial two-stage adaptive alternate genetic fireworks algorithm for solving stochastic type-II simple assembly line balancing problem

Due to various interference factors, a pre-planned assembly scheme and its cycle time can be disturbed, resulting in the failure of product delivery on schedule. However, when assembly data of the production line can be obtained in real time, the balance of the assembly line could be dynamically adjusted in case of experiencing serious interferences to optimize its cycle time in time and improve its production efficiency. Therefore, this paper proposes a dynamic rebalancing framework by integrating real time manufacturing data into a novel serial two-stage adaptive alternate genetic fireworks algorithm for solving a stochastic type-II simple assembly line balancing problem (SSALBP-II). MES (manufacturing execution system) is used to obtain some real time data such as the resource status, operation information and task information and to judge abnormal phenomenon and the overdue delivery caused by interferences. On this basis, the stochastic type-II simple assembly line balance model is constructed, with a new serial two-stage adaptive alternate genetic fireworks algorithm (STAGFA). This new algorithm can incorporate both genetic algorithm and fireworks algorithm to solve the model according to the transformation of population diversity discrimination index. Through the comparison between STAGFA and other algorithms such as fireworks algorithm, genetic algorithm, other improved intelligent algorithms, it is proved that the STAGFA is effective and superior in solving the assembly line (re)balance problem. Then, the rebalanced scheme verified by simulation is dispatched to control the physical assembly line and realize dynamic rebalancing cycle time effectively.

Multi-Drone Optimal Mission Assignment and 3D Path Planning for Disaster Rescue

In a three-dimensional (3D) disaster rescue mission environment, multi-drone mission assignments and path planning are challenging. Aiming at this problem, a mission assignment method based on adaptive genetic algorithms (AGA) and a path planning method using sine–cosine particle swarm optimization (SCPSO) are proposed. First, an original 3D digital terrain model is constructed. Second, common threat sources in disaster rescue environments are modeled, including mountains, transmission towers, and severe weather. Third, a cost–revenue function that considers factors such as drone performance, demand for mission points, elevation cost, and threat sources, is formulated to assign missions to multiple drones. Fourth, an AGA is employed to realize the multi-drone mission assignment. To enhance convergence speed and optimize performance in finding the optimal solution, an AGA using both the roulette method and the elite retention method is proposed. Additionally, the parameters of the AGA are adjusted according to the changes in the fitness function. Furthermore, the improved circle algorithm is also used to preprocess the mission sequence for AGA. Finally, based on the sine–cosine function model, a SCPSO is proposed for planning the optimal flight path between adjacent task points. In addition, the inertia and acceleration coefficients of linear weights are designed for SCPSO so as to enhance its performance to escape the local minimum, explore the search space more thoroughly, and achieve the purpose of global optimization. A multitude of simulation experiments have demonstrated the validity of our method.

Enhanced adaptive neuro-fuzzy inference system using genetic algorithm: a case study in predicting electricity consumption

Energy forecasting is crucial for efficient energy management and planning for future energy needs. Previous studies have employed hybrid modeling techniques, but insufficient attention has been given to hyper-parameter tuning and parameter selection. In this study, we present a hybrid model, which combines fuzzy c-means clustered adaptive neuro-fuzzy inference system (ANFIS) and genetic algorithm (GA), named GA–ANFIS–FCM, to model electricity consumption in Lagos districts, Nigeria. The model is simulated using the algorithms’ control settings, and the best model is identified after assessing their performance using renowned statistical indicators. To further narrow down the best viable model, the impact of the core parameter of the GA on the GA–ANFIS–FCM optimal model is examined by varying the crossover percentage in the range of 0.2–0.6. Firstly, the results reveal the better performance of the hybridized ANFIS model than the standalone ANFIS model. Additionally, the best model is obtained with the GA–ANFIS–FCM model with four clusters at a crossover percentage of 0.4, with mean absolute percentage error (MAPE), mean absolute error (MAE), coefficient of root mean square error (CVRMSE), root mean square error (RMSE) values of 7.6345 (signifying a forecast accuracy of 92.4%), 706.0547, 9.4913, and 918.6518 during the testing phase, respectively. The study demonstrates the potential of the proposed model as a reliable tool for energy forecasting.

An 18‐bit 1‐GS/s time interleaved analog‐to‐digital converter with timing skew calibration based on an adaptive genetic algorithm

SummaryA foreground calibration method for timing skews of the time interleaved analog‐to‐digital converter (TIADC) system is proposed in this paper. In this method, the adaptive genetic algorithm (AGA) is used to replace the traditional least mean square (LMS) method to detect the timing skews, and the variable delay lines are used to calibrate the timing skews. This method can accomplish the detection and calibration of positive and negative timing skews, and the reference channels are not required. In addition, compared with other methods, this method has higher detection speed and detection accuracy. In this paper, a four‐channel 18‐bit 1‐GS/s TIADC system is modeled, and the performance of the proposed method is further verified. The results show that the convergence speed and the calibration accuracy of this method are higher, the signal‐to‐noise ratio (SNR) of the TIADC system is improved by 65.9 dB after calibration, and this method is also effective when the frequency of the input signal is close to the Nyquist frequency.

