Archives of Computational Methods in Engineering最新文献

Reduced-order models (ROMs) offer compact representations of complex engineering systems governed by partial differential equations or high-dimensional ordinary differential equations enabling efficient simulations of otherwise computationally intensive problems. These models are typically constructed by projecting the high-dimensional governing equations onto reduced subspaces derived using techniques such as Singular Value Decomposition (SVD) or Proper Orthogonal Decomposition (POD). However, conventional ROMs struggle with nonlinear systems due to the high computational cost of repeatedly accessing high-dimensional solution spaces for nonlinear term evaluations. Hyper-reduction methods address this challenge by efficiently approximating nonlinear term evaluations, significantly improving ROM performance. They are also essential for solving large parametric linear problems that lack an efficient parameter-affine decomposition. This paper provides a comprehensive overview of hyper-reduction algorithms, emphasizing both their theoretical foundations and practical implementations in academic research and industry. With the rapid advancement of data-driven methods, reduced-order modeling has become indispensable for analyzing and simulating large-scale systems, including fluid dynamics, thermal processes, and structural mechanics. As the demand for efficient computational tools in science and engineering continues to grow, a detailed discussion of hyper-reduction techniques is both timely and valuable. The paper explores state-of-the-art hyper-reduction techniques, including discrete empirical interpolation methods (DEIM), energy-conserving sampling and weighting (ECSW), and emerging machine learning-based approaches. A nonlinear parametric heat conduction example is presented to illustrate the implementation of these methods. The analysis evaluates their strengths and weaknesses using standard metrics, providing insights into their practical utility. Finally, the paper concludes by discussing future research directions and potential applications of hyper-reduction, including its integration with real-time simulations and digital twin systems.

In silico models and computational tools are invaluable instruments that complement experiments to improve our understanding of complex phenomena such as cancer evolution. This work offers a perspective on different approaches that can be used for mathematical modeling of glioblastoma, the most common and lethal brain cancer, in microfluidic devices, the most biomimetic in vitro cell culture technique nowadays. These approaches range from purely knowledge-based solutions to data-driven, and hence completely model-free, algorithms. In particular, we focus on hybrid approaches, which combine physically-based and data-driven strategies, demonstrating how this integration can enhance the understanding we get from simulation by revealing the underlying model structure and thus, in turn, the prospective biological mechanism.

The exact formulation of the subloading surface model is provided in this article for the description of the elastoplastic and elasto–viscoplastic deformations not only for the monotonic but also the cyclic loading processes at general deformation rate ranging from the quasi-static to the impact loading in a unified manner by the subloading-overstress model. Here, it is noteworthy that even the elastoplastic deformation can be described more exactly by the present elasto–viscoplastic constitutive equation, noting that the elastoplastic constitutive equation is limited to the description of the quasi-static deformation behavior, but the purely quasi-static deformation does not exist actually. Therefore, the elastoplastic constitutive equation can be disused only by using the subloading-overstress model. It will be extended to describe the temperature-dependence for metals, since the elasto–viscoplastic deformation behavior is influenced by the temperature in general. Then, the validity of the extended subloading-overstress model for the prediction of the temperature-dependent elasto–viscoplastic deformation of metals will be verified by the comparisons with the test data of metals for various isothermal and/or non-isothermal deformations in the monotonic and the cyclic loading processes.

This article systematically analyzes the evolution of AI-based traffic optimization techniques from 2019 to 2024, addressing the critical challenge of urban mobility in increasingly congested cities. While traditional traffic management methods have relied on fixed systems and basic machine learning, recent years have seen a significant shift toward advanced AI solutions including deep neural networks, generative adversarial networks (GANs), and hybrid optimization models. Through a structured five-stage methodology examining 46 research papers, this study evaluates various approaches based on accuracy, efficiency, and real-world applicability. The findings show Recurrent Neural Networks achieving 95% accuracy in traffic pattern classification and GANs reaching 98% accuracy in traffic density recognition. Hybrid models combining neural networks with optimization algorithms have demonstrated exceptional adaptability, achieving R² values of 0.999 in traffic flow prediction. Implementation of graph-based frameworks and integration of multi-modal data sources improved prediction accuracy and reduced travel times. Advancements in reinforcement and transfer learning enhance scalability, positioning AI-powered systems as key drivers of efficient and sustainable urban mobility.

