genetic algorithm

This article presents a novel hybrid model for credit scoring of banking customers, combining Genetic Algorithm, Bayesian Optimization, and the XGBoost machine learning model. The primary goal of this model is to enhance accuracy and efficiency in credit risk assessment and reduce the costs associated with prediction errors. In this study, real-world data from banking customers were utilized, and after preprocessing, including normalization and handling of missing data, the Genetic Algorithm was employed for optimal feature selection. Subsequently, Bayesian Optimization was applied as an advanced tool to fine-tune the hyperparameters of XGBoost. The results indicate the superior performance of the proposed model compared to conventional credit rating methods. The hybrid model achieved an accuracy of 79.3% and demonstrated excellent classification performance for both creditworthy and non-creditworthy customers, particularly in high-risk categories. Statistical analyses and performance comparisons with existing methods confirm the positive impact of feature selection and optimized hyperparameter tuning. This model can serve as a practical tool for banks and financial institutions to mitigate credit risk and improve customer management.

The ease of converting electricity into other energies has caused human activities to be highly dependent on it. Environmental issues, limited resources, unfair distribution of resources and the strong dependence of countries' economies on energy have caused more attention to be paid to optimizing the mix of electricity producing resources. Also, different goals are proposed in policy-making and comprehensive planning, which are in conflict with each other; so It’s important to optimize the composition of the energy production portfolio based on what purpose. In this research, the electricity generation portfolio is a combination of fossil, renewable and nuclear energy groups, and optimization is done using the genetic algorithm method with MATLAB software in single-objective (GA) and multi-objective (NSGA-II) modes.The comparison of optimization methods showed that the share of energies for single-objective optimization based on economic index is different from modes based on environmental index and Multi-purpose optimization. Through multi-objective optimization, a set of solutions is introduced, each of which fulfills one of the objectives at an acceptable level. In other words, achieving general optimal points is more likely than single-objective mode. single-objective optimization usually introduces local optimal points and leads to misguidance and moving away from the main goal.

Clustering is one of the basic techniques in data mining and machine learning, which is used to divide a set of data into homogeneous subsets. There are different methods for clustering, each of which has its own strengths and weaknesses. One of the main challenges in clustering is finding the optimal number of clusters and optimal allocation of data to these clusters. Genetic algorithm, as an optimization method based on natural evolution, has a high ability to solve complex problems and search for large solution spaces and can be used as an effective tool in clustering. The purpose of this article is to investigate the efficiency and accuracy of genetic algorithm in data classification and compare it with traditional clustering methods for classification. In order to evaluate the performance of this algorithm, several insurance data sets are used and the obtained results are analyzed with different criteria such as accuracy. Also, different parameters of the genetic algorithm are examined and their effects on the final performance of the algorithm are studied in order to determine the most optimal settings for data classification.

In this research, to form chromosomes, at first, the number of clusters was determined. Considering that each cluster center had as many features as the number of features in the data set, the length of each chromosome was determined by multiplying the number of clusters by the number of features. New and diverse methods were used for Crossover, Mutation and Survival processes. Also, the evaluation criterion similar to the K-means algorithm was chosen to optimize the clustering performance. This innovative approach led to improving the accuracy and efficiency of the classification process.

By applying the method described in this article to three insurance data sets for fraud detection, we have interesting results with 12% improvement in F1 and 10% increase in accuracy in the first data set, 1% improvement in F1 and 1% improvement in accuracy in the first data set. Second and finally, 1% improvement in F1 and 2% improvement in the accuracy of the third data set compared to the K-means method and other methods have been achieved. Due to the 2-mode data in this data set, the problem is solved for two clusters using the algorithm and the best label for each cluster is selected according to the real labels of the data and the result is presented as the results of classification problems. Additionally, significant improvements in metrics such as ARI and other clustering evaluation criteria have been achieved, and remarkable progress has been made compared to the standard genetic algorithm.

Genetic Algorithm is able to solve complex problems without definite solution and can perform better in data clustering than traditional methods such as K-means. By combining probabilities and randomness, this approach provides the possibility to examine more points as cluster centers and improve clustering performance. The results show that this method works better than the famous methods in some cases and provides a suitable structure for data clustering.

