Questions tagged [conjugate-gradient]

A popular krylov subspace method for solving linear systems of equations, particularly those that exhibit symmetric positive definiteness.

25
votes
3answers
19k views

BFGS vs. Conjugate Gradient Method

What considerations should I be making when choosing between BFGS and conjugate gradient for optimization? The function I am trying to fit with these variables are exponential functions; however, the ...
23
votes
3answers
3k views

What is the principle behind the convergence of Krylov subspace methods for solving linear systems of equations?

As I understand it, there are two major categories of iterative methods for solving linear systems of equations: Stationary Methods (Jacobi, Gauss-Seidel, SOR, Multigrid) Krylov Subspace methods (...
17
votes
3answers
3k views

Problems where Conjugate gradient works much better than GMRES

I am interested in cases where Conjugate gradient works much better than GMRES method. In general, CG is preferable choice in many cases of SPD (symmetric-positive-definite) because it requires less ...
11
votes
4answers
794 views

Calculating determinant while solving $Ax=b$ using CG

I am solving $Ax=b$ for a huge sparse positive definite matrix $A$ using the conjugate gradient (CG) method. It is possible to compute the determinant of $A$ using the information produced during the ...
10
votes
2answers
6k views

Gradient descent and conjugate gradient descent

For a project, I have to implement these two methods and compare how they perform on different functions. It looks like the conjugate gradient method is meant to solve systems of linear equations of ...
9
votes
1answer
1k views

Adaptive gradient descent step size when you can't do a line search

I have an objective function $E$ dependent on a value $\phi(x, t = 1.0)$, where $\phi(x, t)$ is the solution to a PDE. I am optimizing $E$ by gradient descent on the initial condition of the PDE: $\...
9
votes
2answers
4k views

How to remove Rigid Body Motions in Linear Elasticity?

I want to solve $K u = b$ where $K$ is my stiffness matrix. However some constraints may be missing an therefore some rigid body motion may be still present in the system (due to eigenvalue zero). ...
8
votes
2answers
3k views

Genetic algorithm vs conjugate gradient method

I am trying to optimize some force-field parameters in a molecular framework so that the result of simulation comes as close as it can to the experimental structure. In the past, I have written a ...
8
votes
2answers
2k views

What's the difference between conjugate gradient method and biconjugate gradient method

What's the difference between these two methods? Can a problem be solved by one method will be able to solved by the other? Can both/or one of them be parallelized with OpenMP and/or MPI?
8
votes
2answers
1k views

What is the worst case complexity of Conjugate Gradient?

Let $A\in \mathbb{R}^{n\times n}$, symmetric and positive definite. Suppose it takes $m$ units of work to multiply a vector by $A$. It is well known that performing the CG algorithm on $A$ with ...
8
votes
2answers
261 views

Why does conjugate gradient work with this nonsymmetric preconditioner?

In this previous thread the following multiplicative way to combine symmetric preconditioners $P_1$ and $P_2$ for the symmetric system $Ax=b$ was suggested: \begin{align} P_\text{combo}^{-1} :=& ...
8
votes
0answers
124 views

How to find a good preconditioner to the system $(A^T A + \lambda I) x = A^T b$?

The system in the title has a damper factor $\lambda > 0$ and the matrix $A$ is sparse and rectangular, with a structure I can exploit to solve matrix vector products very fast. My current solver, ...
7
votes
1answer
3k views

What are the differences between the different gradient-based numerical optimization methods?

I am interested in the specific differences of the following methods: The conjugate gradient method (CGM) is an algorithm for the numerical solution of particular systems of linear equations. The ...
6
votes
1answer
223 views

What is required of the objective function in order to use Gauss Newton method?

From what I understand, the Gauss-Newton method is used to find a search direction, then the step size, etc., can be determined by some other method. In addition to that, are the following ...
6
votes
1answer
321 views

Caveats of Hessian free method

Hessian free iterative optimization techniques like Newton-CG, do not explicitly compute the Hessian but instead approximate the product of the Hessian with a vector through finite difference. The ...
6
votes
1answer
552 views

Estimate extreme eigenvalues with CG

CG may be used to estimate the extremal eigenvalues of a SPD matrix (by computing eigenvalues of tridiagonal matrix associated with the Lanczos algorithm). After a few iterations the largest ...
5
votes
2answers
1k views

Sufficient conditions to ensure convergence of the conjugate gradient method

I know that a conjugate gradient method is guaranteed to converge to the solution of a linear system if the matrix is positive definite. I'm working with a family of matrices that have the following ...
5
votes
2answers
159 views

Would recalculating the residual in the conjugate gradient method help?