Active contour and active surface models are image segmentation methods which offer a solid mathematical background, reduced computational time, smooth boundaries and, in many cases, also robustness in presence of noise. In other cases, due to the complexity of the images, active contour-surface models do not provide good results. However, their performance can be improved by taking into account more strategic image features that affect the evolution of the active contours-surfaces. This review seeks to explore the features used in literature for this goal, the related topic of feature reduction/selection, and the type of images involved. Considerations about limitations and possible future extensions are also presented.

Coronary artery disease is one of the main cardiovascular illnesses impacting the whole human population. It has been established that this illness is the main cause of mortality in both developed and developing nations. Chest discomfort and a reduction in blood flow to the heart are symptoms of this disease, which is brought on by plaque buildup in the blood arteries. In the past two decades, the domains of artificial intelligence (AI) like machine learning (ML) and deep learning (DL) have opened up new directions in the field of cardiovascular medicine. These methods have swiftly widened its spheres in medicine, from the automatic interpretation of cardiac rhythm abnormalities to aiding in complicated decision-making, and it has shown to be a promising tool for supporting clinicians in making treatment decisions. This study presents several clinical facets of coronary artery disorders, including risk factors, illness diagnostics, and therapeutic approaches. Additionally, the study discusses current developments and noteworthy advancements in AI-based computer-aided diagnosis (CAD) of coronary artery disease. Various key and novel insights and challenges in using CAD for cardiovascular disease have also been discussed.

Leveraging the concepts of the traditional Runge–Kutta method, the Runge–Kutta Optimizer (RUN) is a new meta-heuristic algorithm created for global optimization applications. The RUN method is addressed in this paper along with its mechanisms to optimize search strategies and improve the quality of solutions. The RUN algorithm is effective in solving complex, nonlinear optimization problems because it efficiently balances exploration and exploitation using a combination of random elements and deterministic rules. Variations of the Runge–Kutta algorithm are presented, with an emphasis on modifications that improve the performance of the method on a range of problem sets. By examining a variety of fields, the study highlights the potential application of the algorithm in fields such as engineering, computer science and medicine. A comprehensive analysis of the algorithm methodology and in-depth evaluations of the 2011 CEC benchmark functions provide empirical evidence of the algorithm’s effectiveness and efficiency compared to traditional optimization techniques, as well as its superior performance over a number of state-of-the-art techniques. Convergence analysis shows that RUN leads to faster convergence rates and consistently finds optimal or near-optimal solutions. In addition, a set of real-world engineering challenges, such as design optimization and parameter estimates, are used to test the suitability of the algorithm. With advancement in computing speed and solution accuracy, the effectiveness of the proposed RUN algorithm makes it a proposed methodology for a wide range of optimization problems. Finally, some future directions on potential research plans are included in the paper.

Convolutional neural networks have become essential in computer vision, especially for image classification. They depend heavily on hyperparameters, and there is no practical way to manually tune these numerous settings through trial and error. This made it necessary for automated methods, especially those that come with metaheuristic algorithms, to optimize the hyperparameters and build good network architectures. Metaheuristic algorithms provide an easy way of determining the best hyperparameters by generating and testing various combinations using intuitive strategies and principles of solution-finding. This review provides a comprehensive discussion of convolutional neural networks, such as their layers, architectural designs, types, and ways of improvement, with a focus on optimization using metaheuristic algorithms. It highlights prominent algorithms and recent studies aimed at improving hyperparameter selection. By combining results of current and future research, this review should be a helpful resource for researchers, serving as the basis for further research and innovation in automated hyperparameter optimization using metaheuristic approaches, contributing significantly to further development in this field. The study concludes that metaheuristic algorithms significantly enhance the performance of convolutional neural networks with a simple yet effective replacement for manual tuning and high future prospects for automated optimization breakthroughs.