Search

In this article, a local coupling multitrace domain decomposition method (LCMT-DDM) based on surface integral equation (SIE) formulations is proposed to analyze electromagnetic scattering from multilayered dielectric objects. Different from the traditional SIE-DDM, where the interactions between subdomains are accounted for using global radiation coupling, LCMT-DDM uses a local coupling scheme. The original multilayered object is decomposed into several independent domains, i.e., the exterior region (free space) and many homogeneous interior regions (dielectrics). The boundaries of subdomains are all touching faces, where only the Robin transmission conditions (RTCs) are enforced to ensure the field continuity. Hence, each subdomain only couples with its neighboring regions, which makes the DDM system a highly sparse matrix, especially when the number of subdomains is large. In each subdomain, the electric-field integral equation (EFIE) and the magnetic-field integral equation (MFIE) for dielectrics are used as the governing equations. By imposing RTCs, well-conditioned equations are formed in each subdomain without invoking the combined-field integral equation (CFIE), which usually causes accuracy issues in dielectric modeling. Since the subdomain matrices are diagonally dominant, the flexible generalized minimal residual (FGMRES) technique is used to accelerate the iterative solution of the whole DDM system. Moreover, an effective preconditioner that can be recursively constructed is proposed. [ABSTRACT FROM AUTHOR]

In this paper, we present a new multiscale d omain decomposition algorithm for computing solutions of static Eikonal equations. In our new method, the decomposition of the domain does not depend on the slowness function in the Eikonal equati on or the boundary conditions. The novelty of our new method is a coupling of coarse grid and fine grid solvers to propagate information along the characteristics of the equation efficiently. The method involves an iterative parareal-like update scheme in order to stabilize the method and speed up convergence. One can view the new method as a general framework where an effective coarse grid s olver is computed "on the fly" from coarse and fine grid solutions that are computed in previous iterations. We study the optimal weights used to define the effective coarse grid solver and the stable update scheme via a model problem. To demonstrate the framework, we develop a specific scheme using Cartesian grids and the fast sweeping method for solving Eikonal equations. Numerical examples are given to show the method's effectiveness on Eikonal equations involving a variety of multiscale slowness functions. [ABSTRACT FROM AUTHOR]

In this paper we study convergence estimates for a multigrid algorithm with smoothers of successive subspace correction (SSC) type, applied to symmetric elliptic PDEs under no regularity assumptions on the solution of the problem. The proposed analysis provides three main contributions to the existing theory. The first novel contribution of this study is a convergence bound that depends on the number of multigrid smoothing iterations. This result is obtained under no regularity assumptions on the solution of the problem. A similar result has been shown in the literature for the cases of full regularity and partial regularity assumptions. Second, our theory applies to local refinement applications with arbitrary level hanging nodes. More specifically, for the smoothing algorithm we provide subspace decompositions that are suitable for applications where the multigrid spaces are defined on finite element grids with arbitrary level hanging nodes. Third, global smoothing is employed on the entire multigrid space with hanging nodes. When hanging nodes are present, existing multigrid strategies advise to carry out the smoothing procedure only on a subspace of the multigrid space that does not contain hanging nodes. However, with such an approach, if the number of smoothing iterations is increased, convergence can improve only up to a saturation value. Global smoothing guarantees an arbitrary improvement in the convergence when the number of smoothing iterations is increased. Numerical results are also included to support our theoretical findings. [ABSTRACT FROM AUTHOR]

Summary: A multiscale extension for a parallel noninvasive mixed domain decomposition method is presented. After briefly exposing our noninvasive implementation of the Latin method, we present how the scalability of the algorithm is obtained by the partial verification of the constitutive law of the interfaces. We propose a new interpretation of the classical macrostrategy that imposes the overall balance of interface's forces. We also propose a new study of the macrostrategy that consists in enforcing the coarse continuity of the interfaces' displacement. [ABSTRACT FROM AUTHOR]

Based on overlapping domain decomposition, a new class of parallel split least‐squares (PSLS) mixed finite element methods is presented for solving parabolic problem. The algorithm is fully parallel. In the overlapping domains, the partition of unity is applied to distribute the corrections reasonably, which makes that the new method only needs one or two iteration steps to reach given accuracy at each time step while the classical Schwarz alternating methods need many iteration steps. The dependence of the convergence rate on the spacial mesh size, time increment, iteration times, and subdomains overlapping degree is analyzed. Some numerical results are reported to confirm the theoretical analysis. [ABSTRACT FROM AUTHOR]

Summary: An efficient method for solving large nonlinear problems combines Newton solvers and domain decomposition methods. In the domain decomposition method framework, the boundary conditions can be chosen to be primal, dual, or mixed. The mixed approach presents the advantage to be eligible for the research of an optimal interface parameter (often called impedance), which can increase the convergence rate. The optimal value for this parameter is usually too expensive to be computed exactly in practice: An approximate version has to be sought, along with a compromise between efficiency and computational cost. In the context of parallel algorithms for solving nonlinear structural mechanical problems, we propose a new heuristic for the impedance, which combines short‐ and long‐range effects at a low computational cost. [ABSTRACT FROM AUTHOR]

