Maximum Iteration Research Articles

Reference-free calibration method for asynchronous rotation in robotic CT.

Geometry calibration for robotic CT system is necessary for obtaining acceptable images under the asynchrony of two manipulators. We aim to evaluate the impact of different types of asynchrony on images and propose a reference-free calibration method based on a simplified geometry model. We evaluate the impact of different types of asynchrony on images and propose a novel calibration method focused on asynchronous rotation of robotic CT. The proposed method is initialized with reconstructions under default uncalibrated geometry and uses grid sampling of estimated geometry to determine the direction of optimization. Difference between the re-projections of sampling points and the original projection is used to guide the optimization direction. Images and estimated geometry are optimized alternatively in an iteration, and it stops when the difference of residual projections is close enough, or when the maximum iteration number is reached. In our simulation experiments, proposed method shows better performance, with the PSNR increasing by 2%, and the SSIM increasing by 13.6% after calibration. The experiments reveal fewer artifacts and higher image quality. We find that asynchronous rotation has a more significant impact on reconstruction, and the proposed method offers a feasible solution for correcting asynchronous rotation.

Classification of Malignancy of Lung Cancer Using Backpropagation Algorithm on CT-Scan Images

In this study, we investigate the classification of lung cancer CT scan images based on malignancy level using a backpropagation artificial neural network (ANN). Lung cancer is a deadly disease characterized by the growth of abnormal lung cells. The proposed method involves preprocessing to enhance image quality, followed by feature extraction using the Gray Level Co-occurrence Matrix (GLCM) method with angle variations of 0°, 45°, 90°, 135°, and d=1. The extracted features include energy, contrast, correlation, and homogeneity. The energy value range in malignant cancer is 0.27 to 0.81, while in benign cancer it is 0.26 to 0.73. The contrast in benign cancer ranges from 1.38 to 11.87, while in malignant cancer it is 1.47 to 13.67. The image correlation for malignant cancer is between 0.63 to 0.94, while for benign cancer it is 0.69 to 0.96. Homogeneity in malignant cancer has a value range between 0.67 to 0.91, while in benign cancer it ranges from 0.70 to 0.92. The classification of lung cancer malignancy is restricted to benign and malignant levels using a network architecture of [4 10 2], maximum iteration of 100000, and learning rate of 0.001. The accuracy of the testing data from the ANN is between 90% and 100%. These results demonstrate the effectiveness of the GLCM method and backpropagation algorithm in accurately classifying the malignancy level of lung cancer, which could aid in the early detection and treatment of the disease.

Reduced Outer Space Algorithm for Globally Computing Affine Sum-of-Ratios Problems

This paper investigates the affine sum-of-ratios problem (SRP) which has numerous applications in many fields of economy and engineering. For globally computing the affine SRP, based on equivalent transformation and new affine relaxation problem, a reduced outer space branch-and-bound algorithm is designed. The designed algorithm has been proven to converge to a global optimal solution of the problem (SRP) eventually. Meanwhile, the maximum iteration time for the algorithm in the worst case is estimated by analyzing its computational complexity. Finally, numerical experimental results indicate that the presented algorithm is robust and effective.

Prediction of Buildings’ Settlements Induced by Deep Foundation Pit Construction Based on LSTM-RA-ANN

In view of the shortcomings of existing methods for predicting the settlement of surrounding buildings caused by deep foundation pit construction, this study uses the monitoring data of a foundation pit project in Shanghai and divides the construction process of the pit into three working conditions, that is, enclosure construction, earthwork excavation, and basement support construction. The attention mechanism and residual update are integrated into the artificial neural network (ANN) model, and the root-mean-square error, average absolute error, and determination coefficient are used as the evaluation indices of the model. The artificial neural network prediction model LSTM-RA-ANN for building settlements in deep foundation pit construction was then established. The prediction performance of the model was also analysed under different working conditions, and the influences of the main factors (including the soil parameter, monitoring point location, activation function, hyperparameter, and input number) on the evaluation index was further explored. The results indicate that the performances of the established LSTM-RA-ANN model are closely related to the construction conditions, the predicted settlements agree well with the monitored ones in three working conditions with the greatest errors occurring at a later time of the working conditions, and the prediction accuracy of the great–small order corresponds to basement support, enclosure construction, and earthwork excavation respectively. The farther the monitoring point is from the edge of the pit, the better the model performance is. The activation function, initial learning rate, and maximum iteration batch have a great influence on the evaluation indices of the model, while the number of input points has little effect on the evaluation indices. These results may serve as a reference for the safe construction and normal operation of foundation pit engineering.