Railway tracks are among the most critical and costly components of any railroad. Optimal and timely maintenance allocation ensures system reliability and effective resource utilization. This article presents a mathematical model designed for condition-based preventive maintenance of railway tracks. The objective is to minimize total maintenance costs while maintaining system availability and reliability. The cost function considers maintenance and renewal expenses, track possession costs, and unplanned failure costs during the planning horizon. The network is divided into segments, with the standard deviation of the longitudinal level calculated as the track quality index. Prediction and recovery models are developed to assess variations in the track quality index over the planning period. Preventive operations, including inspections, tamping, and ballast replacement, are assigned to track segments based on the allowable index limits and other constraints. The expected number of random failures is estimated using the non-homogeneous Poisson process, and preventive operations aim to reduce the incidence of such failures.

Given the complexity and non-linearity of the mathematical model, a meta-heuristic method based on a genetic algorithm was developed in MATLAB software. Initial random solutions were generated using a heuristic method. As a case study, input parameters were collected and modeled for one kilometer of railway in the Arak region. The efficiency of the proposed genetic algorithm was then compared with the results obtained using GAMS software.

For the case study, the mathematical model was solved for various problem dimensions using both methods. When GAMS software was able to achieve the optimal solution, both methods produced identical results. When GAMS could only report a feasible solution, the genetic algorithm outperformed it. For higher dimensions where GAMS provided a solution, the genetic algorithm produced results with objective function values that were 40% better.

The planning of preventive maintenance operations for railway tracks during the project period is achieved through solving the proposed mathematical model. This allocation is based on the quality index of the tracks and considers the reducing effect of preventive operations on the expected number of random failures, ensuring the required minimum availability and reliability of the system. The variability of the model's input parameters allows for consideration of influencing heterogeneous factors based on the tracks, traffic, and environment, enabling adaptation to different conditions or the implementation of various strategies. Additionally, the proposed genetic algorithm can efficiently solve different dimensions of problem instances.

One of the major challenges of the performance audit is identifying the factors affecting the correct implementation and how these factors affect the operational process improvement of executive bodies. The current research aims to implement the performance audit in the Iran Audit Court. This research is applied. The present research has two communities. The first group comprises specialists and experts (audit organization auditors, Audit Court auditors, Audit Court accountants, and financial managers). The second community consists of National Audit Court employees and managers during the research period. In the first community, the snowball technique was used to determine the sample size, and 14 people were selected based on the principle of information saturation. In the second community, the convenience sampling method was used to determine the sample size using Cochran's formula (n=315). Eight effective factors in implementing performance audit in the National Audit Court were identified based on the theoretical basis. Regarding the Delphi-phase type 3 approach, the identified factors were confirmed. Then, 18 neural network, genetics, heuristic, and meta-heuristic approaches were investigated to investigate the effect of the effective factors on the implementation of performance audits in the National Audit Court. The gray wolf approach was more accurate than other methods. According to the results of the gray wolf method, infrastructure variables (0.3440); efficiency (0.1849); effectiveness (0.1134); economic efficiency (0.2947); cultivation (0.3657); specialist human resources (0.5984); organization performance (0.4609); management and leadership (0.6804) have a positive and significant effect on the implementation of performance audit.

Flood is one of the most destructive natural disasters with social, economic and environmental consequences. Therefore, comprehensive flood management is necessary to reduce the effects of floods on human life and livelihood. The main goal of this study is to investigate the application of the maximum entropy model (Entropy Maxent) in R software for flood susceptibility mapping in Fars province (Jahrom urban basin). First, by using the information of natural resources of Fars province and field visits, flood-prone points (50 points) were determined. In the next step, environmental variables such as altitude, slope, and distance from the river, drainage density, average annual rainfall, land use, soil type, and geology were selected by performing the multiple collinearity test, and vegetation cover and topographic humidity index were removed. Among the selected environmental variables, the three factors of height distance from the waterway and land use have had the greatest impact in the modeling process. After that, the receiver operating characteristic curve (ROC) was drawn for the flood sensitivity map, and the value of training data (0.943) and test data (0.932) was obtained. In the following, the model was optimized and upgraded using the genetic algorithm. As a result, this flood susceptibility map can be useful for researchers and planners in flood mitigation strategies.

