What is the state of art gradient based algorithms in convex optimization solving non-smooth piece-wise linear functions? Thank you.

EDIT: It is different from one of my previous post in the sense that here, I am focusing on piece-wise linear functions:D


There are two cases I can think of (assuming minimization):

  • your non-smooth piecewise linear function is convex, in which case, standard reformulations result in a convex program (or even a linear program), so you can consult a textbook in convex optimization such as the one by Boyd to find applicable algorithms.

  • your non-smooth piecewise linear function is nonconvex; then the reformulation I mentioned above doesn't apply, and you probably need to provide more about the problem structure.

| cite | improve this answer | |
  • $\begingroup$ My problem is convex. The problem is that, I intuitively think I shall use Nesterov's accelerated gradient method... But the book hasn't introduced that... And I am wondering whether there is even improvements on Nesterov's method to specially tackle convex piece-wise linear problems... $\endgroup$ – user40780 Mar 21 '15 at 0:28

Your Answer

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

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