We introduce domain decomposition methods of Schwarz waveform relaxation (WR) type for fractional diffusion-wave equations. We show that the Dirichlet transmission conditions among the subdomains lead to slow convergence. So, we construct optimal transmission conditions at the artificial interfaces and we prove that optimal Schwarz WR methods on N subdomains converge in N iterations both on infinite spatial domains and on finite spatial domains. We also propose optimal transmission conditions when the original problem is spatially discretized and we prove the same result found in the continuous case. [ABSTRACT FROM AUTHOR]

The choice of the transmission conditions (TCs) across the interfaces between subdomains and its free parameters has a significant impact on the convergence of domain decomposition methods (DDM), especially when applied to propagation and scattering problems. In this paper, we study the convergence behavior of the optimized Schwarz DDM applied to electromagnetic scattering problem in two dimensions for different TCs based on a polynomial approximation of the Dirichlet-to-Neumann map on the Fourier domain. We then propose a methodology to obtain the free parameters of the TC based on the numerical optimization of the spectral radius of the Schwarz iteration matrix on the Fourier domain, taking into account the number of subdomains and some geometrical characteristics such as length and overlap size. To confirm the improvements in convergence, we perform numerical experiments and compare the iteration count for different TCs. [ABSTRACT FROM AUTHOR]

A mixed parallel strategy for the solution of homogenization-based multi-scale constitutive problems undergoing finite strains is proposed. The approach aims to reduce the computational time and memory requirements of non-linear coupled simulations that use finite element discretization at both scales (FE2). In the first level of the algorithm, a non-conforming domain decomposition technique, based on the FETI method combined with a mortar discretization at the interface of macroscopic subdomains, is employed. A master-slave scheme, which distributes tasks by macroscopic element and adopts dynamic scheduling, is then used for each macroscopic subdomain composing the second level of the algorithm. This strategy allows the parallelization of FE2 simulations in computers with either shared memory or distributed memory architectures. The proposed strategy preserves the quadratic rates of asymptotic convergence that characterize the Newton-Raphson scheme. Several examples are presented to demonstrate the robustness and efficiency of the proposed parallel strategy. [ABSTRACT FROM AUTHOR]

A hybrid coupled strategy was proposed for a coupled neutronics and thermal-hydraulics analysis of a full PWR core, using the continuous-energy Reactor Monte Carlo code (RMC) and the sub-channel code COBRA-TF (CTF). In the code system with hybrid coupling, CTF was invoked and controlled by RMC internally, without external interface. The On-The-Fly cross sections treatment of RMC was used to reduce the complexity of the coupled code as well as to reduce the memory requirement. The domain decomposition parallel technique was developed in CTF to improve the efficiency of full core sub-channel calculations, and the PWR preprocessor of CTF can reduce the complexity of full core modeling establishment. The coupled codes were applied to steady-state simulations of the Benchmark for Evaluation And Validation of Reactor Simulations (BEAVRS) in the hot, full power condition at the full-core level to reveal the effects of the coupling on the full-core power distribution in both axial and radial directions. The influences of other important parameters including the neutron population and the boron concentration were also investigated. The results proved the effectiveness and high fidelity of the coupled system. More systematic and detailed analyses can be performed based on realistic operating conditions of PWR full core with the coupled codes system. [ABSTRACT FROM AUTHOR]

We introduce a novel substructuring approach for solving the incompressible Stokes equations for the case of enclosed flows. We employ a simple distribution of the global pressure constraint to subdomains which allows for a natural decomposit ion into Stokes subdomain problems with Dirichlet data which are well-posed and inf-sup stable. This approach yields a saddle-point problem on the interface r involving an operator which is conti nuous and coercive on H1/2 (Γ) and which is restricted to the interface trace space of functions satisfying the incompressibility constraint. We derive the form of the constraints explicitly, both for the cont i nuous and for the discrete case. This allows us to design directly a class of interface preconditioners of constraint type, thus avoiding the need to formulate a coarse level problem. Our analysis indicates that the resulting solution method has performance independent of the mesh-size, while numerical results point to a mild dependence on the number of subdomains. We illustrate the technique on s ome standard test problems and for a range of domains, meshes, and decompositions. [ABSTRACT FROM AUTHOR]

