Article
Physics, Mathematical
Jiuyang Liang, Pei Liu, Zhenli Xu
Summary: This paper presents a fast and highly accurate algorithm for solving Poisson's equations in a cuboidal domain with mixed boundary conditions. The algorithm utilizes numerical quadratures and the fast multipole algorithm to achieve high convergence rates, even for piecewise continuous source terms, by representing the solution as a summation of point sources in free space.
COMMUNICATIONS IN COMPUTATIONAL PHYSICS
(2021)
Article
Chemistry, Physical
Safiyye Kavak, Ajinkya Anil Kadu, Nathalie Claes, Ana Sanchez-Iglesias, Luis M. Liz-Marzan, Kees Joost Batenburg, Sara Bals
Summary: In this study, a method for automating the volumetric segmentation process applied to three-dimensional reconstructions of nanoparticle assemblies is proposed. The performance of different automated segmentation methods in accurately extracting quantitative descriptors for morphological characterization is evaluated. The results show that the spherical Hough transform exhibits superior performance in accurately extracting quantitative descriptors, allowing for an objective, efficient, and reliable volumetric segmentation of complex nanoparticle assemblies.
JOURNAL OF PHYSICAL CHEMISTRY C
(2023)
Article
Mathematics, Applied
Yiying Fang, Ying Jiang
Summary: In this paper, a fully discrete fast multiscale Galerkin method is developed to solve the boundary integral equation derived from the interior Dirichlet problem in a domain with corners. The integral operator in the equation is split into two operators: a noncompact operator with a singular kernel and a compact operator with a piecewise smooth kernel. The paper presents two fast schemes to evaluate the entries of representation matrices for the noncompact and compact operators.
JOURNAL OF SCIENTIFIC COMPUTING
(2023)
Article
Multidisciplinary Sciences
Shangting You, Yi Xiang, Henry H. Hwang, David B. Berry, Wisarut Kiratitanaporn, Jiaao Guan, Emmie Yao, Min Tang, Zheng Zhong, Xinyue Ma, Daniel Wangpraseurt, Yazhi Sun, Ting-yu Lu, Shaochen Chen
Summary: Three-dimensional bioprinting techniques are widely used for fabricating 3D-engineered tissues, but face challenges in maintaining high cell density, cell viability, and fabrication resolution simultaneously. A novel approach using iodixanol in bioink was developed to reduce light scattering and improve fabrication resolution. Tissues with high cell density and fine vascular networks were successfully fabricated, and showed viability and endothelialization after 14 days of culture in a perfusion system.
Article
Computer Science, Artificial Intelligence
Shifeng Li, Yan Cheng, Yunfeng Liu, Yuqiang Yang
Summary: This paper proposes a method to detect abnormal events from videos using integral image and Bayesian framework. By treating regular cubes in the videos as events and calculating motion histograms, the method can effectively detect abnormal events in complex scenes.
NEURAL PROCESSING LETTERS
(2022)
Article
Geochemistry & Geophysics
Wenwu Tang, Jianxin Liu, Changying Ma, Feng Zhou
Summary: In this study, a finite-element (FE) forward modeling algorithm for controlled-source electromagnetic (CSEM) with vector and scalar potentials was developed. A new preconditioner constructed from an incomplete Cholesky decomposition with zero fill-ins was introduced to mitigate the null space and nonuniqueness of the potentials caused by the curl operator. The new iterative solver based on preconditioned quasiminimal residual method showed improved efficiency and eliminated the unstable and nonuniqueness issues of the potentials.
Article
Computer Science, Information Systems
Chunfeng Chen, Changyu Hu, Jianjiang Zhou
Summary: The existing methods for calculating electromagnetic scattering of power lines do not consider their periodicity. In this study, we propose a fast electromagnetic scattering calculation method that combines the integral equation discontinuous Galerkin (IEDG) method and the characteristic modes-Sherman-Morrison-Woodbury algorithm (CM-SMWA) for power lines with stranded structure. By discretizing the electric field integral equation (EFIE) using the IEDG method, we can handle non-conformal grids and increase the flexibility of the CM-SMWA. Taking advantage of the periodic property of power lines, we can significantly reduce the calculation time compared to conventional methods.
