scholarly journals A Globally Convergent Filter-Type Trust Region Method for Semidefinite Programming

2012 ◽  
Vol 2012 ◽  
pp. 1-13
Author(s):  
Aiqun Huang ◽  
Chengxian Xu
Keyword(s):  
Semidefinite Programming ◽  
Large Scale ◽  
Linear Equations ◽  
Trust Region Method ◽  
Trust Region ◽  
Filter Method ◽  
System Of Linear Equations ◽  
Convergence Results ◽  
Filter Type ◽  
Interior Methods

When using interior methods for solving semidefinite programming (SDP), one needs to solve a system of linear equations at each iteration. For problems of large size, solving the system of linear equations can be very expensive. In this paper, based on a semismooth equation reformulation using Fischer's function, we propose a filter method with trust region for solving large-scale SDP problems. At each iteration we perform a number of conjugate gradient iterations, but do not need to solve a system of linear equations. Under mild assumptions, the convergence of this algorithm is established. Numerical examples are given to illustrate the convergence results obtained.

scholarly journals Global Convergence of a New Nonmonotone Filter Method for Equality Constrained Optimization

2013 ◽  
Vol 2013 ◽  
pp. 1-9
Author(s):  
Ke Su ◽  
Wei Liu ◽  
Xiaoli Lu
Keyword(s):  
Optimization Problems ◽  
Trust Region Method ◽  
Trust Region ◽  
Filter Method ◽  
Type Method ◽  
Numerical Tests ◽  
Starting Point ◽  
Filter Type ◽  
Acceptability Criterion ◽  
The Given

A new nonmonotone filter trust region method is introduced for solving optimization problems with equality constraints. This method directly uses the dominated area of the filter as an acceptability criterion for trial points and allows the dominated area decreasing nonmonotonically. Compared with the filter-type method, our method has more flexible criteria and can avoid Maratos effect in a certain degree. Under reasonable assumptions, we prove that the given algorithm is globally convergent to a first order stationary point for all possible choices of the starting point. Numerical tests are presented to show the effectiveness of the proposed algorithm.

scholarly journals AN IMPROVED ADAPTIVE TRUST-REGION METHOD FOR UNCONSTRAINED OPTIMIZATION

2014 ◽  
Vol 19 (4) ◽  
pp. 469-490 ◽  
Author(s):  
Hamid Esmaeili ◽  
Morteza Kimiaei
Keyword(s):  
Unconstrained Optimization ◽  
Large Scale ◽  
Optimization Problems ◽  
Trust Region Method ◽  
Trust Region ◽  
Standard Test ◽  
Test Problems ◽  
Unconstrained Optimization Problems ◽  
Large Scale Problems ◽  
Adaptive Radius

In this study, we propose a trust-region-based procedure to solve unconstrained optimization problems that take advantage of the nonmonotone technique to introduce an efficient adaptive radius strategy. In our approach, the adaptive technique leads to decreasing the total number of iterations, while utilizing the structure of nonmonotone formula helps us to handle large-scale problems. The new algorithm preserves the global convergence and has quadratic convergence under suitable conditions. Preliminary numerical experiments on standard test problems indicate the efficiency and robustness of the proposed approach for solving unconstrained optimization problems.

scholarly journals Numerical solutions to integral equations equivalent to differential equations with fractional time

2010 ◽  
Vol 20 (2) ◽  
pp. 261-269 ◽  
Author(s):  
Bartosz Bandrowski ◽  
Anna Karczewska ◽  
Piotr Rozmej
Keyword(s):  
Differential Equations ◽  
Integral Equations ◽  
Large Scale ◽  
Numerical Solutions ◽  
Symmetric Matrix ◽  
Linear Equations ◽  
Time Variable ◽  
Gmres Method ◽  
System Of Linear Equations ◽  
Variable Equation

