Search

Fluid–structure interaction simulations can be performed in a partitioned way, by coupling a flow solver with a structural solver. However, Gauss–Seidel iterations between these solvers without additional stabilization efforts will converge slowly or not at all under common conditions such as an incompressible fluid and a high added mass. Quasi-Newton methods can then stabilize and accelerate the coupling iterations, while still using the solvers as black boxes and only accessing data at the fluid–structure interface. In this review, the IQN-ILS, IQN-MVJ, IBQN-LS, MVQN, IQN-IMVLS and IQN-ILSM methods are reformulated in the generalized Broyden framework to illustrate their similarities and differences. Also related coupling techniques are reviewed and a performance comparison is provided where available.

This article investigates two aspects of the generalized Broyden quasi-Newton method that have a major impact on its convergence: the initial approximation of the Jacobian and the presence of nonlinearities in the secant conditions. After reformulating the common representation of generalized Broyden, a straightforward interpretation is given. This leads to a natural extension of the method in which an application-dependent physics-based surrogate model is used as initial approximation of the (inverse) Jacobian. A carefully chosen surrogate has the potential to greatly reduce the required number of iterations. The behavior of generalized Broyden depends strongly on the parameter that determines how many secant conditions are satisfied by the Jacobian approximation. Respecting all secant conditions reduces it to Anderson acceleration; a single one to Broyden's original method. An analysis demonstrates that these two variants behave very differently when nonlinearities are present in the secant conditions: they are ignored by Broyden, but can destabilize Anderson. On the other hand, the analysis shows that Broyden tends to neglect small linear information, possibly reducing convergence speed. To mitigate stability problems with Anderson acceleration, a practical method to detect and remove nonlinear secant information is introduced next. Finally, we solve a steady free-surface-flow problem using several generalized Broyden variants, testing the influence of the surrogate, the nonlinearities and the combination thereof. The results agree with the theoretical predictions, showing large differences in convergence behavior. Furthermore, the proposed method effectively negates the problems related to nonlinearities in this case.

This dissertation describes optimization methods and matrix factorizations for large-scale quasi-Newton trust-region methods. The proposed methods are applicable to convex and non-convex optimization problems, and are described in algorithms for software implementations.

In this article, we introduce the Free surface Quasi-Newton method with Least-Squares and Surrogate (FreQ-LeSS) to solve steady free surface flows, as encountered in, for example, marine applications. The method has two important advantages over other methods. First of all it is fast thanks to the efficient iterative scheme based on quasi-Newton iterations that make use of an advanced surrogate model. Second, it has good compatibility because it poses no special requirements on the flow solver that is called in the algorithm. The excellent convergence speed of the FreQ-LeSS method is demonstrated for a shallowly submerged submarine: good quality wave patterns are obtained in few iterations. These results match with ISIS-CFD, validating the FreQ-LeSS method. The total cost of solving this steady free surface problem is only a few times that of solving a steady single-phase flow with fixed free surface position. We conclude that FreQ-LeSS promises to be much faster than the two-phase time-stepping methods typically used to solve these problems, and provides a sharper interface at the same time thanks to the fitting approach. In the present work, we restrict ourselves to problems without surface-penetrating objects, to avoid the associated complications related to the contact-line motion.

The purpose of this work is to construct a robust numerical scheme for a class of nonlinear free boundary identification problems. First, a shape optimization problem is constructed based on a least square functional. Schauder’s fixed point theorem is manipulated to show the existence solution for the state solution. The existence of an optimal solution of the optimization problem is proved. The proposed numerical scheme is based on the Radial Basis Functions method as a discretization approach, the minimization process is a hybrid Differential Evolution heuristic method and the quasi-Newton method. At the end we establish some numerical examples to show the validity of the theoretical results and robustness of the proposed scheme.

In this paper, an active set quasi-Newton method for bound constrained nonlinear equation is proposed. By using this active set technique, we only need to solve a reduced dimension linear equation at each iteration to generate the search direction. The algorithm is a combination of the quasi-Newton method and projection method. Firstly we use the quasi-Newton step as the trial step and then use a projection technique to generate the next iteration point. Our key observation is that the algorithm generates a bounded iteration sequence automatically even if the bounds are equal to infinity and the global convergence is obtained in the sense that the whole sequence converges to the stationary point. The numerical tests show the efficiency of the algorithm.

