I am using the convex-concave procedure to solve a problem. For the first few iterations, the code works fine. However as we proceed, some iterations yield Inaccurate/Solved status. Furthermore, the value of an optimization variable does not converge. Does anyone have a solution to this problem?
The convex-concave procedure is not guaranteed to converge to anything, let alone a global or even local optimum of the original problem…Among other things, it has no safeguarding of steps, such as by trust region or line search; and therefore is not robust.
Choice of starting point does matter,. Scaling may matter.
Thanks for the reply.
Do you have any idea of how to overcome the issues in CCP?
Use a good off-the-shelf nonlinear optimizer, rather than trying to gin up your own.
The CCP results you may have seem may be for a biased set of problems for which some crude procedure happens to work o.k.