We present an algorithmically efficient and parallelized domain decomposition based approach to solving Poisson’s equation on irregular domains. Our technique employs the Schur complement method, which permits a high degree of parallel efficiency on multicore systems. We create a novel Schur complement preconditioner which achieves faster convergence, and requires less computation time and memory. This domain decomposition method allows us to apply different linear solvers for different regions of the flow. Subdomains with regular boundaries can be solved with an FFT-based Fast Poisson Solver. We can solve systems with 1,0243 degrees of freedom, and demonstrate its use for the pressure projection step of incompressible liquid and gas simulations. The results demonstrate considerable speedup over preconditioned conjugate gradient methods commonly employed to solve such problems, including a multigrid preconditioned conjugate gradient method. [ABSTRACT FROM AUTHOR]

The effective conductivity of a block of composite can be extracted from the Dirichlet-to-Neumann Poincaré–Steklov operator (PSO) for that block. In this paper, a domain decomposition method for computing the PSO over a finite element mesh is discussed. A new numerical strategy is introduced to accelerate the computation of this operator, using the Schur complement to calculate the PSO for the smallest subdomains, then recursively merging subdomain PSOs up to the full domain. At each step of the algorithm, information extraneous to the PSO is discarded. The effective conductivity values computed by this method are identical to those obtained from a basic Finite Element Method, an order of magnitude faster and with much less computer memory consumed. As proof of concept, effective conductivity measurements are presented for a percolating random fractal-like microstructure across a range of phase fractions. [ABSTRACT FROM AUTHOR]

We analyse temporal approximation schemes based on overlapping domain decompositions. As such schemes enable computations on parallel and distributed hardware, they are commonly used when integrating large-scale parabolic systems. Our analysis is conducted by first casting the domain decomposition procedure into a variational framework based on weighted Sobolev spaces. The time integration of a parabolic system can then be interpreted as an operator splitting scheme applied to an abstract evolution equation governed by a maximal dissipative vector field. By utilizing this abstract setting, we derive an optimal temporal error analysis for the two most common choices of domain decomposition-based integrators: namely, alternating direction implicit schemes and additive splitting schemes of first and second order. For the standard first-order additive splitting scheme we also extend the error analysis to semilinear evolution equations, which may have only mild solutions. The theoretical results are finally illustrated by numerical experiments. [ABSTRACT FROM AUTHOR]

Earth System Models are critical tools for the study of our climate and its future trends. These models are in constant evolution and their growing complexity entails an incrementing demand of the resources they require. Since the cost of using these state-of-the-art models is huge, looking closely at the factors that are able to impact their computational performance is mandatory. In the case of the state-of-the-art ocean model NEMO (Nucleus for European Modelling of the Ocean), used in many projects around the world, not enough attention has been given to the domain decomposition. In this work we show the impact that the selection of a particular domain decomposition can have on computational performance and how the proposed methodology substantially improves it. [ABSTRACT FROM AUTHOR]

Although the HDMR decomposition has become an important tool for the understanding of high dimensional functions, the fundamental conjecture underlying its practical utility is still open for theoretical analysis. In this paper, we introduce the HDMR decomposition in conjunction with the Fourier-HDMR approximation leading to the following conclusions: (1) we suggest a type of Fourier-HDMR approximation for certain classes of differentiable functions; (2) utilizing the Fourier-HDMR method, we prove the fundamental conjecture about the dominance of low order terms in the HDMR expansion under relevant conditions, and we also obtain error estimates of the truncated HDMR expansion up to order u; (3) we prove the domain decomposition approximation theorem which shows that the global Fourier-HDMR approximation is not always optimal for a given accuracy order; (4) and finally, a piecewise Fourier-HDMR approach is discussed for high dimensional modeling. These results help to further understand how to efficiently represent the high dimensional functions. [ABSTRACT FROM AUTHOR]

We describe a new implementation of a two-level overlapping Schwarz preconditioner with energy-minimizing coarse space (GDSW: generalized Dryja--Smith--Widlund) and show numerical results for an additive and a hybrid additive-multiplicative version. Our parallel implementation makes use of the Trilinos software library and provides a framework for parallel two-level Schwarz methods. We show parallel scalability for two- and three-dimensional scalar second-order elliptic and linear elasticity problems for several thousands of cores. We also discuss techniques for the parallel construction of coarse spaces which are also of interest for other parallel preconditioners and discretization methods using energy minimizing coarse functions. We finally show an application in monolithic fluid-structure interaction, where significant improvements are achieved compared to a standard algebraic, one-level overlapping Schwarz method. [ABSTRACT FROM AUTHOR]