Article
Computer Science, Theory & Methods
Michael P. Lingg, Stephen M. Hughey, Balasubramaniam Shanker, Hasan Metin Aktulga
Summary: Evaluation of pair potentials is crucial in physics, especially for Laplace potential and Helmholtz potential. The fast multipole method is a highly scalable algorithm for parallel evaluation of Laplace potential. However, the evaluation of Helmholtz potential faces bottlenecks in computation and communication costs.
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
(2022)
Article
Geochemistry & Geophysics
Xulong Wang, Jianxin Liu, Jian Li
Summary: In order to address the problem of efficient and accurate 3-D modeling of magnetic anomaly, a fast numerical modeling method based on integral equation is developed. It achieves high accuracy and reduces computation time and memory requirement by using a new analytical expression for prismatic magnetic anomaly and taking advantage of the Toeplitz property of the kernel matrix. The applicability of the algorithm is demonstrated by studying the terrain effect on airborne magnetic data in a real complex topography model.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Review
Computer Science, Information Systems
Ming Jiang, Yin Li, Lin Lei, Jun Hu
Summary: This paper reviews a series of fast direct solution methods for electromagnetic scattering analysis and introduces their advantages, applications, and research trends. Three different main methods are discussed, and numerical examples are presented to demonstrate their efficiency and accuracy. Finally, the main challenges and possible strategies of fast direct solution methods are briefly discussed.
Article
Engineering, Electrical & Electronic
Lifeng Wu, Yanwen Zhao, Qiangming Cai, Zhipeng Zhang, Jun Hu
Summary: The method utilizes a high-order vector base to reduce unknowns and combines RBM to avoid repetitive computations, thereby improving computational efficiency and reducing memory costs. Interpolation accuracy, memory cost, and efficiency are demonstrated through two numerical examples.
IEEE ANTENNAS AND WIRELESS PROPAGATION LETTERS
(2021)
Article
Geochemistry & Geophysics
Xingguo Huang, Yong Hu
Summary: The study demonstrates how a Bayesian framework can be utilized to estimate the uncertainty of velocity reconstruction in the context of nonlinear inverse scattering imaging. It relies on Kalman Filter method and direct envelope methods to overcome the challenges of inverse scattering problems. The results highlight the important role played by uncertainty quantification in dealing with the uncertainty of nonlinear inverse scattering problems in various media.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2021)
Article
Engineering, Electrical & Electronic
Yiling Wang, Zaiping Nie
Summary: An efficient thin dielectric sheet (TDS) model is proposed for computing radome targets made from high contrast sandwich structure material. By utilizing dimensional analysis, the intrinsic stratified distribution of diagonal entries in the TDS model has been identified for the first time. Two scaling schemes are introduced to improve the conditioning of the system matrix. This method significantly enhances the computational efficiency of radome targets, facilitating fast simulation in practical engineering applications.
IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION
(2021)
Article
Computer Science, Interdisciplinary Applications
Baoli Yin, Jinfeng Wang, Yang Liu, Hong Li
Summary: This paper analyzes a novel structure preserving difference scheme for the high dimensional nonlinear space-fractional Schrodinger equation, with rigorously proved energy and mass preserving properties. By introducing the fast PF-BiCG algorithm for complex systems, the computing cost is effectively reduced to O(N log N) and memory requirement to O(N), confirmed by numerical experiments.
JOURNAL OF COMPUTATIONAL PHYSICS
(2021)
Article
Mathematics, Applied
Sasan Mohyaddin, Johannes Tausch
Summary: This paper proposes three algorithms for evaluating volume potentials in boundary element methods. The approach is to use a modified fast multipole method for a boundary concentrated volume mesh. By discretizing the volume with nearly O(h-2) degrees of freedom, the algorithm can compute potentials with nearly O(h-2) complexity. The complexity of volume potential calculations has the same asymptotic order as boundary potentials. The proposed algorithms demonstrate accuracy and effectiveness in computing potentials for the Poisson equation in three dimensions.
SIAM JOURNAL ON SCIENTIFIC COMPUTING
(2023)