This study evaluates the performance of the tri-hybrid RBF-GA-SARIMA meta-model in forecasting the frequency of dust storm days in Sistan and Baluchestan Province over a 50-year statistical period (1971–2020). The results obtained from this model were compared, using goodness-of-fit metrics, with those from individual models (SARIMA and RBF) and dual hybrid models (SARIMA-GA, RBF-GA, and RBF-SARIMA). All models demonstrated their peak accuracy and performance across all five stations during the fourth seasonal combination. However, substituting older seasonal data with one or two preceding seasons reduced accuracy and increased relative error in forecasting the FDSD index for the province. This phenomenon is attributed to the deposition of sand and dust particles during earlier seasons, followed by their subsequent mobilization by strong winds, triggering storms in later seasons. Among the evaluated models, the proposed hybrid tri-model exhibited superior accuracy and efficiency, emerging as the most effective approach for predicting the frequency of dust storm days. In contrast, the RBF-SARIMA dual hybrid model displayed the weakest performance, characterized by the lowest R-value and highest RMSE. These findings underscore that the integration of individual models does not inherently enhance the precision of climatic variable modeling.

the purpose of this research is tiering of suppliers and grouping of parts. The innovation of this research is to simultaneously consider the grouping of parts and allocation to suppliers based on the required operational capabilities of the parts and the capabilities available to suppliers and the ranking suppliers.

This research is quantitative-qualitative. Due to the use of experts' opinions and previous researches to collect data, it is qualitative and due to the use of a mathematical model, it is of a quantitative type. The method of collecting data and information is field type and checking related documents. The model is a non-linear mixed integer model with three objectives, including minimizing the cost of allocating parts to suppliers, minimizing the cost of interactions between suppliers, and minimizing management complexity. The model was evaluated with a case study in the electric motor industry and solved by the method of genetic algorithm of NSGA- II on MATLAB software; the validation of the results has also been done.

The results of the model implementation in the case study showed that four suppliers were selected as first-tier suppliers, 13 suppliers were selected as second-tier suppliers, and five suppliers could be removed or considered as substitutes. The results related to the allocation of parts and grouping also resulted in four groups of parts.

When two or more organizations seek to enhance their collaboration, they may opt to form strategic alliances. These alliances aim to progress cooperatively toward shared goals through resource sharing, while maintaining their independence. Maturity models are essential tools for aiding manufacturing organizations in developing their partnerships. However, there is a lack of empirical research on creating a strategic alliance maturity model with clear guidelines. Consequently, no existing model effectively measures the maturity of a strategic alliance, particularly one that can address inaccuracies due to human judgment and inherent evaluation uncertainties. This research aims to design a method to assess the maturity level of strategic alliances, providing a better understanding of the current state of cooperation based on strategic alliance maturity criteria.

This research developed a maturity model using fuzzy logic to evaluate the status of a collaboration at a specific point in time. The strategic alliance maturity model, based on fuzzy logic, was created through a clear and precise procedure and a multi-method approach, including literature reviews, interviews, focus groups, and case studies. A weighted fuzzy inference system combined with the fuzzy data envelopment analysis technique was employed to measure maturity levels using specific indicators. Additionally, a genetic algorithm was applied to generate a set of fuzzy rules. Like all maturity models, the one developed in this research consists of two main components: the maturity levels of the strategic alliance and its maturity dimensions.

The research model defines five levels of maturity: ad-hoc, initial, managed, planned, and optimized strategic alliance. A 44-item list of indicators for measuring strategic alliance maturity was compiled from articles, expert interviews, and analysis of successful and unsuccessful alliances. This list was then categorized into 17 criteria across six dimensions. Essential indicators were identified using the content validity ratio technique, and their relative importance was determined through data envelopment analysis. Expert surveys were used to create fuzzy sets for the variables of the fuzzy inference system. Additionally, a set of fuzzy rules was developed by examining examples of strategic alliances both domestically and internationally, and refining them through expert surveys.

The proposed model has been evaluated and validated through a real case study involving collaboration between a manufacturing organization and its business partner. The research results demonstrate that this approach offers a robust and practical diagnostic tool based on a set of strategic alliance maturity indicators. By analyzing the gaps identified by this model, an action plan can be devised to enhance the maturity level of the strategic alliance.

