Optimization Algorithm Research Articles

ABSTRACT In order to achieve high-resolution measurement of the thickness of thermally grown oxides (TGO) within thermal barrier coatings(TBCs) using terahertz time-domain spectroscopy even beyond its resolution, we proposed a multimodal features extraction approach and an improved dung beetle optimizer algorithm (Penalty-Theft Dung Beetle Optimizer,PDBO) to improve the prediction accuracy and efficiency of the XGBoost model. Initially, multimodal feature datasets were constructed by extracting time-domain, entropy, frequency-domain, and wavelet features from sample signals, with effective features selected using the Maximal Information Coefficient method. Subsequently, the initial population diversity was enhanced through the importing of a Logistic-Tent chaotic mapping initialization strategy. Additionally, a dynamic convergence factor formula, improved by the inverse tangent function, along with a penalty theft strategy, was proposed to bolster the algorithm's global optimization capability. Finally, we compared the prediction performance of XGBoost optimized by DBO, PDBO, and MIC-PDBO for the TGO thickness through simulation and prepared TBCs. The findings indicated that R2 for the multimodal feature dataset during the simulations exceeded 0.98. Additionally, the MIC-PDBO-XGBoost model exhibited exceptional predictive capability, achieving R2 of 0.99 and 0.93 in the simulations and experimental settings, respectively, achieving super-resolution measurements using terahertz time-domain spectroscopy technology.

ABSTRACT The absence of excitation information and the presence of measurement noise pose significant challenges to accurate damage detection. To address these issues, this study introduces a robust output-only damage identification framework that integrates an enhanced response reconstruction scheme with a collaborative swarm optimisation algorithm (CSOA). In the proposed framework, the available structural responses are partitioned into two subsets. The second set measurements are reconstructed from the first set using the enhanced response reconstruction technique with an improved regularised method. The damage identification task is then formulated as an optimisation-based inverse problem, where the objective function is defined by the discrepancy between the measured and reconstructed responses. Subsequently, CSOA is employed as the search tool to identify local damages by minimising this objective function. The proposed approach is validated through numerical studies on a simply supported truss, a cantilever beam and the Guangzhou New TV Tower, as well as experimental verification on an eight-story steel frame. The effects of measurement noise and damage scenarios are systematically investigated. Results demonstrate that the method can reliably and efficiently identify both the location and severity of single and multiple damages, even with noisy output-only data.

Abstract The results of deformation analysis are crucial for the safety of engineering infrastructure and people’s lives. A key element of this analysis is the identification of stable points of the monitoring network, which will constitute the reference for further calculated displacements. This paper proposes a new method of stability analysis aimed at identifying stable groups of reference points in a displacement monitoring network. It is based on coordinate transformation and involves searching for the optimal set of transformation parameters identified with the optimal point in the transformation parameters space. A hybrid algorithm, which combines two optimization algorithms – Hooke–Jeeves and Simulated Annealing – is used to search for the solution. Two variants of the objective function were tested as the elements of the algorithm. Multiple solutions (groups of congruent reference points) can be detected in case they exist. The simulated 2D network with two congruent point groups was used as an example to illustrate the performance of the proposed algorithm. The proposed hybrid algorithm appeared to overperform the individual Hooke–Jeeves and Simulated Annealing algorithms.

