I have a question concerning the derivatives of approximate matrix inverses. I have a system, $$Ax = b$$ which I solve approximately (and with an iterative method) with: $$\Delta x = x - \tilde{A}^{-1}b$$ I would like to take the derivative of this process, i.e. find $$\frac{d\Delta x}{dD} = \frac{dx}{dD} - \frac{d\tilde{A}^{-1}}{dD}b - \tilde{A}^{-1}\frac{db}{dD}$$ I know that if I had an ideal inverse, my expression would be without tildes as I'd have the exact result: $$\frac{d\Delta x}{dD} = \frac{dx}{dD} - {A}^{-1}\frac{dA}{dD}{A}^{-1}b - \tilde{A}^{-1}\frac{db}{dD}$$ But for this, I cannot start with the typical assumption that $$AA^{-1} = I$$ and instead have $$A\tilde{A}^{-1} = C \neq I$$ Does anyone have any ideas for resources or techniques for such problems? Thank you.

  • $\begingroup$ What is $D$?$\;\;\;$ $\endgroup$ – Algebraic Pavel Oct 16 '18 at 8:52
  • $\begingroup$ The variable that I'm taking the derivative with respect to. The state variables depend on it through the nonlinear solve. $\endgroup$ – EMP Oct 16 '18 at 13:04
  • $\begingroup$ what iterative method did you use? $\endgroup$ – Shogun Oct 16 '18 at 22:41
  • 2
    $\begingroup$ I used Point Jacobi as my iterative solver $\endgroup$ – EMP Oct 17 '18 at 23:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.