In order to control the supply and demand of energy and the correct planning in directing the consumption, the monthly gas consumption of the country's industries was investigated with artificial neural network and genetic algorithm in this research for 1402. The country's population information, industrial producer price index, gross domestic product at a fixed price of 90 and gas consumption of the country's industries were investigated as influential variables. The results showed that the best artificial neural network combined with the genetic algorithm is a network with a mutation rate of 0.5, an intersection rate of 0.5, the number of repetitions of 150, and the initial population size of 150. The combination of artificial neural network and genetic algorithm showed that in the spring season there was a total of 2957.7 million barrels of crude oil equivalent, in the summer this figure increased to 3502.6, in the autumn to 4329.9 and in the winter with a growth of 8.15 percent. It will reach 4683 million barrels of crude oil equivalent

To reduce the conflict of interests between managers and shareholders, it is crucial to focus on the sharing of benefits. Managerial remuneration is one way to address this conflict and serves as a tool to align managers' perspectives and performance with the goal of increasing shareholder wealth. Cash rewards for managers should be performance-based to ensure their alignment with shareholder interests. When designing the CEO's salary package in companies, the role of institutional investors is significant. Agency theory highlights the problems that arise when owners delegate the management of the company to managers. To mitigate agency conflicts, managerial rewards should be tied to the value created for shareholders. One of the primary methods for measuring managerial performance is through accounting reports, which act as tools for assessing and motivating managerial performance. Given these points, the aim of this research is to provide a model for predicting the sensitivity of CEO compensation using meta-heuristic algorithms, specifically genetic algorithms and particle swarm optimization.
The statistical population of this research comprises all companies listed on the Tehran Stock Exchange from 1390 to 1400. To select the sample, a systematic elimination method was employed, resulting in a sample of 110 companies. Based on the classification of research according to its purpose, the current study is applied in nature. Additionally, it is a quasi-experimental study within the domain of descriptive research (non-experimental survey). Data collection methods for this research include document analysis, internet research, and library study, depending on the specific requirements. In this research, 12 parameters influencing the sensitivity of CEO compensation were selected: institutional ownership, family ownership, comparability of financial statements, profit management, conditional conservatism, income and cost matching, market added value, corporate acquisition, debt contracts, and cost behavior (categorized into three types: changes in asset returns, changes in sales revenue, and changes in operating costs). These parameters were used as inputs for the data mining model. The sensitivity of CEO service compensation was chosen as the output parameter. Three data mining models were created by separating the cost behavior parameter, and for comparison, three linear regression models were also employed.
The results demonstrate the superiority of the deep neural network model in terms of the coefficient of determination and MSE index. This superiority holds true for all three data mining models compared to the three linear regression models. Among the data mining models, the third model, which incorporates the cost behavior parameter of changes in operational costs, produced the best results. The second model, which includes the cost behavior parameter of changes in sales revenue, achieved the next best results. Finally, the first data mining model, which uses the cost behavior parameter of asset return changes, delivered the weakest results.
The application of deep neural networks, optimized by meta-heuristic algorithms, can create predictive models based on real data, which can be used for management decisions and enhancing service compensation processes in organizations. These methods offer the potential to improve CEO performance and the quality of services provided by organizations by leveraging existing data and artificial intelligence algorithms. Furthermore, this research can assist investors and economic decision-makers in more accurately analyzing and predicting the sensitivity of CEO compensation using deep neural networks and meta-heuristic algorithms.

 The lack of certainty in the movement of the stock market has made forecasting a challenging task in the field of financial time series forecasting. On the other hand, it is not easy to analyze the time series data of stock prices due to non-linearity and high noise. Therefore, the aim of this research is to predict the long-term and short-term trend of the capital market. To achieve this goal, artificial intelligence algorithms of particles and genetics have been used in a comparative manner. The studied variable is the total stock price index in Tehran Stock Exchange in the period of 2016 to 2021 and on a monthly basis. The data have been reviewed after collection using the smoothing method for holidays, and in order to increase the accuracy of the models, the optimal window length of each algorithm has been calculated. The findings indicate that the genetic algorithm by minimizing the prediction error is a suitable algorithm for predicting the short-term and long-term trend of the total price index compared to the particle algorithm in the studied time period.