Introduction: Right ventricular failure (RVF) is a major adverse event following left ventricular assist device (LVAD) implantation. The complex mechanisms involved make it challenging to accurately predict RVF. Although supervised machine learning is useful for predicting complex outcomes, it is often difficult to identify specific factors that increase a patient's risk. This study aimed to assess the risk of RVF in individual patients and identify their unique risk factors using supervised machine learning. Methods: Between June 2010 and January 2024, 482 consecutive patients underwent continuous-flow LVAD implantation at Osaka University Hospital or the National Cerebral and Cardiovascular Center. Of them, 326 who underwent preoperative right heart catheterization and echocardiography were included in the analysis. Important features for predicting the risk of RVF were selected using the χ2 or Mann-Whitney U test, the Gini index in a random forest algorithm, and a literature review. The optimal classification algorithm for this analysis was selected from among the random forest, eXtreme Gradient Boosting, support vector machine, logistic regression, and ensemble learning algorithms by comparison of the area under the curve, accuracy, F1 score, and sensitivity through five-fold cross-validation of the test data. The SHapley Additive exPlanations (SHAP) value was used to assess the individual risk factors for RVF. Results: Thirteen important features (sex, age, non-ischemic cardiomyopathy, body surface area, aspartate aminotransferase level, blood urea nitrogen level, left ventricular end-diastolic dimension, left ventricular ejection fraction, right ventricular stroke work index, central venous pressure, pulmonary capillary wedge pressure, pulmonary pulsatility index, and Interagency Registry for Mechanically Assisted Circulatory Support profile) were selected. Ensemble learning was the most reliable classification algorithm. The area under the curve, accuracy, F1 score, and sensitivity were 0.87, 0.89, 0.77, and 0.80, respectively. The SHAP analysis revealed that impaired right ventricular function assessed by right heart catheterization, poor preoperative condition, and a good ejection fraction were associated with an increased risk in most cases. Conclusions: Supervised machine learning enables the accurate prediction of RVF after LVAD implantation, while SHAP values visualize individual risk factors and may optimize preoperative conditions.

Background: Coronary Artery Disease (CAD) is one of the biggest causes of mortality worldwide. Risk stratification for early detection is essential for the primary prevention of CAD. QRISK3 is known to overestimate future CAD risk in some populations, resulting in unnecessary preventive treatment and reduced cost-effectiveness and safety. Combining machine learning model with the metaheuristic optimisation approach using the PSO algorithm may outperform QRISK3 in predicting CAD. It improves performance by selecting the best-performing subset of features related to clinical outcomes. Research Question: Does the performance of Machine Learning Models Combined with PSO Algorithm for Feature Selection as a Metaheuristic Optimisation Approach in Predicting Coronary Arterial Disease using the UK Biobank dataset outperform the QRISK3 calculator? Aims: This study is to assess the accuracy of QRISK3 in predicting CAD using the UK Biobank dataset. It aims to evaluate the efficacy of machine learning models on the identical dataset for predicting CAD. The work utilises the PSO algorithm for feature selection to identify the optimal subset of features from the UK Biobank dataset. Methods: This study utilises data from the UK Biobank. The dataset consists of 348,015 participants aged 24-84 with no prior diagnosis of CAD. The performance of both QRISK3 and machine learning models was evaluated separately using ROC analysis. Several machine learning models were employed: Logistic Regression, Decision Tree, Random Forest, Naïve Bayes, and Gradient Boosting. The dataset was split into training and test sets with a ratio of 4:1 for the machine learning models. Each model has been developed by adding a PSO algorithm to enhance the model's classification accuracy. Results: Out of the total 348,015 participants, 23,136 individuals (6.64%) were diagnosed with CAD within 10 years following their first visit, while 324,879 individuals (93.4%) did not develop CAD. The AUC value of the QRISK3 prediction is 0.6113, while the combined machine learning models using the PSO algorithm using the gradient boosting model achieve an AUC of 0.7258, showing better performance. Conclusions: This study shows hybrid machine learning models optimised with the PSO algorithm can better predict CAD than QRISK3. These ML models can effectively identify high-risk CAD patients, allowing for more personalised preventative strategies and supporting policymakers in implementing lifestyle change recommendations.

Globally, undiagnosed chronic kidney disease (CKD) is a prevalent asymptomatic disease that leads to a significant morbidity and early mortality burden. Current researches have shown that heart issues, identified to as Cardio-Renal Syndrome (CRS) in research, often emerge in patients with renal disease. This disease has the potential to cause sudden cardiac arrest in its later stages. Research on patients with cardio-vascular issues to determine whether their kidneys are affected is valuable, as chronic kidney disease and cardio-vascular disorders are closely linked. Early diagnosis of CKD can enable patients to slow or even reverse disease progression with the help of medicinal interventions. Therefore, in this study we developed an Enhanced Deep Learning (DL)-based technique for the automatic identification of CKD. Digitized electrocardiogram (ECG) data is gathered from Physionet Database's Fantasia (healthy individuals) and PTB (kidney patients). In order to eliminate noise from the ECG measurements, an adaptive median filter is utilized. The significant features are extracted from preprocessed ECG signals. Next, the extracted features are sent into the Enhanced Attention Mechanism with a Long Short-Term Memory (EALSTM) model to classify if a signal is abnormal (CKD) or normal (Non-CKD). To enhance the effectiveness of the EALSTM, its hyper-parameters are optimized using the Adaptive Dingo Optimization (ADO) algorithm. Based to the results of the experiment, the recommended method achieves outstanding 97.65% accuracy, 98.83% precision, 99.21% sensitivity, 98.04% specificity, 99.22% recall, and 99.02% f-measure. The results indicated that the proposed method significantly outperforms other state-of-the-art methods.