Using an interpretable deep learning model for the prediction of riverine suspended sediment load.

The prediction of suspended sediment load (SSL) within riverine systems is critical to understanding the watershed's hydrology. Therefore, the novelty of our research is developing an interpretable (explainable) model based on deep learning (DL) and Shapley Additive ExPlanations (SHAP) interpretation technique for prediction of SSL in the riverine systems. This paper investigates the abilities of four DL models, including dense deep neural networks (DDNN), long short-term memory (LSTM), gated recurrent unit (GRU), and simple recurrent neural network (RNN) models for the prediction of daily SSL using river discharge and rainfall data at a daily time scale in the Taleghan River watershed, northwestern Tehran, Iran. The performance of models was evaluated by using several quantitative and graphical criteria. The effect of parameter settings on the performance of deep models on SSL prediction was also investigated. The optimal optimization algorithms, maximum iteration (MI), and batch size (BC) were obtained for modeling daily SSL, and structure of the model impact on prediction remarkably. The comparison of prediction accuracy of the models illustrated that DDNN (with R2 = 0.96, RMSE = 333.46) outperformed LSTM (R2 = 0.75, RMSE = 786.20), GRU (R2 = 0.73, RMSE = 825.67), and simple RNN (R2 = 0.78, RMSE = 741.45). Furthermore, the Taylor diagram confirmed that DDNN has the highest performance among other models. Interpretation techniques can address the black-box nature of models, and here, SHAP was applied to develop an interpretable DL model to interpret of DL model's output. The results of SHAP showed that river discharge has the strongest impact on the model's output in estimating SSL. Overall, we conclude that DL models have great potential in watersheds to predict SSL. Therefore, different interpretation techniques as tools to interpret DL model's output (DL model is as black-box model) are recommended in future research.

Forecasting Maximum Water Level Data for Post Sangkuliman using An Artificial Neural Network Backpropagation Algorithm

Neural Network (NN) is an information processing system that has characteristics similar to biological neural networks. One of the algorithms in NN is Backpropagation Neural Network (BPNN). BPNN is an excellent method for dealing with complex pattern recognition problems. In this research, maximum water level forecasting was carried out at Sangkuliman Post using a Backpropagation Neural Network. This research aims to obtain modeling for forecasting maximum water level, as well as forecasting results using the best model. The research results show that the best model is five neurons in hidden layer 1 and 3 neurons in hidden layer 2 with the backpropagation algorithm, the activation function used is binary sigmoid, the learning rate is 0.001, and the maximum iteration is 10,000,000 with the smallest RMSE result being 1.816. The forecast results for the following 12 periods are 1.672, 1.779, 1.523, 1.271, 1.752, 1.692, 1.335, 1.479, 1.750, 1.779, 1.340, 1.269, and 1.754. Forecasting results can be used by various parties in decision-making and planning in multiple fields, as an example to see the patterns of biological and vegetable life around Sangkuliman Post. Based of forecasting results, certain months show an increase in maximum water levels.