Numerical solutions to integral equations equivalent to differential equations with fractional timeThis paper presents an approximate method of solving the fractional (in the time variable) equation which describes the processes lying between heat and wave behavior. The approximation consists in the application of a finite subspace of an infinite basis in the time variable (Galerkin method) and discretization in space variables. In the final step, a large-scale system of linear equations with a non-symmetric matrix is solved with the use of the iterative GMRES method.

scholarly journals A Newton-Like Trust Region Method for Large-Scale Unconstrained Nonconvex Minimization

2013 ◽  
Vol 2013 ◽  
pp. 1-6 ◽  
Author(s):  
Yang Weiwei ◽  
Yang Yueting ◽  
Zhang Chenhui ◽  
Cao Mingyuan
Keyword(s):  
Large Scale ◽  
Trust Region Method ◽  
Trust Region ◽  
Test Problems ◽  
Nonconvex Minimization ◽  
Trust Region Subproblem ◽  
Newton Equation ◽  
Limited Memory ◽  
Quasi Newton ◽  
Approximate Hessian

We present a new Newton-like method for large-scale unconstrained nonconvex minimization. And a new straightforward limited memory quasi-Newton updating based on the modified quasi-Newton equation is deduced to construct the trust region subproblem, in which the information of both the function value and gradient is used to construct approximate Hessian. The global convergence of the algorithm is proved. Numerical results indicate that the proposed method is competitive and efficient on some classical large-scale nonconvex test problems.

scholarly journals Convergence of relaxed chaotic parallel iterative methods

1994 ◽  
Vol 50 (1) ◽  
pp. 167-176 ◽  
Author(s):  
Peter E. Kloeden ◽  
Dong-Jin Yuan
Keyword(s):  
Iterative Methods ◽  
Large Scale ◽  
Linear Equations ◽  
Sufficient Conditions ◽  
System Of Linear Equations ◽  
Parallel Iterative Methods ◽  
Parallel Iterative

Sufficient conditions involving uniform multisplittings are established for the convergence of relaxed and AOR versions of asynchronous or chaotic parallel iterative methods for solving a large scale nonsingular system of linear equations Ax = b.

scholarly journals A New Version of the Accelerated Overrelaxation Iterative Method

2014 ◽  
Vol 2014 ◽  
pp. 1-6
Author(s):  
Shi-Liang Wu ◽  
Yu-Jun Liu
Keyword(s):  
Iterative Method ◽  
Linear Equations ◽  
Positive Definite ◽  
System Of Linear Equations ◽  
Positive Definite Matrices ◽  
Symmetric Positive Definite ◽  
Convergence Results ◽  
Aor Method ◽  
Symmetric Positive Definite Matrices ◽  
Overrelaxation Iterative Method

Hadjidimos (1978) proposed a classical accelerated overrelaxation (AOR) iterative method to solve the system of linear equations, and discussed its convergence under the conditions that the coefficient matrices are irreducible diagonal dominant,L-matrices, and consistently orders matrices. In this paper, a new version of the AOR method is presented. Some convergence results are derived when the coefficient matrices are irreducible diagonal dominant,H-matrices, symmetric positive definite matrices, andL-matrices. A relational graph for the new AOR method and the original AOR method is presented. Finally, a numerical example is presented to illustrate the efficiency of the proposed method.

scholarly journals Publicly verifiable and efficiency/security-adjustable outsourcing scheme for solving large-scale modular system of linear equations

2019 ◽  
Vol 8 (1) ◽  
Author(s):  
Panpan Meng ◽  
Chengliang Tian ◽  
Xiangguo Cheng
Keyword(s):  
Coding Theory ◽  
Large Scale ◽  
High Efficiency ◽  
Linear Equations ◽  
Coefficient Matrix ◽  
Secure Computation ◽  
Modular System ◽  
System Of Linear Equations ◽  
Public Verifiability ◽  
Cloud Server