The rapid increase of containerized applications in cloud environments has highlighted the critical need for efficient resource management and energy optimization. This article extends our previous work with an aim to enhance the performance of cloud systems. We propose an Extended Directed Container Placement (E_DCP) mechanism, a novel approach designed to enhance container placement efficiency in cloud systems when the number of container increases significantly. Leveraging the Whale Optimization Algorithm (WOA), the mechanism utilizes a scoring mechanism to evaluate various solutions with an aim to identify the best among them. By optimizing multiple objectives, including search time, resource utilization and energy efficiency, the mechanism achieves superior outcomes in heterogeneous and homogeneous cloud infrastructures in comparison to recent methods. The mechanism optimizes this solution to minimize overutilized physical machines. Extensive simulations demonstrate significant improvements in search time and resource utilization with acceptable energy consumption level.

Abstract. Traditional machine defect identification methods have problems such as poor adaptability and low accuracy. As for solving the problems, a large number of improved machine defect identification techniques based on artificial intelligence have been developed; however, all of these methods focus on the neural networks and ignore the connections and influences of the data nodes. This paper proposes a feature-enhanced method based on a data dynamic network for bearing defect identifications. The research preprocesses the data by means of optimization algorithms and defines nodes and edges of data for the construction of a data dynamic network. In order to further improve the accuracy of the machine bearing defect identification, the data are also washed for the energy analysis to be coupled with the data dynamic network for feature enhancement. The training sets and test sets are defined with different data coupling techniques. A performance evaluation of the proposed method is carried out by means of the evaluation function for more effective detection of bearing defects.

Abstract. To address constrained optimization problems in mechanical design, this study proposes an enhanced gray wolf optimization (GWO) algorithm. First, a novel individual memory optimization strategy is developed to expand the population's exploration scope and mitigate the risk of individuals pursuing misguided search trajectories. Second, a position update strategy incorporating differential variation is proposed to balance the local and global search capabilities of individual populations. Lastly, a discrete crossover strategy is proposed to promote information diversity across individual dimensions within the population. By integrating these three improvement strategies with the GWO, a novel improved gray wolf optimization (IGWO) algorithm is developed, which not only preserves robust global and local search capabilities but also demonstrates accelerated convergence performance. To validate the effectiveness, feasibility, and generalizability of the proposed algorithms, three representative mechanical design optimization cases and the Z3 parallel mechanism scale parameter optimization case were employed. Empirical findings reveal that the IGWO algorithm effectively resolves the targeted optimization problem, demonstrating superior performance relative to other benchmark algorithms in comparative analyses.

Abstract Topology optimization using bio-inspired algorithms has gained significant attention in recent decades, particularly for problems involving non-derivative search spaces. This paper introduces a novel bio-inspired algorithm for structural topology optimization, termed the bacterial chemotaxis-based topology optimization algorithm (BCBTOA). Inspired by bacterial chemotaxis, the method models material removal as a guided search process within a two-dimensional continuous domain, with the objective of minimizing structural compliance. To address common numerical challenges such as checkerboarding and mesh dependency, a chemotaxis-based regularization scheme is implemented. The algorithm requires only a single control parameter, R, which governs cavity size and material distribution. Numerical experiments on benchmark problems demonstrate that BCBTOA generates distinct, binary layouts without gray regions and produces mesh-independent solutions for optimal R values. Several communication models between artificial bacteria were investigated, with a modified exponential model yielding the best performance. In addition, different initialization strategies were evaluated, showing that random initialization consistently produced topologies with lower compliance. Overall, the algorithm achieves performance comparable to established methods such as sequential element rejection and admission (SERA) and soft bidirectional evolutionary structural optimization (BESO).