Optimizing Clustering of Indonesian Text Data Using Particle Swarm Optimization Algorithm: A Case Study of the Quran Translation

The Quran considered the holy book for Muslims, contains scientific and historical facts affirming Islam's truth, beauty, and influence on human life. Consequently, the Quran text and its translations are valuable sources for text mining research, particularly for studying the interrelationship of its verses. One approach to grouping objects using certain algorithms is clustering, with K-Means Clustering being a prominent example. However, clustering results are often suboptimal due to the random selection of centroids. To address this, the study proposes using the Particle Swarm Optimization (PSO) algorithm, which selects centroids based on PSO results. The hybrid PSO algorithm initiates a single iteration of the K-means algorithm. It concludes either upon reaching the maximum iteration limit or when the average shift in the center of the mass vector falls below 0.0001. Evaluation of the clustering results from the three models indicates that the K-Means algorithm produced the lowest Sum of Squared Error (SSE) value of 1032.19. Additionally, the hybrid PSO algorithm generated the highest Silhouette value of 0.258 and the lowest quantization value of 0.00947. Further evaluation using a confusion matrix showed that K-Means clustering had an accuracy rate of 81.7%, K-Means with PSO had 82.5%, and the combination of K-Means with hybrid PSO yielded the highest accuracy rate of 91.1% among the three grouping model.

Identifying risk zones and landscape features that affect common leopard depredation.

Human-wildlife conflict (HWC) is a pressing issue worldwide but varies by species over time and place. One of the most prevalent forms of HWC in the mid-hills of Nepal is human-common-leopard conflict (HLC). Leopard attacks, especially in forested areas, can severely impact villagers and their livestock. Information on HLC in the Gorkha district was scarce, thus making it an ideal location to identify high-risk zones and landscape variables associated with such events. Registered cases were collected and reviewed from the Division Forest Office (DFO) during 2019-2021. Claims from DFO records were confirmed with herders and villagers via eight focus group discussions. To enhance modeling success, researchers identified a total of 163 leopard attack locations on livestock, ensuring a minimum distance of at least 100 meters between locations. Using maximum entropy (MaxEnt) and considering 13 environmental variables, we mapped common leopard attack risk zones. True Skill Statistics (TSS) and area under receiver-operator curve (AUC) were used to evaluate and validate the Output. Furthermore, 10 replications, 1,000 maximum iterations, and 1000 background points were employed during modeling. The average AUC value for the model, which was 0.726 ± 0.021, revealed good accuracy. The model performed well, as indicated by a TSS value of 0.61 ± 0.03. Of the total research area (27.92 km2), about 74% was designated as a low-risk area, 19% as a medium-risk area, and 7% as a high-risk area. Of the 13 environmental variables, distance to water (25.2%) was the most significant predictor of risk, followed by distance to road (16.2%) and elevation (10.7%). According to response curves, the risk of common leopard is highest in the areas between 1.5 to 2 km distances from the water sources, followed by the closest distance from a road and an elevation of 700 to 800 m. Results suggest that managers and local governments should employ intervention strategies immediately to safeguard rural livelihoods in high-risk areas. Improvements include better design of livestock corrals, insurance, and total compensation of livestock losses. Settlements near roads and water sources should improve the design and construction of pens and cages to prevent livestock loss. More studies on the characteristics of victims are suggested to enhance understanding of common leopard attacks, in addition to landscape variables. Such information can be helpful in formulating the best management practices.

Classifying toddler stunting based on anthropometric data using the learning vector quantization 3 (LVQ 3) method

Based on anthropometric data, this study classified news stunts using the Learning Vector Quantization 3 (LVQ 3) technique. Kaggle's secondary data consists of 6500 data points with seven main variables. The research phases include data gathering, preprocessing, transformation, normalization, modeling, testing, and assessment. Normalization normalizes data value ranges, while data transformation transforms non-numeric variables into numerical representations. The two layers of the LVQ-3 architecture—input and competitive—are used for modeling. Based on the data class and learning level modification, the neurons' weights are changed. We run experiments with different values for the LVQ 3 parameters, such as window, maximum iteration, and learning level. Confusion, precision, recall, and F1 score matrixes were used in the evaluation process. A learning level of 0.1 to 0.6 produced the best results, with a consistent accuracy of 0.72. The greatest accuracy of 0.7420 was obtained by dividing the data by 70:30 using a window between 0.3 and 1. The findings from this study show that, despite being influenced by the common use of variables and the lack of data variance, LVQ 3 performs quite well and has significant accuracy. To identify the value that best matches the data set conditions, more parameter exploration is required.