We propose two enhancements of quasi-Newton methods used to accelerate coupling iterations for partitioned fluid-structure interaction. Quasi-Newton methods have been established as flexible, yet robust, efficient and accurate coupling methods of multi-physics simulations in general. The coupling library preCICE provides several variants, the so-called IQN-ILS method being the most commonly used. It uses input and output differences of the coupled solvers collected in previous iterations and time steps to approximate Newton iterations. To make quasi-Newton methods both applicable for parallel coupling (where these differences contain data from different physical fields) and to provide a robust approach for re-using information, a combination of information filtering and scaling for the different physical fields is typically required. This leads to good convergence, but increases the cost per iteration. We propose two new approaches - pre-scaling weight monitoring and a new, so-called QR3 filter, to substantially improve runtime while not affecting convergence quality. We evaluate these for a variety of fluid-structure interaction examples. Results show that we achieve drastic speedups for the pure quasi-Newton update steps. In the future, we intend to apply the methods also to volume-coupled scenarios, where these gains can be decisive for the feasibility of the coupling approach., Deutsche Forschungsgemeinschaft (DFG, German Research Foundation), Germany’s Excellence Strategy

We present a generalised phase field-based formulation for predicting fatigue crack growth in metals. The theoretical framework aims at covering a wide range of material behaviour . Different fatigue degradation functions are considered and their influence is benchmarked against experiments. The phase field constitutive theory accommodates the so-called AT1 , AT2 and phase field-cohesive zone ( PF-CZM ) models. In regards to material deformation, both non-linear kinematic and isotropic hardening are considered, as well as the combination of the two. Moreover, a monolithic solution scheme based on quasi-Newton algorithms is presented and shown to significantly outperform staggered approaches. The potential of the computational framework is demonstrated by investigating several 2D and 3D boundary value problems of particular interest. Constitutive and numerical choices are compared and insight is gained into their differences and similarities. The framework enables predicting fatigue crack growth in arbitrary geometries and for materials exhibiting complex (cyclic) deformation and damage responses. The finite element code developed is made freely available at www.empaneda.com/codes .

International audience; Newton's method is an ubiquitous tool to solve equations, both in the archimedean and non-archimedean settings --- for which it does not really differ. Broyden was the instigator of what is called "quasi-Newton methods". These methods use an iteration step where one does not need to compute a complete Jacobian matrix nor its inverse. We provide an adaptation of Broyden's method in a general non-archimedean setting, compatible with the lack of inner product, and study its Q and R convergence. We prove that our adapted method converges at least Q-linearly and R-superlinearly with R-order $2^{\frac{1}{2m}}$ in dimension m. Numerical data are provided.

Steady free surface flows are of interest in the fields of marine and hydraulic engineering. Fitting methods are generally used to represent the free surface position with a deforming grid. Existing fitting methods tend to use time-stepping schemes, which is inefficient for steady flows. There also exists a steady iterative method, but that one needs to be implemented with a dedicated solver. Therefore a new method is proposed to efficiently simulate two-dimensional (2D) steady free surface flows, suitable for use in conjunction with black-box flow solvers. The free surface position is calculated with a quasi-Newton method, where the approximate Jacobian is constructed in a novel way by combining data from past iterations with an analytical model based on a perturbation analysis of a potential flow. The method is tested on two 2D cases: the flow over a bottom topography and the flow over a hydrofoil. For all simulations the new method converges exponentially and in few iterations. Furthermore, convergence is independent of the free surface mesh size for all tests.

A fast parallelable Jacobi iteration type optimization method for non-smooth convex composite optimization is presented. Traditional gradient-based techniques cannot solve the problem. Smooth approximate functions are attempted to be used as a replacement of those non-smooth terms without compromising the accuracy. Recently, proximal mapping concept has been introduced into this field. Techniques which utilize proximal average based proximal gradient have been used to solve the problem. The state-of-art methods only utilize first-order information of the smooth approximate function. We integrate both first and second-order techniques to use both first and second-order information to boost the convergence speed. A convergence rate with a lower bound of O([Formula presented]) is achieved by the proposed method and a super-linear convergence is enjoyed when there is proper second-order information. In experiments, the proposed method converges significantly better than the state of art methods which enjoy O([Formula presented]) convergence. National Research Foundation (NRF) This work was conducted within the Rolls-Royce@NTU Corporate Lab with support from the National Research Foundation (NRF) Singapore un-der the Corp Lab@University Scheme and Energy Research Institute@NTU under Interdisciplinary Graduate School in Nanyang Technological University.

