A forward-advancing wave expansion method for numerical solution of large-scale sound propagation problems

We present an important extension of the arbitrary high-order discontinuous Galerkin (DG) finite-element method to model 2-D elastic wave propagation in highly heterogeneous material. In this new approach we include space-variable coefficients to describe smooth or discontinuous material variations inside each element using the same numerical approximation strategy as for the velocity–stress variables in the formulation of the elastic wave equation. The combination of the DG method with a time integration scheme based on the solution of arbitrary accuracy derivatives Riemann problems still provides an explicit, one-step scheme which achieves arbitrary high-order accuracy in space and time. Compared to previous formulations the new scheme contains two additional terms in the form of volume integrals. We show that the increasing computational cost per element can be overcompensated due to the improved material representation inside each element as coarser meshes can be used which reduces the total number of elements and therefore computational time to reach a desired error level. We confirm the accuracy of the proposed scheme performing convergence tests and several numerical experiments considering smooth and highly heterogeneous material. As the approximation of the velocity and stress variables in the wave equation and of the material properties in the model can be chosen independently, we investigate the influence of the polynomial material representation on the accuracy of the synthetic seismograms with respect to computational cost. Moreover, we study the behaviour of the new method on strong material discontinuities, in the case where the mesh is not aligned with such a material interface. In this case second-order linear material approximation seems to be the best choice, with higher-order intra-cell approximation leading to potential instable behaviour. For all test cases we validate our solution against the well-established standard fourth-order finite difference and spectral element method.

This paper is concerned with the numerical approximation of some linear stochastic partial differential equations with additive noises. A special representation of the noise is considered, and it is compared with general representations of noises in the infinite dimensional setting. Convergence analysis and error estimates are presented for the numerical solution based on the standard finite difference and finite element methods. The effects of the noises on the accuracy of the approximations are illustrated. Results of the numerical experiments are provided.

The parabolic equation (PE) method is widely used in radiowave propagation predictions. It has the advantages of high efficiency and stability, but it will lead to greater predicting errors in some situations, because the effects of transverse terrain gradients are not modeled. This problem can be solved by extending the 2D PE to the three-dimensional (3D) PE. However, the computing efficiency will degrade because of large scale matrix operations. In this paper, a new method is presented, in which the 3D PE is decomposed into two 2D PEs. It increases the computational efficiency and accuracy effectively. To verify the capability of the proposed method in radiowave propagation prediction, an experiment platform was set up. The computational results using this new method are compared with the experimental and Method of Moment (MoM) numerical computational results. Good agreements are achieved in the comparison. The study of radiowave propagation under complex environments is vitally important for improving the effective signal coverage in wireless communication and television broadcasting with plenty of urban buildings as well as mountainous terrains. It presents a big challenge due to the multiple wave propagation mechanisms involving the line of sight, diffraction and scattering propagation and the characteristic of large scale computation. The major issues refer to how to keep the computational accuracy and conducting the large scale computation at the same time. To meet the request for large scale computation of actual engineering problems, the modified empirical formulas are often used for analysis and computation with the drawback of poor computational accuracy, and the variance of the predicted propagation loss is generally 2 to 5 dB for the flat terrain, but for the complex terrain, the variance of the predicted propagation loss is usually about 5 to 10 dB. Many computational methods have been developed to improve the predicted accuracy. The commonly used methods are geometric diffraction methods, path integration methods and FDTD. It is hard to meet the requirement of engineering computation because these methods are computationally expensive and require high data accuracy of the terrain database at the same time (1). The Parabolic Equation (PE) method is a computational method for radiowave propagation loss besides the full wave method and ray tracing method, and has become a research focus. The PE method has high potential in radiowave propagation of mobile communication and broadcasting and television signal coverage as it can not only deal with complex terrain boundary conditions accurately, but also deal with complex atmosphere waveguide structure. Compared with the full wave method, the PE method deals with an unidirectional wave propagation, in which the equation can be solved by the step marching method in the propagation direction. The PE method can reduce the computational complexity of the full wave equation.

Bilevel learning for large-scale flexible flow shop scheduling

The surface electric field analysis of the converter valve shield system is a large-scale electrostatic field problem, which is difficult to analyse. The fast multipole boundary element method (FMBEM), which is suitable for solving large-scale problems, can accelerate the computation speed and conserve memory. However, the coefficient matrix implicitly formed by using the FMBEM is sometimes ill-conditioned, especially for large-scale problems; thus, the convergence of iteration is poor. In this paper, a fast solver is proposed to improve efficiency. First, an adaptive GMRES( m ) with variant restart parameter is adjusted for the Galerkin FMBEM. In addition, the sparse approximate inverse preconditioner is improved, and a new sparsity pattern is proposed for the multiscale problem derived from the converter valve shield system. The numerical results show that the accuracy can meet the engineering requirements compared with the finite element method. Compared with other solvers and preconditioners, the algorithm can achieve a satisfactory convergence rate and reduce the computation time. In addition, a single bridge shield system of ±160 kV converter valve is successfully analysed using the proposed method.