Abstract This paper presents the optimization of hybrid energy storage system for an electric vehicle, by using particle swarm optimization and genetic algorithm techniques including the design of conventional controller. It utilizes the steady-state filtered power as the reference output power of the battery. To regulate the steady-state current output of the battery, the output power of the ultracapacitor is adjusted dynamically with help of a proportional-integral-derivative controller, such that the power difference controlled structure is obtained. The PID controller parameters are optimized through the particle swarm optimization algorithm, and genetic algorithm. The output is compared with the federal test procedure drive cycle, and the optimized power output and state of charge of the battery is obtained. In this paper the result obtained shows that the SoC of battery is increased by 30.12% with the help of PSO-PID and 0.1% with the help of GA-PID and almost similar results were obtained using conventional PID controller. The battery power ratio in the total power demand is 0.6667 with PSO-PID, 0.8871 with GA-PID and 0.9109 with conventional PID controller. The results obtained shows that the proposed control strategy PSO-PID is capable of eliminating the deviation in power quickly and accordingly achieving the best suited global optimization of EV. In comparison with the optimized PID strategy the result obtained by the proposed strategy shows improvement in the energy consumption and battery life of EV.

The present study focuses on wind power forecasting (WPF) models based on deep neural networks (DNNs), aiming to evaluate current approaches, identify gaps, and provide insights into their importance for the integration of Renewable Energy Sources (RESs). The systematic review was conducted following the methodology of Kitchenham and Charters, including peer-reviewed articles from 2020 to 2024 that focused on WPF using deep learning (DL) techniques. Searches were conducted in the ACM Digital Library, IEEE Xplore, ScienceDirect, Springer Link, and Wiley Online Library, with the last search updated in April 2024. After the first phase of screening and then filtering using inclusion and exclusion criteria, risk of bias was assessed using a Likert-scale evaluation of methodological quality, validity, and reporting. Data extraction was performed for 120 studies. The synthesis established that the state of the art is dominated by hybrid architectures (e.g., CNN-LSTM) integrated with signal decomposition techniques like VMD and optimization algorithms such as GWO and PSO, demonstrating high predictive accuracy for short-term horizons. Despite these advancements, limitations include the variability in datasets, the heterogeneity of model architectures, and a lack of standardization in performance metrics, which complicate direct comparisons across studies. Overall, WPF models based on DNNs demonstrate substantial promise for renewable energy integration, though future work should prioritize standardization and reproducibility. This review received no external funding and was not prospectively registered.

ABSTRACT Reusing recycled aggregates from building and demolition wastes to substitute natural aggregates ( NA ) in concrete has been proposed as a way to efficiently manage waste and sustainably fulfil the growing demand for aggregates. While these issues may be mitigated by strengthening RAC with fiber-reinforced polymers ( FRP ), it may be challenging to predict the stress-strain behaviors of FRP -confined RAC (abbreviated as FRPC ( RAC )). Characterizing the characteristics of will need high-performing algorithms as additional experimental information becomes available. Boosting algorithms named categorical boosting, histogram boosting, and adaptive boosting are used in this study to forecast the stress-strain characteristics of FRPC ( RAC ) named ultimate strain ( ε cu ) and ultimate strength ( f cu ). The performance of the mentioned boosting methods is deeply affected by their hyperparameters that can be tuned by optimization algorithms, where a newly developed optimization algorithm is considered for this purpose called the pufferfish algorithm ( PA ). The Shapley Additive explanation ( SHAP ) was used to identify the crucial parameters in the models. 240 experimental specimens' worth of data were collected. In order to prepare the dataset to introduce to the models, several pre-processing steps are accomplished, such as Kolmogorov-Smirnov () test, evaluation, and feature importance. Different aspects of analysis depicted that the order of workability for developed PA -based models from best to weakest is in CaB − PA > AdB − PA > HiGB − PA , where the outperformed algorithm could gain the determination coefficient at 0.9948 and 0.9932 in the training and testing portions of the dataset.