A novel adaptive dynamic GA combined with AM to optimize ANN for multi-output prediction: Small samples enhanced in industrial processing

Multi-output prediction based on small samples (SSs) produced in industrial processing plays an important role in improving processing efficiency and reducing production costs. However, insufficient sample capacity will result in inaccurate and unstable multi-output regression prediction. To solve this problem, this paper jointly optimizes an artificial neural network (ANN) by employing an adaptive dynamic genetic algorithm and adaptive moment estimation (ADGA-AM-ANN) based on enhanced SSs. The data are enhanced by adding noise to the output data to prevent overfitting due to insufficient sample capacity. The adaptive dynamic genetic algorithm is employed to grope for the optimal parameters of the ANN in the global range. This prevents the ANN from falling into a local minimum in the training process, so as to improve the overall accuracy of the model in simultaneously predicting multiple variables, and to promote its internal prediction stability. Adaptive moment estimation is employed to optimize the gradient of optimal parameters to accelerate convergence and deepen the external prediction stability of the model. The cases of cutting and laser processes are used to verify the accuracy and stability of the proposed method.

Optimization of High-Speed Rail Express Transportation Plan Considering Products’ Variable Time Requirements

The global express market has expanded significantly in recent years. High-speed rail express freight transportation has been implemented between multiple cities in China, significantly improving transportation efficiency and effectively meeting demand. In this paper, an optimization model for HSR (high-speed railway) express transport plan is constructed using a space-time-state network, and an improved adaptive genetic algorithm is designed to solve the model. Reasonable transport schemes could be arranged for each cargo flow with the objective of maximizing operational revenue. The numerical experiment based on the Harbin–Dalian high-speed railway in China verifies the effectiveness and applicability of the proposed model. The results show that for HSR operators, appropriate transport strategies can be planned in terms of delivery priorities and transportation mode selection to make full use of the remaining capacity of HSR trains and increase operational revenue.

A New Performance Optimization Method for Linear Motor Feeding System

The linear motor feeding system is a typical electromechanical coupling system. Conventional characteristic analyses of electromechanical coupling often overlook the influence of flexible deformation in critical components of the linear motor feeding system. Moreover, when employing genetic algorithms to optimize servo system PID control parameters, slow convergence, nonconvergence, or premature convergence problems may arise. To address these issues, this paper proposes a new performance optimization method for a linear motor feeding system. The method uses a combination of “multi-body theory + finite element” to accurately account for the flexible deformation of critical components of the feeding system, establishes a rigid–flexible electromechanical coupling model of the linear motor feeding system, and optimizes the PID parameters of the established model with an improved adaptive genetic algorithm. Simulation results demonstrate that, when utilizing an adaptive genetic algorithm to optimize the rigid–flexible electromechanical coupling model and a control system model that disregards flexible body deformation, the system achieves stability in 0.02 s and 0.027 s with overshoots of 13% and 27%, respectively. These outcomes confirm the accuracy and importance of considering flexible body deformation in the optimization performance of a linear motor feeding system. At the same time, the time required to reach the steady state of the rigid–flexible electromechanical coupling model optimized by the adaptive genetic algorithm is shortened from 0.035 s to 0.02 s. The sinusoidal signal response curve of the optimized system does not exhibit any peak overshoot compared with that of the nonoptimized system, and the response speed is also faster. These results demonstrate the effectiveness of the rigid–flexible electromechanical coupling model optimized by the nonlinear adaptive genetic algorithm. The displacement response curves of the linear motor feeding system under different workbench loads are obtained through experiments and compared with those obtained from simulations to verify the established model and the correctness of the proposed method.

A Model-Assisted Combined Machine Learning Method for Ionospheric TEC Prediction

In order to improve the prediction accuracy of ionospheric total electron content (TEC), a combined intelligent prediction model (MMAdapGA-BP-NN) based on a multi-mutation, multi-cross adaptive genetic algorithm (MMAdapGA) and a back propagation neural network (BP-NN) was proposed. The model combines the international reference ionosphere (IRI), statistical machine learning (SML), BP-NN, and MMAdapGA. Compared with the IRI, SML-based, and other neural network models, MMAdapGA-BP-NN has higher accuracy and a more stable prediction effect. Taking the Athens station in Greece as an example, the root mean square errors (RMSEs) of MMAdapGA-BP-NN in 2015 and 2020 are 2.84TECU and 0.85TECU, respectively, 52.27% and 72.13% lower than the IRI model. Compared with the single neural network model, the MMAdapGA-BP-NN model reduced RMSE by 28.82% and 24.11% in 2015 and 2020, respectively. Furthermore, compared with the neural network optimized by a single mutation genetic algorithm, MMAdapGA-BP-NN has fewer iterations ranging from 10 to 30. The results show that the prediction effect and stability of the proposed model have obvious advantages. As a result, the model could be extended to an alternative prediction scheme for more ionospheric parameters.