Finite Element Method (FEM1) is pervasively used in most of 3D product design analysis, in which Computer Aided Design (CAD) models need to be converted in to mesh models first and then enriched with some material features and boundary conditions data, etc. The interaction between CAD models and FEM models is intensive. Boundary Element Method (BEM) has been expected to be advantageous in large-scale problems in recent years owing to its reduction of the dimensionality and its reduced complexity in mesh generation. However, the BEM application has so far been limited to relatively small problems due to the memory and computational complexity for matrix buildup are O(N2). The fast multipole BEM (FMBEM) combined with BEM and fast multipole method (FMM) can overcome the defect of the traditional BEM, and provides an effective method to solve the large-scale problem. Combining GPU parallel computing with FMBEM can further improve its efficiency significantly. Based on the three-dimensional elastic mechanics problems, the parallelisms of the multipole moment (ME), multipole moment to multipole moment (M2M) translation, multipole moment to local expansion (M2L) translation, local expansion to local expansion (L2L) translation and near-field direct calculation were analyzed respectively according to the characteristics of the FMM, and the parallel strategies under CUDA were presented in this paper. Three main major parts are included herein: (1) FMBEM theory in 3D elastostatics, (2) the parallel FMBEM algorithm using CUDA, and (3) comparison the GPU parallel FMBEM with BEM, FEM and FMBEM respectively by engineering examples. Numerical example results show the 3D elastostatics GPU FMBEM not only can speed up the boundary element calculation process, but also save memory which can be effective to solve the large-scale engineering problems.

Efficient optimization of a large-scale biorefinery system using a novel decomposition based approach

Lanczos' generalized derivative: Insights and Applications

Fluorescence diffuse optical tomography is a powerful tool for the investigation of molecular events in studies for new therapeutic developments. Here, the emphasis is put on the mathematical problem of tomography, which can be formulated in terms of an estimation of physical parameters appearing as a set of Partial Differential Equations (PDEs). The standard polynomial Finite Element Method (FEM) is a method of choice to solve the diffusion equation because it has no restriction in terms of neither the geometry nor the homogeneity of the system, but it is time consuming. In order to speed up computation time, this paper proposes an alternative numerical model, describing the diffusion operator in orthonormal basis of compactly supported wavelets. The discretization of the PDEs yields to matrices which are easily computed from derivative wavelet product integrals. Due to the shape of the wavelet basis, the studied domain is included in a regular fictitious domain. A validation study and a comparison with the standard FEM are conducted on synthetic data.

A new coarse-mesh computational method for the numerical solution of heat conduction and fluid flow problems is formally developed and applied to sample problems. The method is based upon formal use of Green's functions, which are defined locally over subdomains of the original system under consideration. The formal development of the local Green's function method for the solution of heat conduction problems is presented and discussed. Numerical solutions of sample problems for one-dimensional heat conduction with constant thermal conductivity, one-dimensional heat conduction with temperature-dependent thermal conductivity, and two-dimensional heat conduction with constant thermal conductivity are given, and these results are compared with results obtained using the finite difference and finite element methods. The formal development of the local Green's function method for the solution of fluid flow problems is then also presented and discussed; the numerical solution of a sample problem for simple one-dimensional incompressible fluid flow with viscous heating is also given.

This paper proposes a general method of solution for problems, involving the allocation of jobs to machines and the sequencing of jobs on machines, in which a number of jobs, each consisting of a number of operations that may be performed on a given set of machines, is to be completed in some optimal fashion. Considering constraints on machine availability, job completion dates, and zoning, the paper demonstrates the underlying relation of a wide variety of machine-job problems with reference to their graphical representation, states a branch-and-bound method of solution, solves an illustrative example, and gives the mode of application to specific problems (for several of these the method reduces to methods published elsewhere). The algorithm represents not only a general framework of theoretical interest, but also a practical approach to certain specific problems.

We present a new fast finite element method which is called the generalized-simplified marker and cell method (GSMAC) as an extention of the SMAC finite difference method of Amsden and Harlow. There are several new features of the GSMAC method, which is distinguished from other iterative finite element methods. We use two important concepts, orthogonal decomposition and cycle-to-cycle self-adjustment, which allow a stable and accurate calculation of unsteady flow fields at high Reynolds numbers without any artificial modification. As a result, the GSMAC method has proved to be faster than the other finite element methods and to require only the same amount of CPU time and storage as the standard finite difference method. The results of calculation in a square cavity agree very well with those of Ghia et al.. Especially for Re=5000 and 10000, the same degree of accuracy can be obtained with 1/26.4 of mesh points.

We present an approach for efficient numerical solution of optimization problems governed by parabolic partial differential equations. The main ingredients are: space-time finite element discretization, second-order optimization algorithms, and storage reduction techniques. We discuss the combination of these components for the solution of large-scale optimization problems.

A large scale low-profile traveling-wave narrow-wall slotted waveguide array with 2160 slots is analyzed by parallel method of moments (MoM) out-of-core solver based on the higher order basis functions (HOBs). Its accuracy is validated through comparison with the finite element method (FEM) adopted in the commercial software Ansoft HFSS. An out-of-core two-dimensional block-cyclic distribution scheme for the large dense full matrix filling and an out-of-core LU (lower/upper) decomposition matrix equation solver are utilized under the message passing interface (MPI). Numerical results are shown to demonstrate the memory cost effectiveness, time efficiency of the proposed method and its applicability. The method presented in this paper is able to solve challenging realistic problems in large scale.

The performance of electrical machines is largely dictated by the action of current and flux in the core length. The field in a cross-section obeys Poisson's equation and approximate solutions have been obtained by finite difference and element methods. The finite difference method requires a large number of nodes and is slow to converge as permeability is variable. The finite element method is more flexible being more readily fitted to iron-air boundaries and has better convergence. However, it is difficult to formulate a legitimate variational formulation for transient conditions in the presence of dissipation. Here, discrete equations are formed by applying Ampere's circuital law around each node. Careful choice of contour lines give a current distribution superior to that obtained with finite elements. Fast convergence is obtained and the method is applicable under transient conditions.