Summary Integrated nested Laplace approximation provides accurate and efficient approximations for marginal distributions in latent Gaussian random field models. Computational feasibility of the original Rue et al. (2009) methods relies on efficient approximation of Laplace approximations for the marginal distributions of the coefficients of the latent field, conditional on the data and hyperparameters. The computational efficiency of these approximations depends on the Gaussian field having a Markov structure. This note provides equivalent efficiency without requiring the Markov property, which allows for straightforward use of latent Gaussian fields without a sparse structure, such as reduced rank multi-dimensional smoothing splines. The method avoids the approximation for conditional modes used in Rue et al. (2009), and uses a log determinant approximation based on a simple quasi-Newton update. The latter has a desirable property not shared by the most commonly used variant of the original method.

In the present work, some numerical and computational aspects of COSMO-based activity coefficient models were explored. The residual contribution in such models rely on the so called self-consistency equation. This equation does not have a closed-form solution and is usually solved by the successive substitution method. The performance of a classical Newton-Raphson method was tested in solving the self-consistency equation. The results obtained by the Newton implementation and by successive substitution agreed within the convergence tolerance. The CPU times for solving the model using both methods also were compared. Contradicting the usual experience, it was observed that the Newton method becomes slower than successive substitution when the number of components (or number of COSMO segments) in the mixture increases. An analysis of the number of floating point operations required showed the same, Newton’s method will be faster only for small systems.

We investigate the potential of quasi-Newton methods in facilitating convergence of monolithic solution schemes for phase field fracture modelling. Several paradigmatic boundary value problems are addressed, spanning the fields of quasi-static fracture, fatigue damage and dynamic cracking. The finite element results obtained reveal the robustness of quasi-Newton monolithic schemes, with convergence readily attained under both stable and unstable cracking conditions. Moreover, since the solution method is unconditionally stable, very significant computational gains are observed relative to the widely used staggered solution schemes. In addition, a new adaptive time increment scheme is presented to further reduce the computational cost while allowing to accurately resolve sudden changes in material behavior, such as unstable crack growth. Computation times can be reduced by several orders of magnitude, with the number of load increments required by the corresponding staggered solution being up to 3000 times higher. Quasi-Newton monolithic solution schemes can be a key enabler for large scale phase field fracture simulations. Implications are particularly relevant for the emerging field of phase field fatigue, as results show that staggered cycle-by-cycle calculations are prohibitive in mid or high cycle fatigue. The finite element codes are available to download from www.empaneda.com/codes .

We present a Newton type algorithm to find parametric surfaces of prescribed mean curvature with a fixed given boundary. In particular, it applies to the problem of minimal surfaces. The algorithm relies on some global regularity of the spaces where it is posed, which is naturally fitted for discretization with isogeometric type of spaces. We introduce a discretization of the continuous algorithm and present a simple implementation using the recently released isogeometric software library igatools. Finally, we show several numerical experiments which highlight the convergence properties of the scheme. Fil: Chicco Ruiz, Anibal Leonardo. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Santa Fe. Instituto de Matemática Aplicada del Litoral. Universidad Nacional del Litoral. Instituto de Matemática Aplicada del Litoral; Argentina Fil: Morin, Pedro. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Santa Fe. Instituto de Matemática Aplicada del Litoral. Universidad Nacional del Litoral. Instituto de Matemática Aplicada del Litoral; Argentina Fil: Pauletti, Miguel Sebastian. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Santa Fe. Instituto de Matemática Aplicada del Litoral. Universidad Nacional del Litoral. Instituto de Matemática Aplicada del Litoral; Argentina

This article presents BiqCrunch , an exact solver for binary quadratic optimization problems. BiqCrunch is a branch-and-bound method that uses an original, efficient semidefinite-optimization-based bounding procedure. It has been successfully tested on a variety of well-known combinatorial optimization problems, such as Max-Cut, Max- k -Cluster, and Max-Independent-Set. The code is publicly available online; a web interface and many conversion tools are also provided.

In this review article, the focus is on partitioned simulation techniques for strongly coupled fluid-structure interaction problems, especially on techniques which use at least one of the solvers as a black box. First, a number of analyses are reviewed to explain why Gauss-Seidel coupling iterations converge slowly or not at all for fluid-structure interaction problems with strong coupling. This provides the theoretical basis for the fast convergence of quasi-Newton and multi-level techniques. Second, several partitioned techniques that couple two black-box solvers are compared with respect to implementation and performance. Furthermore, performance comparisons between partitioned and monolithic techniques are examined. Subsequently, two similar techniques to couple a black-box solver with an accessible solver are analyzed. In addition, several other techniques for fluid-structure interaction simulations are studied and various methods to take into account deforming fluid domains are discussed.