The accurate prediction of transmural stresses in arterial walls requires on the one hand robust and efficient numerical schemes for the solution of boundary value problems including fluid-structure interactions and on the other hand the use of a material model for the vessel wall that is able to capture the relevant features of the material behavior. One of the main contributions of this paper is the application of a highly nonlinear, polyconvex anisotropic structural model for the solid in the context of fluid-structure interaction, together with a suitable discretization. Additionally, the influence of viscoelasticity is investigated. The fluid-structure interaction problem is solved using a monolithic approach; that is, the nonlinear system is solved (after time and space discretizations) as a whole without splitting among its components. The linearized block systems are solved iteratively using parallel domain decomposition preconditioners. A simple - but nonsymmetric - curved geometry is proposed that is demonstrated to be suitable as a benchmark testbed for fluid-structure interaction simulations in biomechanics where nonlinear structural models are used. Based on the curved benchmark geometry, the influence of different material models, spatial discretizations, and meshes of varying refinement is investigated. It turns out that often-used standard displacement elements with linear shape functions are not sufficient to provide good approximations of the arterial wall stresses, whereas for standard displacement elements or F-bar formulations with quadratic shape functions, suitable results are obtained. For the time discretization, a second-order backward differentiation formula scheme is used. It is shown that the curved geometry enables the analysis of non-rotationally symmetric distributions of the mechanical fields. For instance, the maximal shear stresses in the fluid-structure interface are found to be higher in the inner curve that corresponds to clinical observations indicating a high plaque nucleation probability at such locations. Copyright © 2015 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

In this work we are interested in the search of interface conditions to couple hydrostatic and nonhydrostatic ocean models. To this aim, we consider simplified systems and use a time discretization to handle linear equations. We recall the links between the two models (with the particular role of the aspect ratio δ = H/L «1) and introduce an iterative method based on the Schwarz algorithm (widely used in domain decomposition methods). The convergence of this method depends strongly on the choice of interface conditions: this is why we look for exact absorbing conditions and their approximations in order to provide tractable and efficient coupling algorithms. [ABSTRACT FROM AUTHOR]

A class of preconditioners based on balancing domain decomposition by constraints methods is introduced in the Portable, Extensible Toolkit for Scientific Computation (PETSc). The algorithm and the underlying nonoverlapping domain decomposition framework are described with a specific focus on their current implementation in the library. Available user customizations are also presented, together with an experimental interface to the finite element tearing and interconnecting dual-primal methods within PETSc. Large-scale parallel numerical results are provided for the latest version of the code, which is able to tackle symmetric positive definite problems with highly heterogeneous distributions of the coefficients. Current limitations and future extensions of the preconditioner class are also discussed. [ABSTRACT FROM AUTHOR]

Trapped modes in a hard cylindrical tube with a local axisymmetric enlargement or bulge and filled with a uniform acoustic medium is studied. The governing Helmholtz equation in the cylindrical coordinate system is employed to deal with this problem through the domain decomposition method and matching technique. The trapped modes and the corresponding frequencies less than the threshold frequency or cut-off frequency are derived. It is found that in addition to the fundamental mode, the second- and higher-order trapped modes exist and depend on the geometry parameters of the local bulge. The effects of the bulge radius and width on the frequencies are discussed. The local bulge leads to a decrease of the frequencies and the corresponding vibration mode is localized near the bulge. A multimodal analysis is made and frequency band gap of generalized trapped modes is also studied. A frequency band gap depends on the radius of a bulge and is independent of its width. The obtained results can be extended to analyze bound states in quantum wires. [ABSTRACT FROM AUTHOR]

A new domain decomposition scheme based on equivalence principle algorithm (EPA) is proposed for analyzing electromagnetic scattering from connected structures. The EPA transforms the unknown of inside scatterers to the unknown on the equivalence surfaces which enclose the scatterers. The challenge arises when the equivalence surface intercepts the connected scatterers, because the current will be singular at the connection that is hard to model. In this paper, a new scheme based on source reconstruction method (SRM) is introduced to deal with the connected problems. This scheme can avoid the current singularity and keep the good convergence of EPA equations. Furthermore, generalized EPA (GEPA) is proposed to improve the theory of EPA and the ability to solve real life problems. Numerical results demonstrate the accuracy and efficiency of the proposed method. [ABSTRACT FROM AUTHOR]

A bound is obtained for the condition number of a BDDC algorithm for problems posed in H(curl) in two dimensions, where the subdomains are only assumed to be uniform in the sense of Peter Jones. For the primal variable space, a continuity constraint for the tangential average over each interior subdomain edge is imposed. For the averaging operator, a new technique named deluxe scaling is used. Our optimal bound is independent of jumps in the coefficients across the interface between the subdomains and depends only on a few geometric parameters of the decomposition. Numerical results that verify the result are shown, including some with subdomains with fractal edges and others obtained by a mesh partitioner. [ABSTRACT FROM AUTHOR]

Piezoelectric actuators are in common use for control of distributed parameter systems. We consider the topology optimization of a multiphysic model in piezoelectricity into two spatial dimensions. The topological derivative of a tracking-type shape functional is derived in its closed form for the purpose of shape optimization of piezoelectric actuators. The optimum design procedure is applied to a micromechanism which transforms the electrical energy supplemented via its piezoceramic part into the elastic energy of an actuator. The domain decomposition technique and the Steklov-Poincaré pseudodifferential boundary operator are employed in the asymptotic analysis of the shape functional defined on a part of the boundary of the elastic body under consideration. The new method of sensitivity analysis is general and can be used for shape-topological optimization in a broad class of multiphysics models. Our numerical results confirm the efficiency of the proposed approach to optimum design in multiphysics. [ABSTRACT FROM AUTHOR]