Banks face credit risk in the process of providing facilities based on the amount of resources provided. In the meantime, facility portfolio management can be effective in reducing credit risk by optimally allocating resources to economic sectors. In this research, the portfolio of Sina Bank's facilities is determined by using the Markowitz modern portfolio model and meta-heuristic algorithms of genetics and firefly. Comparing the performance of the models indicates the greatest efficiency of the genetic algorithm model in optimizing the bank's facilities portfolio. the results of the estimation of this model show that the service & commercial, housing & construction sectors have the largest share in the optimal portfolio of the bank's facilities. Industry & mining, agriculture & water sectors are considered risky assets of Sina Bank. the process of granting facilities has not been optimal. To reduce the credit risk of that bank's facilities, 52.4% should be allocated to the service & commercial sector, 40.7% to the housing & construction sector, 3.5% to the industry & mining sector, 3.4% to be allocated to agriculture & water sector.

Resilience has been defined as "a process of ability, or the result of successful adaptation despite threatening conditions", which plays an important role in dealing with the stresses and threats of life and its adverse effects. Also, resilient cities are vulnerable to events. mitigate extremes and respond creatively to environmental, economic and social changes in order to increase their long-term sustainability. Therefore, the purpose of this research is to predict the resilience of Qaemshahr city against natural hazards. The method used in this research is to use GIS software in order to determine points with more resilience and use genetic algorithm (GA) and particle swarm intelligence (PSO) methods and their programming in MATLAB environment, in order to Using it, the most important components of the city's resilience and the effective strategies to strengthen resilience will be determined. The present research was conducted in two stages, in the first stage, first, using the AHP model, the degree of resilience of urban areas against the 3 hazards of floods and earthquakes. and the landslide was determined and after that the final risk map (resilience level) was prepared by combining all the mentioned components. In the second stage, in order to determine the most important components of the city's resilience, this time 32 indicators of the resilience components were compiled based on 4 criteria (physical-spatial, economic, social, institutional) and analyzed using the genetic algorithm method. Finally, 3 points of the city (Shahid Watani Stadium, Telar Park, Siraj Park) were selected and prioritized as areas with more resilience. Then, in order to confirm and compare the research findings, the PSO method was used by the genetic algorithm (GA). As a result, it can be said that there is a significant difference in terms of resilience among the city's regions, that the increase in the resilience of Qaimshahr city The first stage requires strengthening and modifying the structures of the natural environment and then maintaining the safe environment.

Restatement of Financial Statements , which means re-presenting last year's financial information in order to correct incorrect information or increase comparability, is one of the common phenomena in Iran, which causes investors to lose confidence in the credibility and competence of management and the quality of reported profits. The purpose of this research is to present the development model of Benish fraud in companies admitted to the Tehran Stock Exchange between 2008 and 2019. The data of 265 companies were extracted using the selected and final indicators of Benish's (1999) fraud model, and neural network and genetic algorithms were also used to develop the prediction model. The results of the research indicate that based on the confusion matrix, the accuracy and efficiency of the improved Benish model with the combined algorithm of neural network and genetics in predicting unrepresented companies was 78.15%, and re-presented companies was 29.63%. . Also, the overall accuracy of the model presented with the combined algorithm of neural network and genetics was 73.21%.

With the passage of four generations of theoretical and experimental development of the smart city, it was expected that by relying on technologies such as the Internet of Things, artificial intelligence, cloud computing, and the like, this theory would be the way to face the problem and solve the complexities of resilience planning. But countless recent researches show the opposite. The root of the problem is not in the application of technology, but in the theoretical integration of smart city management and urban resilience. In the meantime, the resilience approach is a new point of view that instead of focusing on reducing vulnerability and looking at confrontation, it emphasizes increasing the resilience and flexibility of cities against natural and human hazards. The purpose of this research is to evaluate the feasibility of smart urban management to improve the resilience of Urmia city. To achieve this goal, a descriptive-analytical research method was used based on the analysis of a questionnaire that was completed by 30 urban experts and specialists. The results of the statistical analysis show that Urmia city is relatively favorable in terms of social and infrastructural resilience, but it does not have a favorable situation in terms of institutional-management, economic and physical-environmental resilience. In general, the overall resilience of Urmia city is lower than the optimal level according to the experts, so that the calculated value of the overall resilience of Urmia city is 2.90, which is less than the average level (3).