In this paper, we demonstrate a local convergence of an adaptive scalar solver which is practical for strongly diagonal dominant Jacobian problems such as in some systems of nonlinear equations arising from the application of a nonoverlapping domain decomposition method. The method is tested to a nonlinear interface problem of a multichip heat conduction problem. The numerical results show that the method performs slightly better than a Newton–Krylov method.

Fluid-structure interaction (FSI) can be simulated in a monolithic way by solving the flow and structural equations simultaneously and in a partitioned way with separate solvers for the flow equations and the structural equations. A partitioned quasi-Newton technique which solves the coupled problem through nonlinear equations corresponding to the interface position is presented and its performance is compared with a monolithic Newton algorithm. Various structural configurations with an incompressible fluid are solved, and the ratio of the time for the partitioned simulation, when convergence is reached, to the time for the monolithic simulation is found to be between 1/2 and 4. However, in this comparison of the partitioned and monolithic simulations, the flow and structural equations have been solved with a direct sparse solver in full Newton-Raphson iterations, only relatively small problems have been solved and this ratio would likely change if large industrial problems were considered or if other solution strategies were used.

A new smoothing quasi-Newton method for nonlinear complementarity problems is presented. The method is a generalization of Thomas’ method for smooth nonlinear systems and has similar properties as Broyden's method. Local convergence is analyzed for a strictly complementary solution as well as for a degenerate solution. Presented numerical results demonstrate quite similar behavior of Thomas’ and Broyden's methods.

International audience; Magnetic resonance electric properties tomography is a non-destructive imaging modality that maps the spatial distribution of the electrical conductivity and permittivity of the human body using standard clinical magnetic resonance imaging systems. From the B+1 magnetic field maps and the local form of the Maxwell equations, several schemes have been derived to provide direct approximated formulas but they suffer from instabilities. In this paper, we propose to address it as an inverse problem solved by a constrained optimization algorithm where we exploit the weak formulation of the electric Helmholtz equation and a Lagrangian approach. We derive the associated adjoint field equation and employ a quasi-Newton minimization scheme. We also take advantage of a regularisation strategy based on geometrical a priori information for defining large zones into which the electric parameters are known to be piece-wise constant.

An algorithm for the simulation of unsteady, viscous, stratified compressible flows, which remains valid at all speeds, is presented. The method is second-order accurate in both space and time and is independent of the Mach number. In order to remove the stiffness of the numerical problem due to the large disparity between the flow speed and the acoustic wave speed at low Mach number, an approximate Newton method, based on artificial compressibility, is proposed. Additionally, a modified advection upstream splitting method (AUSM +) scheme is used, which permits accurate computations of both compressible and incompressible flows. A detailed description of the method and an efficiency comparison with other approximate Newton methods described in the literature are given. Furthermore, it is shown that the accuracy of the algorithm is not dependent on the Mach number through the computations of various benchmark test cases. Copyright (C) 2000 John Wiley & Sons, Ltd.

International audience; Computing the minimal polyadic decomposition (also often referred to as canonical decomposition, or sometimes Parafac) amounts to finding the global minimum of a coercive polynomial in many variables. In the case of arrays with nonnegative entries, the low-rank approximation problem is well posed. In addition, due to the large dimension of the problem, the decomposition can be rather efficiently calculated with the help of preconditioned nonlinear conjugate gradient algorithms, as subsequently shown, if equipped with an algebraic calculation of the globally optimal stepsize in low dimension. Other algorithms are also studied (gradient and quasi-Newton approaches) for comparisons. Two versions of each algorithm are considered: the Enhanced Line Search version (ELS), and the backtracking version alternating with ELS. Computer simulations are provided and demonstrate the good behavior of these algorithms dedicated to nonnegative arrays, compared to others put forward in the literature. Finally, applications in the context of data analysis illustrate various algorithms. The main advantage of the suggested approach is to explicitly take into account the nonnegative nature of the loading matrices in the problem parameterization, instead of enforcing positive entries by projection. According to the experiments we have run, such an approach also happens to be more robust with respect to possible modeling errors.

International audience; This paper deals with the minimal polyadic decomposition (also known as canonical decomposition or Parafac) of a 3-way array, assuming each entry is positive. In this case, the low-rank approximation problem becomes well-posed. The suggested approach consists of taking into account the nonnegative nature of the loading matrices directly in the problem parameterization. Then, the three gradient components are derived allowing to efficiently implement the decomposition using classical optimization algorithms. In our case, we focus on the conjugate gradient algorithm, well matched to large problems. The good behaviour of the proposed approach is illustrated through computer simulations in the context of data analysis and compared to other existing approaches.