Minimal residual method

The Minimal Residual Method or MINRES is a Krylov subspace method for the iterative solution of symmetric linear equation systems. It was proposed by mathematicians Christopher Conway Paige and Michael Alan Saunders in 1975.^[1]

In contrast to the popular CG method, the MINRES method does not assume that the matrix is positive definite, only the symmetry of the matrix is mandatory.

The GMRES method is essentially a generalization of MINRES for arbitrary matrices. Both minimize the 2-norm of the residual and do the same calculations in exact arithmetic when the matrix is symmetric. MINRES is a short-recurrence method with a constant memory requirement, whereas GMRES requires storing the whole Krylov space, so its memory requirement is roughly proportional to the number of iterations. On the other hand, GMRES tends to suffer less from loss of orthogonality.^{[1] [2]}

The MINRES method iteratively calculates an approximate solution of a linear system of equations of the form $${\displaystyle Ax=b,}$${\displaystyle Ax=b,} where ${\displaystyle A\in \mathbb {R} ^{n\times n}}${\displaystyle A\in \mathbb {R} ^{n\times n}} is a symmetric matrix and ${\displaystyle b\in \mathbb {R} ^{n}}${\displaystyle b\in \mathbb {R} ^{n}} a vector.

For this, the norm of the residual ${\displaystyle r(x):=b-Ax}${\displaystyle r(x):=b-Ax} in a ${\displaystyle k}${\displaystyle k}-dimensional Krylov subspace $${\displaystyle V_{k}=x_{0}+\operatorname {span} \{r_{0},Ar_{0}\ldots ,A^{k-1}r_{0}\}}$${\displaystyle V_{k}=x_{0}+\operatorname {span} \{r_{0},Ar_{0}\ldots ,A^{k-1}r_{0}\}} is minimized. Here ${\displaystyle x_{0}\in \mathbb {R} ^{n}}${\displaystyle x_{0}\in \mathbb {R} ^{n}} is an initial value (often zero) and ${\displaystyle r_{0}:=r(x_{0})}${\displaystyle r_{0}:=r(x_{0})}.

More precisely, we define the approximate solutions ${\displaystyle x_{k}}${\displaystyle x_{k}} through $${\displaystyle x_{k}:=\mathrm {argmin} _{x\in V_{k}}\|r(x)\|,}$${\displaystyle x_{k}:=\mathrm {argmin} _{x\in V_{k}}\|r(x)\|,} where ${\displaystyle \|\cdot \|}${\displaystyle \|\cdot \|} is the standard Euclidean norm on ${\displaystyle \mathbb {R} ^{n}}${\displaystyle \mathbb {R} ^{n}}.

Because of the symmetry of ${\displaystyle A}${\displaystyle A}, unlike in the GMRES method, it is possible to carry out this minimization process recursively, storing only two previous steps (short recurrence). This saves memory.

Note: The MINRES method is more complicated than the algebraically equivalent Conjugate Residual method. The Conjugate Residual (CR) method was therefore produced below as a substitute. It differs from MINRES in that in MINRES, the columns of a basis of the Krylov space (denoted below by ${\displaystyle p_{k}}${\displaystyle p_{k}}) can be orthogonalized, whereas in CR their images (below labeled with ${\displaystyle s_{k}}${\displaystyle s_{k}}) can be orthogonalized via the Lanczos recursion. There are more efficient and preconditioned variants with fewer AXPYs. Compare with the article.

First you choose ${\displaystyle x_{0}\in \mathbb {R} ^{n}}${\displaystyle x_{0}\in \mathbb {R} ^{n}} arbitrary and compute $${\displaystyle {\begin{aligned}r_{0}&=b-Ax_{0}\\p_{0}&=r_{0}\\s_{0}&=Ap_{0}\end{aligned}}}$${\displaystyle {\begin{aligned}r_{0}&=b-Ax_{0}\\p_{0}&=r_{0}\\s_{0}&=Ap_{0}\end{aligned}}}

Then we iterate for ${\displaystyle k=1,2,\dots }${\displaystyle k=1,2,\dots } in the following steps:

In the case of positive definite matrices, the convergence rate of the MINRES method can be estimated in a way similar to that of the CG method.^[3] In contrast to the CG method, however, the estimation does not apply to the errors of the iterates, but to the residual. The following applies:

$${\displaystyle \|r_{k}\|\leq 2\left({\frac {{\sqrt {\kappa (A)}}-1}{{\sqrt {\kappa (A)}}+1}}\right)^{k}\|r_{0}\|,}$${\displaystyle \|r_{k}\|\leq 2\left({\frac {{\sqrt {\kappa (A)}}-1}{{\sqrt {\kappa (A)}}+1}}\right)^{k}\|r_{0}\|,}

where ${\displaystyle \kappa (A)}${\displaystyle \kappa (A)} is the condition number of matrix ${\displaystyle A}${\displaystyle A}. Because ${\displaystyle A}${\displaystyle A} is normal, we have $${\displaystyle \kappa (A)={\frac {\left|\lambda _{\text{max}}(A)\right|}{\left|\lambda _{\text{min}}(A)\right|}},}$${\displaystyle \kappa (A)={\frac {\left|\lambda _{\text{max}}(A)\right|}{\left|\lambda _{\text{min}}(A)\right|}},} where ${\displaystyle \lambda _{\text{max}}(A)}${\displaystyle \lambda _{\text{max}}(A)} and ${\displaystyle \lambda _{\text{min}}(A)}${\displaystyle \lambda _{\text{min}}(A)} are maximal and minimal eigenvalues of ${\displaystyle A}${\displaystyle A}, respectively.

function[x, r] = minres(A, b, x0, maxit, tol)
x=x0;
r=b-A*x0;
p0=r;
s0=A*p0;
p1=p0;
s1=s0;
foriter=1:maxit
p2=p1;p1=p0;
s2=s1;s1=s0;
alpha=r'*s1/(s1'*s1);
x=x+alpha*p1;
r=r-alpha*s1;
if(r'*r<tol^2)
break
end
p0=s1;
s0=A*s1;
beta1=s0'*s1/(s1'*s1);
p0=p0-beta1*p1;
s0=s0-beta1*s1;
ifiter>1
beta2=s0'*s2/(s2'*s2);
p0=p0-beta2*p2;
s0=s0-beta2*s2;
end
end
end

• Minimal Residual Method, Wolfram MathWorld, Jul 26, 2022.

• Numerical linear algebra

• CS1 maint: location
• Articles with short description
• Short description matches Wikidata