Timeline for Any technique for linearization, or linear approximation?
Current License: CC BY-SA 4.0
13 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 2, 2021 at 19:53 | comment | added | Navid Hashemi | Hi Mark, I have changed my question to this new one, do you have any suggestion for improvement? | |
| Aug 2, 2021 at 19:50 | history | edited | Navid Hashemi | CC BY-SA 4.0 |
added 708 characters in body
|
| Aug 2, 2021 at 19:39 | comment | added | Navid Hashemi | Thanks Mark, for the valuable comment | |
| Aug 1, 2021 at 23:23 | comment | added | Mark L. Stone | That's not as proof of comvexity. An alternating variable scheme, as you seem to be describing, is a possible solution technique. I don't think it will generally come with a guarantee of converging to anything, let alone a global, or even local minimum. Nevertheless, if your computational experience with that approach is favorable, it may be that it works well in your case | |
| Aug 1, 2021 at 21:41 | comment | added | Navid Hashemi | I am pretty sure it is convex even if it is not linear. because I replaced $y^T Y^{-1}y$ with a new variable $s$ and solved the optimization. and then updated $s$ with the solution of $Y$ and $y$. I witnessed the solution of this itereative approach always converges to a unique solution no matter what I selected as the initial condition for $s$. | |
| Jul 30, 2021 at 19:49 | comment | added | Mark L. Stone | I don't think there is an y (exact) linearization given that Y and y are both variables. That is a (non-convex) nonlinear semidefinite constraint. Any linear approximation would be at best locally "valid". | |
| S Jul 30, 2021 at 18:31 | history | suggested | Jukka Kohonen | CC BY-SA 4.0 |
Clarify "it" & fix spelling.
|
| Jul 30, 2021 at 18:11 | review | Suggested edits | |||
| S Jul 30, 2021 at 18:31 | |||||
| Jul 30, 2021 at 17:58 | history | edited | YCor | CC BY-SA 4.0 |
removed capitals from title
|
| Jul 30, 2021 at 16:21 | comment | added | Navid Hashemi | My first aim is to find a linear equivallent to this constraint, but if that is impossible, I am interested in a linear constraint which is a good approximation of this nonlinear constraint. | |
| Jul 30, 2021 at 16:15 | comment | added | Ben McKay | What are we approximating? | |
| Jul 30, 2021 at 15:25 | review | First posts | |||
| Jul 30, 2021 at 18:09 | |||||
| Jul 30, 2021 at 15:20 | history | asked | Navid Hashemi | CC BY-SA 4.0 |