The interface control domain decomposition (ICDD) method of Discacciati, Gervasio, and Quarteroni [SIAM J. Control Optim., 51 (2013), pp. 3434-3458, doi:10.1137/120890764; J. Coupled Syst. Multiscale Dyn., 1 (2013), pp. 372-392, doi:10.1166/jcsmd.2013.1026] is proposed here to solve the coupling between Stokes and Darcy equations. According to this approach, the problem is formulated as an optimal control problem whose control variables are the traces of the velocity and the pressure on the internal boundaries of the subdomains that provide an overlapping decomposition of the original computational domain. A theoretical analysis is carried out, and the well-posedness of the problem is proved under certain assumptions on both the geometry and the model parameters. An efficient solution algorithm is proposed, and several numerical tests are implemented. Our results show the accuracy of the ICDD method, its computational efficiency, and robustness with respect to the different parameters involved (grid size, polynomial degrees, permeability of the porous domain, thickness of the overlapping region). The ICDD approach turns out to be more versatile and easier to implement than the celebrated model based on the Beavers, Joseph, and Saffman coupling conditions. [ABSTRACT FROM AUTHOR]

We present a solver for the 2D high-frequency Helmholtz equation in heterogeneous acoustic media, with online parallel complexity that scales optimally as O ( N L ) , where N is the number of volume unknowns, and L is the number of processors, as long as L grows at most like a small fractional power of N . The solver decomposes the domain into layers, and uses transmission conditions in boundary integral form to explicitly define “polarized traces”, i.e., up- and down-going waves sampled at interfaces. Local direct solvers are used in each layer to precompute traces of local Green's functions in an embarrassingly parallel way (the offline part), and incomplete Green's formulas are used to propagate interface data in a sweeping fashion, as a preconditioner inside a GMRES loop (the online part). Adaptive low-rank partitioning of the integral kernels is used to speed up their application to interface data. The method uses second-order finite differences. The complexity scalings are empirical but motivated by an analysis of ranks of off-diagonal blocks of oscillatory integrals. They continue to hold in the context of standard geophysical community models such as BP and Marmousi 2, where convergence occurs in 5 to 10 GMRES iterations. While the parallelism in this paper stems from decomposing the domain, we do not explore the alternative of parallelizing the systems solves with distributed linear algebra routines. [ABSTRACT FROM AUTHOR]

This article investigates the swept rule of space–time domain decomposition, an idea to break the latency barrier via communicating less often when explicitly solving time-dependent PDEs. The swept rule decomposes space and time among computing nodes in ways that exploit the domains of influence and the domain of dependency, making it possible to communicate once per many timesteps without redundant computation. The article presents simple theoretical analysis to the performance of the swept rule which then was shown to be accurate by conducting numerical experiments. [ABSTRACT FROM AUTHOR]

Quantifying uncertainty effects of coefficients that exhibit heterogeneity at multiple scales is among many outstanding challenges in subsurface flow models. Typically, the coefficients are modeled as functions of random variables governed by certain statistics. To quantify their uncertainty in the form of statistics (e.g., average fluid pressure or concentration) Monte-Carlo methods have been used. In a separate direction, multiscale numerical methods have been developed to efficiently capture spatial heterogeneity that otherwise would be intractable with standard numerical techniques. Since heterogeneity of individual realizations can differ drastically, a direct use of multiscale methods in Monte-Carlo simulations is problematic. Furthermore, Monte-Carlo methods are known to be very expensive as a lot of samples are required to adequately characterize the random component of the solution. In this study, we utilize a stochastic representation method that exploits the solution structure of the random process in order to construct a problem dependent stochastic basis. Using this stochastic basis representation a set of coupled yet deterministic equations is constructed. To reduce the computational cost of solving the coupled system, we develop a multiscale domain decomposition method utilizing Robin transmission conditions. In the proposed method, enrichment of the solution space can be performed at multiple levels that offer a balance between computational cost, and accuracy of the approximate solution. [ABSTRACT FROM AUTHOR]

Domain decomposition methods often exhibit very poor performance when applied to engineering problems with large heterogeneities. In particular, for heterogeneities along domain interfaces, the iterative techniques to solve the interface problem are lacking an efficient preconditioner. Recently, a robust approach, named finite element tearing and interconnection (FETI)-generalized eigenvalues in the overlaps (Geneo), was proposed where troublesome modes are precomputed and deflated from the interface problem. The cost of the FETI-Geneo is, however, high. We propose in this paper techniques that share similar ideas with FETI-Geneo but where no preprocessing is needed and that can be easily and efficiently implemented as an alternative to standard domain decomposition methods. In the block iterative approaches presented in this paper, the search space at every iteration on the interface problem contains as many directions as there are domains in the decomposition. Those search directions originate either from the domain-wise preconditioner (in the simultaneous FETI method) or from the block structure of the right-hand side of the interface problem (block FETI). We show on two-dimensional structural examples that both methods are robust and provide good convergence in the presence of high heterogeneities, even when the interface is jagged or when the domains have a bad aspect ratio. The simultaneous FETI was also efficiently implemented in an optimized parallel code and exhibited excellent performance compared with the regular FETI method. Copyright © 2015 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