Biological 2,4,6-trinitrotoluene removal by extended aeration activated sludge: optimization using artificial neural network

Serious health issues can result from exposure to the nitrogenous pollutant like 2,4,6-trinitrotoluene (TNT), which is emitted into the environment by the munitions and military industries, as well as from TNT-contaminated wastewater. The TNT removal by extended aeration activated sludge (EAAS) was optimized in the current study using artificial neural network modeling. In order to achieve the best removal efficiency, 500 mg/L of chemical oxygen demand (COD), 4 and 6 h of hydraulic retention time (HRT), and 1–30 mg/L of TNT were used in this study. The kinetics of TNT removal by the EAAS system were described by the calculation of the kinetic coefficients K, Ks, Kd, max, MLSS, MLVSS, F/M, and SVI. Adaptive neuro fuzzy inference system (ANFIS) and genetic algorithms (GA) were used to optimize the data obtained through TNT elimination. ANFIS approach was used to analyze and interpret the given data, and its accuracy was around 97.93%. The most effective removal efficiency was determined using the GA method. Under ideal circumstances (10 mg/L TNT concentration and 6 h), the TNT removal effectiveness of the EAAS system was 84.25%. Our findings demonstrated that the artificial neural network system (ANFIS)-based EAAS optimization could enhance the effectiveness of TNT removal. Additionally, it can be claimed that the enhanced EAAS system has the ability to extract wastewaters with larger concentrations of TNT as compared to earlier experiments.

Analysis and Evaluation of Global Temperature Changes and the Prediction of Future Global Temperature

Temperatures have continued to rise in recent years, many areas have shown extreme temperatures, and global warming has been a perennial topic. We use various optimization prediction and optimization evaluation methods in this paper to determine the trends and causes of global temperature variation and make effective suggestions for them. We first constructed a first-order differential temperature model and found that the temperature rise in March 2022 was larger than that observed in the past 10 years. By constructing a prediction model (BP) based on BP neural network and a grey temperature prediction model (GAGM) optimized based on adaptive genetic algorithm, and by importing the data, we find that the average temperature in 2116 and 2145 will reach 20℃. In addition, if the correlation factor evaluation model (FAEM) is constructed in the country, and we conclude that there is a correlation between location, time and temperature. Next, we analyzed key environmental factors, such as volcanic eruptions, forest fires, and COVID-19. We imported the CO 2 concentration and temperature data into the FAEM evaluation model to obtain the results that natural disasters can lead to the temperature increase. Then nine environmental factors, including COND, NYKC, NYXH, GDP, PJRK, LJL, LZMJ, KJSP, and GJJY. Taking these factors as independent variables and temperature as the dependent variables, COND, NYXH, PJRK, LJMJ, JSL are the main causes of global temperature change and KJSP. This study provides an assessment of future global temperature changes, making recommendations for policymakers to optimize the energy mix and plant trees to effectively mitigate the global temperature rise.

Mobile chargers scheduling algorithm for maximum data flow in wireless sensor networks

Most nodes in wireless sensor networks (WSNs) are battery powered. However, battery replacement is inconvenient, which severely limits the application field of the networks. In addition, the energy consumption of nodes is not balanced in WSNs, nodes with low energy will seriously affect data transmission capability. To solve these problems, we utilize mobile chargers (MCs) in WSNs, which can move by itself and charge low-energy nodes. Firstly, we construct a mixed integer linear programming model (MILP) to solve maximum flow problem, which is proved to be NP-hard problem. To maximize flow to the sink nodes, the BottleNeck algorithm is used to generate the initial population for the genetic algorithm. This algorithm takes path as the unit and schedules MCs to charge the lowest energy node first. Then, the improved adaptive genetic algorithm (IAGA) is utilized to simulate the natural evolution process and search for the optimal deployment location for MCs. The experiment results show that IAGA can effectively improve the maximum flow of sink node compared with other methods.

