Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Guys I was reading about CG method to solve the sparse systems. I came across that the method is defined for positive definite symmetric matrices. I was wondering does it converges for negative definite matrices too ?

share|cite|improve this question
Not directly, CG minimizes a convex quadratic. However, you could solve $(-A)x = (-b)$, or modify the algorithm by changing signs at appropriate places (as in the difference between minimization and maximization is 'just' a sign difference). – copper.hat Nov 16 '12 at 4:13

$Ax =b \iff (-A) x = -b$, and if $A$ is symmetric negative definite then $-A$ is symmetric positive definite. So you can use CG to solve $-Ax = -b$.

share|cite|improve this answer
So for convergence we compute 2-norm(b-Ax) < tolerance value. If yes then we say that it has converged to a solution. In this case we need to check for 2-norm( (-b) - (-Ax)) < tolerance, which is as good as 2-norm (Ax - b).. is that right ? – Anonym Nov 16 '12 at 4:16
Yes, sounds good. – littleO Nov 16 '12 at 4:19
ok. And what if we would like to solve for non-symmetric full-rank matrices using CG? In that case can we solve Ax = b as A'Ax = A'b using CG ? – Anonym Nov 16 '12 at 4:26
Thank You all for your help. – Anonym Nov 16 '12 at 4:39
@Anonym: If the Answer littleO gave is satisfactory, you can express your approval by Accepting it, which will also help others who may come across this thread to know it was well resolved. – hardmath Nov 16 '12 at 23:33

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.