AbstractSolving large-scale modular system of linear equations ($\mathcal {LMSLE}$ℒℳSℒE) is pervasive in modern computer and communication community, especially in the fields of coding theory and cryptography. However, it is computationally overloaded for lightweight devices arisen in quantity with the dawn of the things of internet (IoT) era. As an important form of cloud computing services, secure computation outsourcing has become a popular topic. In this paper, we design an efficient outsourcing scheme that enables the resource-constrained client to find a solution of the $\mathcal {LMSLE}$ℒℳSℒE with the assistance of a public cloud server. By utilizing affine transformation based on sparse unimodular matrices, our scheme has three merits compared with previous work: 1) Our scheme is efficiency/security-adjustable. Our encryption method is dynamic, and it can balance the security and efficiency to match different application scenarios by skillfully control the number of unimodular matrices. 2) Our scheme is versatile. It is suit for generic m-by-n coefficient matrix A, no matter it is square or not. 3) Our scheme satisfies public verifiability and achieves the optimal verification probability. It enables any verifier which is not necessarily the client to verify the correctness of the results returned from the cloud server with probability 1. Finally, theoretical analysis and comprehensive experimental results confirm our scheme’s security and high efficiency.

Fast Stokesian dynamics

2019 ◽  
Vol 878 ◽  
pp. 544-597 ◽  
Author(s):  
Andrew M. Fiore ◽  
James W. Swan
Keyword(s):  
Saddle Point ◽  
Large Scale ◽  
Near Field ◽  
Linear Equations ◽  
Far Field ◽  
System Of Linear Equations ◽  
Stokesian Dynamics ◽  
Brownian Force ◽  
Saddle Point Matrix ◽  
Stochastic Forces

We present a new method for large scale dynamic simulation of colloidal particles with hydrodynamic interactions and Brownian forces, which we call fast Stokesian dynamics (FSD). The approach for modelling the hydrodynamic interactions between particles is based on the Stokesian dynamics (SD) algorithm (J. Fluid Mech., vol. 448, 2001, pp. 115–146), which decomposes the interactions into near-field (short-ranged, pairwise additive and diverging) and far-field (long-ranged many-body) contributions. In FSD, the standard system of linear equations for SD is reformulated using a single saddle point matrix. We show that this reformulation is generalizable to a host of particular simulation methods enabling the self-consistent inclusion of a wide range of constraints, geometries and physics in the SD simulation scheme. Importantly for fast, large scale simulations, we show that the saddle point equation is solved very efficiently by iterative methods for which novel preconditioners are derived. In contrast to existing approaches to accelerating SD algorithms, the FSD algorithm avoids explicit inversion of ill-conditioned hydrodynamic operators without adequate preconditioning, which drastically reduces computation time. Furthermore, the FSD formulation is combined with advanced sampling techniques in order to rapidly generate the stochastic forces required for Brownian motion. Specifically, we adopt the standard approach of decomposing the stochastic forces into near-field and far-field parts. The near-field Brownian force is readily computed using an iterative Krylov subspace method, for which a novel preconditioner is developed, while the far-field Brownian force is efficiently computed by linearly transforming those forces into a fluctuating velocity field, computed easily using the positively split Ewald approach (J. Chem. Phys., vol. 146, 2017, 124116). The resultant effect of this field on the particle motion is determined through solution of a system of linear equations using the same saddle point matrix used for deterministic calculations. Thus, this calculation is also very efficient. Additionally, application of the saddle point formulation to develop high-resolution hydrodynamic models from constrained collections of particles (similar to the immersed boundary method) is demonstrated and the convergence of such models is discussed in detail. Finally, an optimized graphics processing unit implementation of FSD for mono-disperse spherical particles is used to demonstrated performance and accuracy of dynamic simulations of $O(10^{5})$ particles, and an open source plugin for the HOOMD-blue suite of molecular dynamics software is included in the supplementary material.

Sign in / Sign up

Export Citation Format

Share Document