Research on Virtual Character Behavior Simulation and Control Algorithm in Digital Movie Production

Abstract Aiming at the inefficient design of virtual character behavior in traditional digital movies, this paper proposes to construct a virtual character behavior and control model. The virtual character’s behavior is constructed using a behavior tree, and perception and decision-making behavior models are created to optimize the visual perception of the virtual character. After the behavioral modeling is completed, the action controller is used to generate the character’s behavioral actions, combined with the physical dynamics method to capture the character’s trajectory, and the feedback control theory is applied to output the character’s joint rotation data information to realize the behavioral control of the virtual character. In the digital movie scene design of firefighters fighting and rescuing, the acquisition and execution time of a single behavioral modeling sample of this paper’s virtual character behavioral modeling method is 0.01s~0.03s, and the maximum running iteration time is 63.83s, which is faster than the agent-based behavioral modeling method. Compared to the traditional virtual character behavioral control method, which has an offset of more than 4mm, the virtual character firefighter generated in this paper has an offset of less than 0.3mm, which provides better control accuracy.

Global optimization algorithm for a class of linear ratios optimization problem

<abstract><p>We presented an image space branch-and-bound algorithm for globally minimizing the sum of linear ratios problem. In the algorithm, a new linearizing technique was proposed for deriving the linear relaxation problem. An image space region reduction technique was constructed for improving the convergence speed of the algorithm. Moreover, by analyzing the computational complexity of the algorithm, the maximum iterations of the algorithm were estimated, and numerical experimental results showed the potential computing benefits of the algorithm. Finally, a practical application problem in education investment was solved to verify the usefulness of the proposed algorithm.</p></abstract>

Nowcasting of Chili Pepper (Capsicum frutescens L.) Prices in East Java Province Using Multi-Layer Perceptron Method

The aims of study is to predict the price of chili pepper at the provincial level in East Java by looking for the best input variable from three types of input variables, price of chili pepper at the regency and city levels, natural factors, and word search index on Google Trends as an approach to the causes of chili pepper price fluctuations. The Multi-Layer Perceptron method, accompanied by a search for the best combination of model parameters is selected to get the model with the best nowcasting ability. The result shows that the best model for nowcasting is characterized by: the input variable is price of chili pepper at the regency and city levels with three hidden layers and 32, 45, and 51 neurons in each hidden layer, maximum iteration is 200 iterations, maximum iteration when the model not increase in performance for applying early stopping is 20 iterations, non-linear activation used is RELU (Rectified Linear Unit), and optimization function used is ADAM optimizer. The accuracy of nowcasting in this study is highly accurated with MAPE smaller than 10%.

A Novel Affine Relaxation-Based Algorithm for Minimax Affine Fractional Program

This paper puts forward a novel affine relaxation-based algorithm for solving the minimax affine fractional program problem (MAFPP) over a polyhedron set. First of all, some new variables are introduced for deriving the equivalence problem (EP) of the MAFPP. Then, for the EP, the affine relaxation problem (ARP) is established by using the two-stage affine relaxation method. The method provides a lower bound by solving the ARP in the branch-and-bound searching process. By subdividing the output space rectangle and solving a series of ARPs continuously, the feasible solution sequence generated by the algorithm converges to a global optimal solution of the initial problem. In addition, the algorithmic maximum iteration in the worst case is estimated by complexity analysis for the first time. Lastly, the practicability and effectiveness of the algorithm have been verified by numerical experimental results.