Purpose This study systematically evaluated the intellectual progress in artificial intelligence (AI)-driven osteoporosis research between 2004 and 2024 by employing scientometric and visualization techniques. Through mapping knowledge domains and identifying emerging trends, it offered actionable recommendations and strategic insights to guide future scholarly endeavors. Methods We queried the Web of Science Core Collection for English-language articles and reviews published between January 1, 2004, and November 30, 2024, using search terms including “osteoporosis,” “deep learning,” “convolutional neural networks,” and “artificial intelligence.” Bibliometric data were processed via VOSviewer (v1.6.20), CiteSpace (v6.3.R1), Scimago Graphica (v 1.0.46), and the R package Bibliometrix to quantify annual publication trends, assess national/institutional contributions, evaluate journal/author impact metrics, and map keyword co-occurrence and burst dynamics. Results The bibliometric analysis identified 408 publications (343 articles, 65 reviews) from 2004 to 2024, with a marked increase in output observed in 2024. China and the United States dominated scholarly productivity and citation impact. Leading institutions included the Technical University of Munich and Seoul National University, while Osteoporosis International emerged as the most influential journal. Prolific authors such as Thomas Baum demonstrated significant academic leadership. Keyword co-occurrence analysis revealed deep learning, artificial intelligence, and diagnosis as core research frontiers, signaling future technological and clinical priorities. Conclusion This study represents the first comprehensive bibliometric analysis of research on artificial intelligence in the field of osteoporosis. It not only outlines the field’s development trajectory and emerging frontiers but also highlights the research focus on AI technologies, particularly deep learning. Furthermore, it emphasizes critical challenges in clinical translation, such as algorithm optimization, model interpretability, and ethical privacy concerns. By systematically identifying key contributors, collaborative networks, and evolving research fronts, this study provides a foundational roadmap for the field. It offers strategic priorities for researchers to address methodological gaps, serves as a reference for clinicians to understand the evolving technological toolkit, and provides a basis for policymakers to promote interdisciplinary collaboration.

Facing the urgent need for personalized and real-time interaction in music education, this study constructed an artificial intelligence-assisted interactive platform that offers targeted support in pitch recognition, rhythm training, and classroom feedback. Through function planning and algorithm optimization, it enables diversified student-teacher communication while leveraging artificial intelligence to capture and analyze teaching data, helping teachers identify students' weaknesses in playing skills or music understanding for strategic teaching adjustments. The platform also overcomes compatibility issues across various hardware and user needs via inclusive design and management, provides precise counseling for individual learning difficulties, and offers real-time monitoring with personalized feedback to deliver technical guidance and emotional support. This research comprehensively explored the platform's requirements, design, evaluation, and user feedback, aiming to inspire the music education field and inform future studies.

Abstract With the continuous expansion of wind and solar complementary power generation systems, introducing energy storage systems to ensure their stability has become crucial. To solve the high cost in current methods, a wind-solar hybrid energy storage model is established, and a grey wolf pigeon swarm optimization algorithm for capacity optimization is constructed. The performance is compared with other comparison algorithms. The accuracy and recall of the designed method were 98.67% and 97.74%. Moreover, the F1 value, precision, loss value and AUC value of the AOC curve were 97.68%, 96.88%, 0.34, and 0.974, respectively, all of which were superior to the comparison algorithms. Subsequently, to verify the effectiveness, it was compared with other algorithms. The cost for optimizing the capacity in area A was 155,534 RMB, which was below the comparison algorithm. The algorithm has good performance and excellent capacity optimization effect. It helps to reduce the capacity optimization cost and provides theoretical basis for research on capacity optimization.

Abstract Accurate assessment of the State of Health (SOH) for Lithium-ion Batteries (LIBs) is critical for ensuring the operational safety of battery management systems. To overcome the accuracy limitations of existing SOH prediction methods, this study proposes an integrated framework that synergizes feature engineering, an Improved Beluga Whale Optimization (IBWO) algorithm, and Gaussian Process Regression (GPR). The methodology involves several key stages. First, Health Features (HFs) are extracted from raw charge/discharge data using a Variational Autoencoder (VAE). Optimal HFs are then selected through Pearson correlation analysis and augmented using a Generative Adversarial Network (GAN). Subsequently, the Variational Mode Decomposition (VMD) algorithm is applied to adaptively decompose the HFs. The decomposed components are predicted and reconstructed using GPR models, effectively suppressing noise within the HFs. Following this, GPR models establish a nonlinear mapping between the reconstructed HFs and the actual SOH to enable prediction. Meanwhile，to address the challenge of GPR hyperparameter selection, an IBWO algorithm is proposed. The IBWO incorporates Circle chaotic mapping initialization, quasi-opposition-based learning, and a dynamic weight adjustment strategy to improve optimization performance. Experimental validation on the NASA and MIT public battery dataset demonstrates the framework's superior performance, with prediction error indicators consistently below 1%, confirming high accuracy and robust generalization capability.