The conjugate gradient method suffers from an accumulation of errors as it continues. For this reason it is unwise to use it as a direct solver. My question is, would it help to recalculate the ...
5
votes
1answer
830 views

Solving a system of nonlinear PDEs by minimization

I have two coupled nonlinear partial differential equations of the form: $ \begin{align} \dot{u} -f(u,u',u'',v,v',v'')=0 \\ \dot{v} -g(u,u',u'',v,v',v'')=0 \end{align} $ The boundary conditions are ...
5
votes
1answer
270 views

Solving a set of linear equations with block structure and weak coupling

I have a standard set of linear equations $Ax=b$ where the Hessian matrix $A$ has the special block structure as shown: $A= \begin{pmatrix} T & U\\ U^T & V \end{pmatrix}$, $x= \begin{...
5
votes
0answers
178 views

Conjugate residual/gradient convergence checking in practice

Let's say we want to solve $Ax=b$ ($A$ symmetric positive /semi/definite) with the conjugate residual/gradient method. $A$ comes from FEM where the mesh is being refined. The exact solution is $x_*$ ...
4
votes
2answers
751 views

nonlinear conjugate gradient for multivariable functions

For the optimization problem $\underset{\mathbf{x}\in \mathbb{R}^n}{\operatorname{argmin}} f(\mathbf{x})$, we can use the following standard nonlinear conjugate gradient method to find the solution: $...
4
votes
2answers
193 views

Looking for a mathematical proof of stability in floating point arithmetic of CG - any reference?

I am looking for a reference - paper, book, discussion, anything that has a mathematical proof for stability of the conjugate gradient method in floating point arithmetic. Something similar for ...
4
votes
2answers
327 views

best way to optimize a function with linear/non-linear parameters

I am trying to fit some raw data using a function of the form $f(r) = \sum_{i=1}^{K} d_kS_k(n_k,\alpha_k,r)$ where $S_k(n_k,\alpha_k,r) = \frac{\alpha_k ^{n_k+3}}{(n_k+2)!}r^{n_k}\exp(-\alpha_kr)$ ...
4
votes
1answer
299 views

Linear equation system: Direct solver works, iterative solver does not

I have to solve for x in b = A*x, where a is sparse. This works fine with Matlab's mldivide: x = A \ b. Since I will have to use an iterative algorithm for very large A, I'm currently testing Matlab's ...
4
votes
1answer
166 views

Is there a nonlinear solver similar to CGNR evaluating only $A^TAx$?

First of all, I am quite new to this field and I excuse myself in advance for any stupid content in this question. In the field of compressed sensing or deblurring I have a nonlinear optimization ...
4
votes
2answers
843 views

Conjugate gradient method to minimize a function

I am having some serious difficulties trying to understand how to use (apply) CG to minimize a function. In all the textbooks and notes, the step size $\alpha$ is give by the following expression $$ \...
4
votes
0answers
192 views

Nonlinear conjugate gradient restart threshold 1/10

Nocedal and Wright on Conjugate Gradient Methods, p. 123, describe a restart strategy ... whenever two consecutive gradients are far from orthogonal $\qquad {{| \nabla f_k^T \ \nabla f_{k-1} |} \...
4
votes
0answers
231 views

Is it possible to predict the null space of a structure from contributing elements null spaces?

I am trying to solve an almost incompressible problem with heterogeneous properties by domain decomposition. Solution with CG converges slowly or divergerces completely. My problem becomes ill-...
3
votes
2answers
5k views

What are some reasons that Conjugate Gradient iteration does not converge?

I would greatly appreciate it if you could share some reasons the Conjugate Gradient iteration for Ax = b does not converge? My matrix A is symmetric positive definite. Thank you so much! Edit with ...
3
votes
2answers
106 views

How can you explain the following bound on the inner product?

I am reading a paper on stability of CG, and I came across the following statement: \begin{equation} \frac{\|A\|\,\|p\|^2}{\langle p,Ap\rangle} \leq \kappa(A) \end{equation} where $\kappa(\cdot)$ is ...
3
votes
1answer
465 views

Which C++ linear algebra library is probably the fastest on solving huge sparse [square matrix] linear system?

I am developing a 2D CFD solver for fluid-particle interaction. To solve Navier-Stokes equations on a grid of size $10000\times 10000$ cells (or >1 million cells), a large linear system $Ax=b$ with $A$...
3
votes
1answer
133 views

CG question: is symmetry always necessary?

Consider the 1D Poisson equation $$\nabla^2 u = f.$$ Using finite difference method on cell corner data and a uniform grid with ghost points, I think we can write the system of equations with Neumann ...
3
votes
1answer
842 views

Is a checkerboard block decomposition of a matrix useful when solving a linear system with a parallel conjugate gradient method?