Deteksi Mata Katarak Berdasarkan Tekstur Gray Gray Level Co-Ocurrence Matrix Dengan Metode Self Organizing Map

Cataracts are a common eye problem in Indonesia. Untreated cataracts are the main cause of blindness and the most dominant vision impairment in Indonesia among people over 50 years old, with a proportion reaching 77.7%. Regular eye examinations are necessary to prevent cataracts, but there are obstacles due to the availability of equipment and the cost of eye examinations. Therefore, an efficient and effective cataract detection system is needed. This study aims to detect cataracts in the eyes by utilizing a dataset of eyes from the internet. This research uses RGB and HSV feature extraction combined with the GLCM extraction method. The Gray Level Co-occurrence Matrix (GLCM) method describes the spatial relationship between pixel intensities in an image. With the GLCM matrix, texture information can be extracted from eye images. The Self Organizing Map (SOM) method performs learning based on feature extraction obtained from previously labeled eye images. Based on the test results, this study successfully detects cataract and normal eyes. With the greatest accuracy achieved in the comparison of 80% training data and 20% test data with dimension size = [2 2] and maximum iterations of 100, the highest accuracy of 90% was obtained.

Fuzzy C-Means in Content-Based Document Clustering for Grouping General Websites Based on Their Main Page Contents

The research aimed to use Fuzzy C-Means clustering in content-based document clustering to classify general websites based on their content. The data used were a table ranking of the most visited websites for Indonesia, taken from https://dataforseo.com/top-1000-websites/ on September 24th, 2022. The research was conducted with two different cases using Fuzzy C-Means clustering, which had two different iteration parameter values, namely 100 and 200 in maximum iteration. The research results on Fuzzy C-Means clustering in content-based document clustering are based on the two cases. These different maximum iteration parameters result in a different amount of website name data in the cluster. They are formed in the first and second clusters only. However, in the other clusters, the numbers are all the same. The results of the cluster research are validated using the silhouette coefficient, with case no. 1 and no. 2 values being 0,977783879 and 0,977788457. The use of Fuzzy C-Means clustering in content-based document clustering has an excellent performance when this method is applied to group general websites based on their content. With that result, content-based clustering can be also applied in other cases. Hence, the results can be considered to be applied to other cases for content-based clustering in the future.

A time-bounded approach for reducing runtime of well placement optimization using PSO and PSOGA algorithms

Determining the optimal drilling locations and well production/injection rates is a highly complex decision in reservoir development. The present study aims to optimize well placement and production/injection rates through the application of intelligent search algorithms. The Particle Swarm Optimization (PSO) and combined Particle Swarm Optimization-Genetic Algorithm (PSOGA) have been chosen as the optimal means to achieve this objective. The Net Present Value (NPV) of the project serves as the objective function, with decision variables including the location of wells, perforation, and flow rates. To evaluate the performance of the algorithms in optimizing well placement and flow rates, a heterogeneous reservoir model has been used. In order to reduce optimization runtime, a novel method has been presented. This method involves changing the termination condition, such that the entire optimization time is considered rather than the maximum iterations. The population size can be added to enhance algorithm convergence. Thus, modified algorithms, namely TPSO and TPSO-GA, have been introduced. The results of determining the termination condition on the TPSO-GA algorithm are quite promising. By setting a 24-h termination condition or a 70% reduction in the basic runtime, and with a population size of 30 agents, the same values of the objective function can be obtained. Furthermore, the impact of two parameters within the objective function, namely the oil price and discount rate, has also been explored.

A spatial branch-reduction-bound algorithm for solving generalized linear fractional problems globally

In various engineering applications, the generalized linear fractional problem (GLFP) is an essential model that has many challenges in both theoretical and practical aspects. Therefore, the main work of this paper is to design an effective spatial algorithm for solving the GLFP efficiently. We start with equivalently converting the GLFP into an equivalent problem (EP) by transforming each fractional equation into one new variable. Next, applying the second-order cone relaxation to the constraint functions and executing a double-layer relaxation on the objective function of the EP, the second-order cone relaxed problem is constructed to underestimate the EP. Then, integrating some region reduction methods, we implement a spatial branch-reduction-bound algorithm. Furthermore, we verified the convergence of the proposed algorithm. Equally important, the maximum iterations of the proposed algorithm in the worst scenario are evaluated by the complexity analysis of the proposed algorithm. Finally, by comparing some algorithms in the current literature, numerical results confirm the feasibility, robustness, and efficiency of the proposed algorithm.