Reliable design of urban surface-underground integrated logistics system network with stochastic demand and social-environmental concern

Transforming urban goods movement from roads to underground is an emerging concept to improve supply chain performance and alleviate overcrowded environments. Existing literature merely discussed the underground logistics system (ULS) network design problem within a simplified boundary, whereas several critical system features (e.g., stochastic freight demand and co-modality) were omitted. To bridge the knowledge gaps, this paper proposes a reliable modeling approach to solve an urban surface-underground integrated logistics network (USUIL) design problem with a holistic consideration of cost-benefit objectives, road-ULS modal split, and demand uncertainties. To begin with, we introduce a city-wide hub-and-spoke USUIL network structure with a two-tier parallel distribution process. Five major external benefits of underground goods movement are formulated using real-world data. Next, a deterministic model is developed to optimize the network decisions including hub-satellite location-allocation, deployment of tunnels and pipelines, logistics service assignment, and flow routing. Then, to deal with the influences of case-wise demand uncertainties, the model is reformulated into a reliable one by incorporating the ideas of solution robustness and stochastic equivalents. A two-stage heuristic-based optimization tool that hybridizes the adaptive immune genetic algorithm, plant growth simulation algorithm, and Clarke-Wright saving algorithm was proposed to solve the problem. Finally, numerical experiments are conducted on the Beijing city case to validate the proposed method, and the best network schemes obtained from different models and scenarios are compared. Results show that a greater robust control factor in our model helps the network to lower the integrated freight transport cost while improving the social-environmental benefit at the expense of deploying more facilities. The reliable model performs well in balancing the conservativeness (i.e., less capacity insufficiency risk) and optimality (i.e., less cost due to redundant facilities) of the USUIL network design when faced with diverse demand cases.

Electric vehicle charging scheduling considering infrastructure constraints

The impacts of large-scale electric vehicles (EVs) charging on the power grid and the lack of charging infrastructure may directly hinder the promotion of EVs. With the limited number of charging piles and maximum instantaneous power at the charging station, how to effectively charge scheduling for EVs and reduce the charging cost for users becomes an important issue. To address this problem, we propose an EV charging scheduling strategy in response to time-of-use price. Here, the least cost of charging is set as the objective function and the limitations of charging piles number and instantaneous power of the stations are constraints. EV charging behavior characteristic is simulated using the Monte Carlo method based on 876,012 sets of historical charging data. Then, after solving the optimization problem by the adaptive genetic algorithm, each EV is assigned a specific charging pile that can meet its charging demand. The experimental results show that the proposed method can achieve better results than the comparative methods while ensuring the safe operation of charging stations. The effect of peak and valley reduction on the grid side is also realized.

Fast, Lightweight, and Efficient Cybersecurity Optimization for Tactical–Operational Management

The increase in frequency and complexity of cyberattacks has heightened concerns regarding cybersecurity and created an urgent need for organizations to take action. To effectively address this challenge, a comprehensive and integrated approach is required involving a cross-functional cybersecurity workforce that spans tactical and operational levels. In this context there can be various combinations of cybersecurity actions that affect different functional domains and that allow for meeting the established requirements. In these cases, agreement will be needed, but finding high-quality combinations requires analysis from all perspectives on a case-by-case basis. With a large number of cybersecurity factors to consider, the size of the search space of potential combinations becomes unmanageable without automation. To solve this issue, we propose Fast, Lightweight, and Efficient Cybersecurity Optimization (FLECO), an adaptive, constrained, and multi-objective genetic algorithm that reduces the time required to identify sets of high-quality cybersecurity actions. FLECO enables productive discussions on viable solutions by the cross-functional cybersecurity workforce within an organization, fostering managing meetings where decisions are taken and boosting the overall cybersecurity management process. Our proposal is novel in its application of evolutionary computing to solve a managerial issue in cybersecurity and enhance the tactical–operational cybersecurity management process.

Classification of Water Source in Coal Mine Based on PCA-GA-ET

In recent years, inrush water has hampered the regular mining of coal mines, and the proper identification of the source of inrush water is critical to the prevention and management of water hazards in mines. This paper extracts the standard water chemistry discriminating ions Na++K+, Ca2+, Mg2+, Cl−, SO42−, and HCO3− from observed water samples. An improved water source discrimination model is proposed which combines algorithms from data mining, classification models, and learning reinforcement. According to the Pearson correlation coefficient, Na++K+ has a strong correlation with HCO3−. To identify the major metrics, we performed principal component analysis (PCA), and the adaptive differential evolutionary genetic algorithm (GA) was utilized to optimize the depth of the extreme tree (ET) and the number of classifiers. Finally, the model distinguished 25 sets of studied samples from various water sources in the Pingdingshan coalfield. Comparative analysis demonstrated the efficacy of each stage of our work. PCA-GA-ET outperformed the conventional approaches, such as the support vector machine, BP artificial neural network, and random forest. The studies revealed that PCA-GA-ET can eliminate the information overlap between data and simplify the data structure and thereby improve the efficiency and accuracy of water source detection. We discovered that by utilizing the evolutionary algorithm to optimize parameters such as the depth of the extreme trees and the number of decision trees, we could get the model to converge faster and to be more stable and more accurate. The results suggest that PCA-GA-ET has good robustness and accuracy and can meet the needs of water source identification.