In this paper, we present a new methodology for a nonoverlapping domain decomposition method using boundary integral equations and adaptive cross approximation in the frequency domain (IE-ACA-DDM). This paper proposes two main novelties. First, the IE-DDM condition number is improved with the development of a formulation based on electric field integral equation (EFIE) implementation over the entire subdomains and on a simplification of the magnetic field integral equation (MFIE) on the interfaces. As a consequence, the number of matrix–vector products is significantly reduced during the iterative resolution. Then, this paper proves that the ACA can be applied and efficiently implemented in IE-DDM methods. Moreover, a hierarchical LU (HLU) decomposition is proposed to accelerate the ACA solver. The IE-ACA-DDM is very suitable for the resolution of multiple excitations and to bypass the low-frequency breakdown. The flexibility, accuracy, and efficiency of IE-ACA-DDM are demonstrated by comparing with reference integral equation solutions and scaled mock-up measurements. [ABSTRACT FROM PUBLISHER]

The MPS is a particle method developed to simulate incompressible flows with free surfaces that has several applications in nonlinear hydrodynamics. Much effort has been done to handle the large amount of particles required for simulating practical problems with desired refinement. However, the efficient use of the currently available computational resources, such as computer cluster, remains as a challenge. The present paper proposes a new strategy to parallelize the MPS method for fully distributed computing in cluster, which enables to simulate models with hundreds of millions of particles and keeps the required runtime within reasonable limits, as shown by the analysis of scalability and performance. The proposed strategy uses a non-geometric dynamic domain decomposition method that provides homogeneous load balancing and for very large models the scalability is supra-linear. Also, the domain decomposition (DD) is carried out only in the initial setup. As a result, the DD method is based on renumbering of particles using an original fully distributed sorting algorithm. Moreover, unlike the usual strategies, none of the processors require access to global data of the particles on any time step. Therefore, the limit for the maximum size of the model depends more on the total memory of the allocated nodes than the quantity of the local memory of each node. Thus, by extending the application of MPS method to very large models, this study contributes to consolidating the method as a practical tool to investigate complex engineering problems. [ABSTRACT FROM AUTHOR]

In this work, we analyze the scalability of inexact two-level balancing domain decomposition by constraints (BDDC) preconditioners for Krylov subspace iterative solvers, when using a highly scalable asynchronous parallel implementation where fine and coarse correction computations are overlapped in time. This way, the coarse-grid problem can be fully overlapped by fine-grid computations (which are embarrassingly parallel) in a wide range of cases. Further, we consider inexact solvers to reduce the computational cost/complexity and memory consumption of coarse and local problems and boost the scalability of the solver. Out of our numerical experimentation, we conclude that the BDDC preconditioner is quite insensitive to inexact solvers. In particular, one cycle of algebraic multigrid (AMG) is enough to attain algorithmic scalability. Further, the clear reduction of computing time and memory requirements of inexact solvers compared to sparse direct ones makes possible to scale far beyond state-of-the-art BDDC implementations. Excellent weak scalability results have been obtained with the proposed inexact/overlapped implementation of the two-level BDDC preconditioner, up to 93,312 cores and 20 billion unknowns on JUQUEEN. Further, we have also applied the proposed setting to unstructured meshes and partitions for the pressure Poisson solver in the backward-facing step benchmark domain. [ABSTRACT FROM AUTHOR]

A special finite element method based on approximate component mode synthesis (ACMS) was introduced in Hetmaniuk and Lehoucq (2010). ACMS was developed for second order elliptic partial differential equations with rough or highly varying coefficients. Here, a parallel implementation of ACMS is presented and parallel scalability issues are discussed for representative examples. Additionally, a parallel domain decomposition preconditioner (FETI-DP) is applied to solve the ACMS finite element system. Weak parallel scalability results for ACMS are presented for up to 1024 cores. Our numerical results also suggest a quadratic–logarithmic condition number bound for the preconditioned FETI-DP method applied to ACMS discretizations. [ABSTRACT FROM AUTHOR]