Gas foil journal bearings (GFJBs) are critical components in high-speed turbomachinery due to their advantages such as increased load capacity, reduced friction, and improved rotordynamic stability. This study presents a comprehensive numerical analysis comparing the static and dynamic performance of herringbone groove textures—namely single, double, and multi-structured patterns—applied to the top foil of GFJBs. The pressure distribution is obtained by solving the Reynolds equation using the finite difference method, and key performance metrics such as load-carrying capacity, frictional torque, power loss, effective stiffness, damping, and system stability are systematically evaluated. Among the patterns, the multi-herringbone texture (Texture 3) consistently outperformed other configurations. To further enhance performance, a multi-objective Particle Swarm Optimization (PSO) algorithm was employed to optimize key groove parameters. The optimized configuration yielded improved static and dynamic behavior, including higher stiffness and damping values and enhanced system stability. Notably, the optimized textured GFJB demonstrated a 26.9% increase in load-carrying capacity (LCC) compared to the plain GFJB. This research highlights the effectiveness of surface textures in enhancing GFJB performance and demonstrates the utility of PSO as a powerful tool for multi-objective optimization in high-speed bearing design.

ABSTRACT This paper presents a Deep Learning-Based Pair Barracuda Swarm Optimization for an Arabic Text-to-Speech Synthesizer Using Applied Linguistics (DLPBSO-ATTSSAP), designed to support visually impaired individuals. Arabic text-to-speech synthesis is challenging due to linguistic complexity and contextual ambiguity. The proposed system begins with multi-level preprocessing to normalize Arabic text, followed by FastText embeddings to capture semantic and syntactic nuances. A Convolutional Variational Autoencoder (CVAE) is employed to learn latent features for accurate text classification, with hyperparameters optimized through the Pair Barracuda Swarm Optimization (PBSO) algorithm. This enhances classification accuracy and system performance. Finally, the WaveNet model converts processed text into natural, human-like speech. Experimental results show that DLPBSO-ATTSSAP outperforms existing methods across key metrics. By integrating deep learning with swarm optimization, the system provides a user-friendly, efficient, and high-quality speech synthesis solution. This work highlights the model’s ability to address language-specific challenges in Arabic and contribute to accessible communication technologies for underrepresented languages.

This study deals with the finite element method-based simulations of a textured surface, two-lobe journal bearing operating with a magnetorheological (MR) lubricant—a smart fluid composed of fine magnetic particles suspended in a carrier fluid. The flow of the MR lubricant within the textured bearing, which incorporates rectangular micro-grooves along axial directions, is modeled using Reynolds’ equation. The Bingham plastic model defines the MR fluid's viscosity, accounting for the influence of yield stress, magnetic field, and shear-strain rate. The Reynolds equation is solved using the finite element method to determine film pressure, film direct stiffness coefficients, threshold speed, journal trajectories and limit cycles. A Multi-Objective Genetic Algorithm (MOGA) is used to optimize rectangular micro-groove attributes (i.e., number, width, depth and length) to maximize the direct film stiffness coefficients and threshold speed of stability. A robust design recommendation is proposed based on Fuzzy-based Multi-Objective Genetic Algorithm (Fuzzy-MOGA) optimization, targeting improved film stiffness coefficients and threshold speed. The results exhibit a stable and resilient optimization landscape, with minimal sensitivity to input variations. The numerical results demonstrate that the combined use of MR lubricant, two-lobe geometry and surface texturing enhances stiffness coefficients (116.7%∼850.8%) and threshold speed (81.3%). The two-lobe bearing with partial surface texturing in the leading half demonstrates the highest dynamic stability, as evidenced by threshold speed, minimal journal center trajectories and compact limit cycles.