Dynamic global power extraction of partially shaded PV system using a hybrid MPSO-PID with anti-windup strategy

Particle Swarm Optimizer (PSO) has three main drawbacks in case of partially shaded photovoltaic systems. First, it converges slowly to the global peak (GP). Second, because PSO is a time invariant, it cannot adhere to a dynamic GP in the presence of various temporal shading patterns and instead, sticks to the first GP. Third, it exhibits high oscillations around GP. To overcome these shortcomings, this paper proposes a novel hybrid modified PSO-PID (MPSO-PID) with anti-windup strategy based maximum power point tracking. The anti-windup strategy aims to guarantee the robustness of the PID controller and eliminate the PID windup outcome. Two anti-windup structures were applied to PID namely: anti-windup using back calculation and anti-windup using conditional integration. For validation purpose, the hybrid MPSO-PID with anti-windup has been compared with conventional PSO and three modified PSO versions namely: PSO with random numbers eliminated (PSO-RNE), PSO with linearly varying parameters (PSO-LVP), and PSO with fixed maximum iterations (PSO-FMI). The obtained simulation and experimental results revealed that the proposed MPSO-PID with anti-windup overcome the previous shortcomings providing outperformance to track dynamic GP with a less convergence time and small oscillations around GP compared to other algorithms. The proposed MPSO-PID technique has a minimum average response time; 0.026 s; to catch the GP under various shading patterns compared to the other PSO versions; CPSO, PSO-LVP, PSO-RNE, and PSO-FMI; where the average response time are 0.026 s, 0.531s, 0.446 s, 0.763 s, and 0.494 s respectively. Finally, the results proved the robustness and effectiveness of PID design for enhancing both the PSO and PV system performance.

Learning Vector Quantization for Robusta and Arabica Coffee Classification

ANN or artificial neural network is a way to solve various kinds of problems to make decisions based on training. One of the methods of JSt which contains competitive and supervised learning. Where this layer will automatically learn the classification of the closest input distances and will be distributed to the same class. there are 2 types of coffee beans that are famous in the world, namely arabica and robusta, for some people or the layman it will be very difficult to distinguish these 2 types of coffee beans apart from the fact that the shape is almost the same the color looks almost the same but there are a number of differences in the two coffee beans which we can see from the shape of the seed. Robusta has a shape that tends to be round and smaller in size, and has a rougher texture. Arabica, on the other hand, is slightly flatter and longer in shape. The size is slightly bigger than Robusta but the texture of Arabica is smoother than Robusta. This is the basis of this study where the images of the two coffee beans will be extracted using the first-order texture feature extraction method based on MU parameters, standard deviation, skewness, energy, entropy, and smoothness. The method for collecting data was in the form of a quantitative method using images from each coffee bean, both Arabica and Robusta, with a total of 130 images. The comparison between training_data and test_data is 80:20. Through research conducted in the form of performance parameters with the best accuracy, including: Learning rate 0.01, max epoch or maximum iteration of 10 and 30%, the amount of training data used is 39 training images and 26 test images resulting in an accuracy presentation of 71% for the training process and error with a percentage of 96% for the test process.

Optimal Load Sharing between Lithium-Ion Battery and Supercapacitor for Electric Vehicle Applications

There has been a suggestion for the best energy management method for an electric vehicle with a hybrid power system. The objective is to supply the electric vehicle with high-quality electricity. The hybrid power system comprises a supercapacitor (SC) bank and a lithium-ion battery. The recommended energy management plan attempts to maintain the bus voltage while providing the load demand with high-quality power under various circumstances. The management controller is built on a metaheuristic optimization technique that enhances the flatness theory-based controller’s trajectory generation parameters. The SC units control the DC bus while the battery balances the power on the common line. This study demonstrates the expected contribution using particle swarm optimization and performance are assessed under various optimization parameters, including population size and maximum iterations. Their effects on controller performance are examined in the study. The outcomes demonstrate that the number of iterations significantly influences the algorithm’s ability to determine the best controller parameters. The results imply that combining metaheuristic optimization techniques with flatness theory can enhance power quality. The suggested management algorithm ensures power is shared efficiently, protecting power sources and providing good power quality.