In this work, we study time-splitting strategies for the numerical approximation of evolutionary reaction–diffusion problems. In particular, we formulate a family of domain decomposition splitting methods that overcomes some typical limitations of classical alternating direction implicit (ADI) schemes. The splitting error associated with such methods is observed to be O ( τ 2 ) in the time step τ . In order to decrease the size of this splitting error to O ( τ 3 ) , we add a correction term to the right-hand side of the original formulation. This procedure is based on the improved initialization technique proposed by Douglas and Kim in the framework of ADI methods. The resulting non-iterative schemes reduce the global system to a collection of uncoupled subdomain problems that can be solved in parallel. Computational results comparing the newly derived algorithms with the Crank–Nicolson scheme and certain ADI methods are presented. [ABSTRACT FROM AUTHOR]

We study convergence of the integral transport matrix method (ITMM) based on a parallel block Jacobi (PBJ) iterative strategy for solving particle transport problems. The ITMM is a spatial domain decomposition method proposed for massively parallel computations. A Fourier analysis of the PBJ-based iterations applied to S N diamond-difference equations in 1D slab and 2D Cartesian geometries is performed. It is carried out for infinite-medium problems with homogeneous material properties. To analyze the performance of the ITMM with the PBJ algorithm and evaluate its potential in scalability we consider a limiting case of one spatial cell per subdomain. The analysis shows that in such limit the spectral radius of the iteration method is one without regard to values of the scattering ratio and optical thickness of the spatial cells. This implies lack of convergence in infinite medium. Numerical results of finite-medium problems are presented. They demonstrate effects of finite size of spatial domain on the performance of the iteration algorithm as well as its asymptotic behavior when the extent of the spatial domain increases. These numerical experiments also show that for finite domains iterative convergence to a finite criterion is achievable in a multiple of the sum of number of cells in each dimension. [ABSTRACT FROM AUTHOR]

This paper introduces a domain decomposition preconditioner for elliptic equations with rough coefficients. The coarse space of the domain decomposition method is constructed via the so-called rough polyharmonic splines (RPS for short). As an approximation space of the elliptic problem, RPS is known to recover the quasi-optimal convergence rate and attain the quasi-optimal localization property. The authors lay out the formulation of the RPS based domain decomposition preconditioner, and numerically verify the performance boost of this method through several examples. [ABSTRACT FROM AUTHOR]

We consider a model for the propagation and absorption of electromagnetic waves (in the time-harmonic regime) in a magnetised plasma. We present a rigorous derivation of the model and several boundary conditions modelling wave injection into the plasma. Then we propose several variational formulations, mixed and non-mixed, and prove their well-posedness thanks to a theorem by Sébelin et al. Finally, we propose a non-overlapping domain decomposition framework, show its well-posedness and equivalence with the one-domain formulation. These results appear strongly linked to the spectral properties of the plasma dielectric tensor. [ABSTRACT FROM AUTHOR]

The optimised Schwarz method and the related two-Lagrange multiplier (2LM) method are nonoverlapping domain decomposition methods which can be used to numerically solve boundary value problems. Local Robin problems are solved on each subdomain in parallel to approximate the global solution, where a careful choice of Robin parameters leads to faster convergence. The 2LM method involves the solution of a nonsymmetric linear system that is usually solved with a Krylov subspace method such as GMRES. The speed of convergence of GMRES can be estimated using a conformal map from the exterior of the field of values of the system matrix to the interior of the unit disc. In this article we consider an elliptic PDE problem with a jump in diffusion coefficients across the interface between the subdomains. We approximate the field of values of the 2LM system matrix by a rectangle, R, in ℂ and provide optimised Robin parameters that ensure R is 'well conditioned' in the sense that GMRES converges quickly. We derive convergence estimates for GMRES and consider the behaviour asymptotically as the mesh size h becomes small and the jump in coefficients becomes large. We observe, for our choice of Robin parameters, that increasing the jump in coefficients increases the convergence rate of GMRES. Numerical experiments are performed to verify the theoretical results. [ABSTRACT FROM AUTHOR]

An efficient direct spectral domain decomposition method is developed coupled with Chebyshev spectral approximation for the solution of 2D, unsteady and incompressible Navier-Stokes equations in complex geometries. In this numerical approach, the spatial domains of interest are decomposed into several non-overlapping rectangular sub-domains. In each sub-domain, an improved projection scheme with second-order accuracy is used to deal with the coupling of velocity and pressure, and the Chebyshev collocation spectral method (CSM) is adopted to execute the spatial discretization. The influence matrix technique is employed to enforce the continuities of both variables and their normal derivatives between the adjacent sub-domains. The imposing of the Neumann boundary conditions to the Poisson equations of pressure and intermediate variable will result in the indeterminate solution. A new strategy of assuming the Dirichlet boundary conditions on interface and using the first-order normal derivatives as transmission conditions to keep the continuities of variables is proposed to overcome this trouble. Three test cases are used to verify the accuracy and efficiency, and the detailed comparison between the numerical results and the available solutions is done. The results indicate that the present method is efficiency, stability, and accuracy. [ABSTRACT FROM AUTHOR]