According to these lecture notes, a checkerboard block decomposition should exhibit better scalability when applied to parallel matrix-vector multiplication (presumably because of greater cache hit ...
3
votes
1answer
267 views

Linear constraints for L-BFGS-B

I know L-BFGS-B only supports simple box constrains of the form: $l_i \leq x_i \leq u_i$, where $l_i$ and $u_i$ are constants. For my specific optimization problem, I need to specify some simple ...
3
votes
2answers
283 views

2D Laplace problem with mixed boundary conditions using Conjugate Gradients

I am being asked for one of my classes to solve 2D Laplace equations with mixed boundary conditions using the Conjugate Gradient method. The equations and conditions are given as: $$ \frac{\partial^...
3
votes
1answer
199 views

Trust-region Newton: implementation issue with Conjugate Gradient calculations

UPDATE: The problem turned out to be the step (refer penultimate paragraph below) where I was factoring out a small value from the vectors of the numerator and denominator and then computed dot ...
3
votes
0answers
32 views

Stochastic conjugate directions to improve convergence in narrow valleys

My question concerns a specific statement in this paper: N. N. Schraudolph and T. Graepel, "Conjugate Directions for Stochastic Gradient Descent," in Int. Conf. Artificial Neural Networks, Berlin, ...
3
votes
0answers
1k views

Understanding MATLAB's `fmincg` optimization function

I'm researching numerical optimization. Recently I've come across a variant of a conjugate gradient method named fmincg. The function is written in MATLAB and is ...
3
votes
0answers
103 views

Conjugate gradient: the 1-norm of the residual

I am trying to solve $Ax=b$ using the conjugate gradient method. However, it is important to me to obtain a bound not only on the usual residual $||b-Ax_k||_2$ but also on the quantity $||b-Ax_k||_1$. ...
3
votes
0answers
89 views

How to implement conjugate gradient method to minimize this nonlinear action?

Given a 2D stochastic differential equation: \begin{align} \dot{x}_{i}=f_{i}(\textbf{x})+g_{ij}\xi_{j}(t), \end{align} where $i=2$, $g_{ij}g_{jk}=2\epsilon\delta_{ik}$ and $f(\textbf{x})=-\nabla\phi(\...
2
votes
1answer
105 views

How can a CG solver solve a non positive definite sparse matrix

I am using the CUSP CG solver and I ran it on couple of sparse matrices from the University of Florida sparse matrix collection. The solver was able to solve non positive definite sparse matrices. My ...
2
votes
1answer
421 views

Hessian-free and Truncated Newton methods

In this paper on Deep Learning for Machine Learning, the approach is referred to as Hessian-free method. That is because the Hessian is never computed explicitly. Instead, the product of the Hessian ...
2
votes
1answer
434 views

PETSc Krylov Subspace and nullspace

How is the nullspace correction implemented in the PETSc conjugate gradient solver? I searched the source code and documentation, but could not find references on the actual implementation. For ...
2
votes
2answers
1k views

Usage of DifferentiableMultivariateFunction and NonLinearConjugateGradientOptimizer

I try to use the matlab fminunc functionality in java and found the optimisation functions in commons-math. I have absolutely no clue how to use them right because the example tells nothing about the ...
2
votes
1answer
44 views

Blowup of error in Conjugate Gradient method with periodic Dirichlet Poisson matrix

My problem is that the L2-Norm of the residual for the periodic Poisson matrix $P$ is initially decreasing but starts to blow up after a certain number of iterations. The blowup happens earlier the ...
2
votes
1answer
66 views

Nonlinear conjugate gradient with orthogonality constraint

I have to solve a set of nonlinear optimization problems in the subspace defined as the orthogonal space to a given vector. More precisely, $$ \arg\min f(\vec x) \qquad \text{with} \qquad \vec x \...
2
votes
1answer
125 views

Solve FEM matrix from coupled system

I'm developing an FEM solver for a coupled system. I have diffusion and potential equations which result in positive definite matrices for each equation, but the coupling makes the overall system ...
2
votes
0answers
60 views

Richardson's Iteration, Gradient Method and Spectral Radius

Richardson's iteration introduce a scalar $\alpha$ to the update formula: $$ \textbf{x}^{(k+1)} = \textbf{x}^{(k)} + \alpha \textbf{r}^{(k)} $$ And compute $\alpha$ by minimizing the spectral radius:...
2
votes
0answers
94 views

Optimization based integration for MPM

I'm considering implementing (just for simplicity) the unconstrained implicit optimization based integration for Material Point Method as described in Chenfanfu Jiang's thesis on MPM (the minimization ...