This paper investigates two FEM–BEM coupling formulations for acoustic fluid–structure interaction (FSI) problems, using the Finite Element Method (FEM) to model the structure and the Boundary Element Method (BEM) to represent a linear acoustic fluid. The coupling methods described interconnect fluid and structure using classical or localized Lagrange multipliers, allowing the connection of non-matching interfaces. First coupling technique is the well known mortar method, that uses classical multipliers and is compared with a new formulation of the method of localized Lagrange multipliers (LLM) for FSI applications with non-matching interfaces. The proposed non-overlapping domain decomposition technique uses a classical non-symmetrical acoustic BEM formulation for the fluid, although a symmetric Galerkin BEM formulation could be used as well. A comparison between the localized methodology and the mortar method in highly non conforming interface meshes is presented. Furthermore, the methodology proposes an iterative preconditioned and projected bi-conjugate gradient solver which presents very good scalability properties in the solution of this kind of problems. [ABSTRACT FROM AUTHOR]

We study non-overlapping Schwarz methods for solving a steady-state diffusion problem in heterogeneous media. Various optimized transmission conditions are determined by solving the corresponding min-max problems; we consider different choices of Robin conditions and second order conditions. To compare the resulting methods, we analyze the convergence in two separate asymptotic regimes: when the mesh size is small, and when the jump in the coefficient is large. It is shown that optimized two-sided Robin transmission conditions are very effective in both regimes; in particular they give mesh independent convergence. Numerical experiments are presented to illustrate and confirm the theoretical results. [ABSTRACT FROM AUTHOR]

In the paper two new locally one-dimensional alternating segment schemes for solving the 2-D parabolic problems are developed. The parabolic problems can be solved efficiently over decomposed sub-domains. Two schemes are proved to be unconditionally stable and numerical experiments show the validity of the schemes. [ABSTRACT FROM AUTHOR]

The FETI-DP (dual-primal finite element tearing and interconnecting) algorithms, proposed by the authors in [SIAM J. Numer. Anal., 51 (2013), pp. 1235-1253] and [Internat. J. Numer. Methods Engrg., 94 (2013), pp. 128-149] for solving incompressible Stokes equations, are extended to three-dimensional problems. A new analysis of the condition number bound for using the Dirichlet preconditioner is given. The algorithm and analysis are valid for mixed finite elements with both continuous and discontinuous pressures. An advantage of this new analysis is that the numerous coarse level velocity components, required in the previous analysis to enforce the divergence-free subdomain boundary velocity conditions, are no longer needed. This greatly reduces the size of the coarse level problem in the algorithm, especially for three-dimensional problems. The coarse level velocity space can be chosen as simple as those coarse spaces for solving scalar elliptic problems corresponding to each velocity component. Both the Dirichlet and lumped preconditioners are analyzed using the same framework in this new analysis. Their condition number bounds are proved to be independent of the number of subdomains for fixed subdomain problem size. Numerical experiments in both two and three dimensions, using mixed finite elements with both continuous and discontinuous pressures, demonstrate the convergence rate of the algorithms. [ABSTRACT FROM AUTHOR]

A new three scale approach for textile models is suggested: a one-dimensional fiber model and a fabric model, with a meso-level in between, i.e. the yarn scale, Goessens et al. (2012). For loose textile substrates this seems appropriate as the yarn level plays an important role. This is because the saturation vapor pressure will influence the release rate from the fibers, and its value will vary over the yarn cross section. Therefore, in this work we present two upscaling techniques for the three step multiscale model. The active component is tracked in the fiber, the yarn, and finally at the fabric level. At the fiber level a one-dimensional reduction to a non-linear diffusion equation is performed, and solved on an as needed basis. The outcome is upscaled via the volume averaging method and used as an input for the yarn level. At this level a one-dimensional model can be applied to calculate the concentration, which on its turn is upscaled using the overlapping domain decomposition as an input for the fabric level model. [ABSTRACT FROM AUTHOR]

Monte Carlo (MC) neutral particle transport codes are considered the gold-standard for nuclear simulations, but they cannot be robustly applied to high-fidelity nuclear reactor analysis without accommodating several terabytes of materials and tally data. While this is not a large amount of aggregate data for a typical high performance computer, MC methods are only embarrassingly parallel when the key data structures are replicated for each processing element, an approach which is likely infeasible on future machines. The present work explores the use of spatial domain decomposition to make full-scale nuclear reactor simulations tractable with Monte Carlo methods, presenting a simple implementation in a production-scale code. Good performance is achieved for mesh-tallies of up to 2.39 TB distributed across 512 compute nodes while running a full-core reactor benchmark on the Mira Blue Gene/Q supercomputer at the Argonne National Laboratory. In addition, the effects of load imbalances are explored with an updated performance model that is empirically validated against observed timing results. Several load balancing techniques are also implemented to demonstrate that imbalances can be largely mitigated, including a new and efficient way to distribute extra compute resources across finer domain meshes. [ABSTRACT FROM AUTHOR]